-*- mode: org -*-
Engineering, manufacturing, logistics
*  14(<-592): Association rules mining based on SVM and its application in simulated moving bed PX adsorption process

In this paper, a novel data mining method is introduced to solve the
multi-objective optimization problems of process industry. A
hyperrectangle association rule mining (HARM) algorithm based on
support vector machines (SVMs) is proposed. Hyperrectangles rules are
constructed on the base of prototypes and support vectors (SVs) under
some heuristic limitations. The proposed algorithm is applied to a
simulated moving bed (SMB) paraxylene (PX) adsorption process. The
relationships between the key process variables and some objective
variables such as purity, recovery rate of PX are obtained. Using
existing domain knowledge about PX adsorption process, most of the
obtained association rules can be explained.

2005


*  18(<- 41): Investigation of a Bridge Pier Scour Prediction Model for Safe Design and Inspection

A novel bridge scour estimation approach that comprises advantages of
both empirical and data-driven models is developed here. Results from
the new approach are compared with existing approaches. Two field
datasets from the literature are used in this study. Support vector
machine (SVM), which is a machine-learning algorithm, is used to
increase the pool of field data samples. For a comprehensive
understanding of bridge-pier-scour modeling, a model evaluation
function is suggested using an orthogonal projection method on a model
performance plot. A fast nondominated sorting genetic algorithm
(NSGA-II) is evaluated on the model performance objective functions to
search for Pareto optimal fronts. The proposed formulation is compared
with two selected empirical models [Hydraulic Engineering Circular
No. 18 (HEC-18) and Froehlich equation] and a recently developed
data-driven model (gene expression programming model). Results show
that the proposed model improves the estimation of critical scour
depth compared with the other models.

2015



*  21(<-314): Research of load identification based on multiple-input multiple-output SVM model selection

In this article, the problem of multiple-input multiple-output (MIMO)
load identification is addressed. First, load identification is proved
in dynamic theory as non-linear MIMO black-box modelling
process. Second, considering the effect of hyper-parameters in
small-size sample problem, a new MIMO Support Vector Machine (SVM)
model selection method based on multi-objective particle swarm
optimization is proposed in order to improve the identification's
performance. The proposed method treats the model selection of MIMO
SVM as a multi-objective optimization problem, and leave-one-out
generalization errors of all output models are minimized
simultaneously. Once the Pareto-optimal solutions are found, the SVM
model with the best generalization ability is determined. The proposed
method is evaluated in the experiment of dynamic load identification
on cylinder stochastic vibration system, demonstrating its benefits in
comparison to the existing model selection methods in terms of
identification accuracy and numerical stability, especially near the
peaks.

2012


*  22(<-138): Applying multiple kernel learning and support vector machine for solving the multicriteria and nonlinearity problems of traffic flow prediction

This article proposes to develop a prediction model for traffic flow
using kernel learning methods such as support vector machine (SVM) and
multiple kernel learning (MKL). Traffic flow prediction is a dynamic
problem owing to its complex nature of multicriteria and
nonlinearity. Influential factors of traffic flow were firstly
investigated; five-point scale and entropy methods were employed to
transfer the qualitative factors into quantitative ones and rank these
factors, respectively. Then, SVM and MKL-based prediction models were
developed, with the influential factors and the traffic flow as the
input and output variables. The prediction capability of MKL was
compared with SVM through a case study. It is proved that both the SVM
and MKL perform well in prediction with regard to the accuracy rate
and efficiency, and MKL is more preferable with a higher accuracy rate
when under proper parameters setting. Therefore, MKL can enhance the
decision-making of traffic flow prediction. Copyright (c) 2012 John
Wiley & Sons, Ltd.

2014


*  25(<-593): A numerical investigation of mixing processes in a novel combustor application

A mixing process in a staggered toothed-indented shaped channel was
investigated. It was studied in two steps: (1) numerical simulations
for different sizes of the boundary-contour were performed by using a
CFD code; (2) these results were used for simulation-data modeling for
prediction of mixing performances across the whole field of changing
geometric and the aerodynamic stream parameters. Support vector
machine (SVM) technique, known as a new type of self learning machine,
was selected to carry out this stage. The suitability of this
application method was demonstrated in comparison with a neural
network (NN) method. The established modeling system was then applied
to some further studies of the prototype mixer including observations
of the mixing performance in three special cases and performing
optimizations of the mixing processes for two conflicting objectives
and hereby obtaining the Pareto optimum sets.

2005


*  39(<-345): A new feature extraction and selection scheme for hybrid fault diagnosis of gearbox

A novel feature extraction and selection scheme was proposed for
hybrid fault diagnosis of gearbox based on S transform, non-negative
matrix factorization (NMF), mutual information and multi-objective
evolutionary algorithms. Time-frequency distributions of vibration
signals, acquired from gearbox with different fault states, were
obtained by S transform. Then non-negative matrix factorization (NMF)
was employed to extract features from the time-frequency
representations. Furthermore, a two stage feature selection approach
combining filter and wrapper techniques based on mutual information
and non-dominated sorting genetic algorithms II (NSGA-II) was
presented to get a more compact feature subset for accurate
classification of hybrid faults of gearbox. Eight fault states,
including gear defects, bearing defects and combination of gear and
bearing defects, were simulated on a single-stage gearbox to evaluated
the proposed feature extraction and selection scheme. Four different
classifiers were employed to incorporate with the presented techniques
for classification. Performances of four classifiers with different
feature subsets were compared. Results of the experiments have
revealed that the proposed feature extraction and selection scheme
demonstrate to be an effective and efficient tool for hybrid fault
diagnosis of gearbox. (C) 2011 Elsevier Ltd. All rights reserved.

2011


*  52(<- 20): Application of multi-stage multi-objective multi-disciplinary agent model based on dynamic substructural method in Mistuned Blisk

A method called multi-stage multi-objective multi-disciplinary agent
model based on dynamic substructural method (MSMOMDAM-DSM) is
proposed. For a large amount of calculation, it can increase the
mistuned blisk's computational efficiency more significantly comparing
with the traditional probability analysis method when the request of
computational accuracy is satisfied. Deterministic analysis is
investigated based on the improved hybrid interface substructural
component modal synthesis method (IHISCMSM). It demonstrates that the
symmetry is broken and the localization phenomenon is observed when
the blisk is mistuned. Meanwhile, the frequency response function
(FRF) manifests that multiple peaks are observed in different
frequencies when the mistuning level is greater than 0% which are
caused by the combined behavior of resonance and mistuning. On the
base of deterministic analysis, the most dangerous point of mistuned
blisk is extracted as response amplitude and the probability analysis
is investigated. The probability density distribution function (PDDF)
of random variable is given and the limit state equation (LSE) of
radial deformation, sampling history, simulation sample, cumulative
distribution function (CDF) and histogram distribution are
obtained. The sensitivity and the scatter diagram are also analyzed
which manifests that some variables are positive while some are
negative and the rotational speed is the importance
degree. Probability design and inverse probability design are also
researched which lay a solid foundation for designing safe and
reasonable blisk. At last, it is verified that the computational
accuracy and efficiency of MSMOMDAM-DSM is superior to support vector
machine-response surface method (SVM-RSM). (C) 2015 Elsevier Masson
SAS. All rights reserved.

2015



*  90(<- 31): Research on the route optimization for fresh air processing of air handling unit in spacecraft launching site

The existing control methods for Air handling units (AHUs) in
spacecraft launching site (SLS) are comparatively dated, the air
processing routes are usually arbitrarily determined in line with
experience, which fails to cope with the coupling and function
redundancy of air condition system and the diversity of outdoor
environment, therefore resulting in tremendous energy waste. This
paper proposes a new route optimization strategy for fresh air
processing-Firstly analyzing the possibly processing routes for fresh
air based on psychrometric chart, then proposing an optimization
algorithm AFSA-GA to optimize the possibly processing routes,
eventually obtaining the best route that requires the least energy
consumption. By adopting the strategy proposed to optimize air
processing route for High-Temperature and High-Humidity working
conditions, it can be proved that the proposed strategy can decrease
considerable amount of energy consumption, and the proposed
optimization algorithm AFSA-GA has the advantages of faster
convergence speed and avoiding premature. (C) 2015 Elsevier Ltd. All
rights reserved.

2015



*  91(<- 51): Multi-objective optimization of the HVAC (heating, ventilation, and air conditioning) system performance

A data-driven approach to optimize the total energy consumption of the HVAC (heating, ventilation, and air conditioning) system in a typical office facility is presented. A multi-layer perceptron ensemble is selected to build the total energy model integrating three indoor air quality models, the facility temperature model, the facility relative humidity model, and the facility CO2 concentration model. To balance the energy consumption and the indoor air quality, a quad-objective optimization problem is constructed. The problem is solved with a modified particle swarm optimization algorithm producing control settings of supply air temperature and static pressure of the air handling unit. By assigning different weights to the objectives to the model, the generated control settings optimize HVAC system with the trade-off between the energy consumption and the facility thermal comfort. Significant energy savings can be obtained even with air quality constraint. (C) 2015 Elsevier Ltd. All rights reserved.

2015





*  92(<- 81): Optimal allocation and adaptive VAR control of PV-DG in distribution networks

The development of distributed generation (DG) has brought new
challenges to power networks. One of them that catches extensive
attention is the voltage regulation problem of distribution networks
caused by DG. Optimal allocation of DG in distribution networks is
another well-known problem being widely investigated. This paper
proposes a new method for the optimal allocation of photovoltaic
distributed generation (PV-DG) considering the non-dispatchable
characteristics of PV units. An adaptive reactive power control model
is introduced in PV-DG allocation as to balance the trade-off between
the improvement of voltage quality and the minimization of power loss
in a distribution network integrated with PV-DG units. The optimal
allocation problem is formulated as a chance-constrained stochastic
programming (CCSP) model for dealing with the randomness of solar
power energy. A novel algorithm combining the multi-objective particle
swarm optimization (MOPSO) with support vector machines (SVM) is
proposed to find the Pareto front consisting of a set of possible
solutions. The Pareto solutions are further evaluated using the
weighted rank sum ratio (WRSR) method to help the decision-maker
obtain the desired solution. Simulation results on a 33-bus radial
distribution system show that the optimal allocation method can fully
take into account the time-variant characteristics and probability
distribution of PV-DG, and obtain the best allocation scheme. (C) 2014
Elsevier Ltd. All rights reserved.

2015


*  93(<- 86): Optimization of Liquid Desiccant Regenerator with Multiobject Particle Swarm Optimization Algorithm

In this paper, a model-based optimization strategy for a liquid
desiccant regenerator operating with lithium chloride solution is
presented. By analyzing the characteristics of the components, such as
electric heater, pump, and fan, energy predictive models for the
components in the regenerator are developed. To minimize the energy
usage while maintaining the regeneration rate within an accepted
level, one multiobjective optimization problem is formulated with two
objectives, the constraints of decision variables, components
interactions, and the outdoor conditions. A multiobjective
optimization strategy based on decreasing inertia weight particle
swarm optimization (DIWPSO) is proposed to obtain the optimal
nondominated solutions of the optimization problem, and a decision
making strategy is introduced to select the final solution, desiccant
solution flow rate, desiccant solution temperature, and the
regenerating air flow rate, to minimize the energy usage in the
regenerator. Experimental studies are carried out on an existing
system to compare the energy consumption and regeneration rate between
the proposed optimization strategy and conventional strategy to
evaluate energy saving performance of the proposed
strategy. Experimental results demonstrate that an average of 8.55%
energy can be saved by implementing the proposed optimization strategy
in liquid desiccant regenerator.

2014


*  94(<-116): Incomplete information-based decentralized cooperative control strategy for distributed energy resources of VSI-based microgrids

This paper presents an effective method to control distributed energy
resources (DERs) installed in a microgrid (MG) to guarantee its
stability after islanding occurrence. Considering voltage and
frequency variations after islanding occurrence and based on stability
criteria, MG pre-islanding conditions are divided into secure and
insecure classes. It is shown that insecure MG can become secure, if
appropriate preventive control is applied on the DERs in different
operating conditions of the MG. To select the most important variables
of MG, which can estimate proper values of output power set points of
DERs, a feature selection procedure known as symmetrical uncertainty
is used in this paper. Among all the MG variables, critical ones are
selected to calculate the appropriate output power of different DERs
for different conditions of the MG. The values of selected features
are transmitted by the communication system to the control unit
installed on each DER to control its output power set point. In order
to decrease the communication system cost, previous researchers have
used local variables to control the set point of different DERs. This
approach decreases the accuracy of the controller because the
controller uses incomplete information. In this paper, multi-objective
approach is used in order to decrease the cost of the communication
system, while keeping the accuracy of the preventive control strategy
in an allowable margin. The results demonstrate the effectiveness of
the proposed method in comparison with other methods.

2014



*  95(<-161): Feasibility study and performance assessment for the integration of a steam-injected gas turbine and thermal desalination system

This study proposes a systematic approach for retrofitting a
steam-injection gas turbine (SIGT) with a multieffect thermal vapor
compression (METVC) desalination system. The retrofitted unit's
product cost of the fresh water (RUPC) was used as a performance
criterion, which comprises the thermodynamic, economic, and
environmental attributes when calculating the total annual cost of the
SIGT-METVC system. For the feasibility study of retrofitting the SIGT
plant with the METVC desalination system, the effects of two key
parameters were analyzed using response surface methodology (RSM)
based on a central composite design (CCD): the steam air ratio (SR)
and the temperature difference between the effects of the METVC system
(Delta T-METVC) on the fresh water production (Q(freshwater)) and the
net power generation (W-net) of the SIGT-METVC system. Multi-objective
optimization (MOO) which minimizes the modified total annual cost
(MTAC) and maximizes the fresh water flow rate was performed to
optimize the RUPC of the SIGT-METVC system. The best Pareto optimal
solution showed that the SIGT-METVC system with five effects is the
best one among the systems with 4-6 effects. This system under optimal
operating conditions can save 21.07% and 9.54% of the RUPC, compared
to the systems with four and six effects, respectively. (C) 2013
Elsevier B.V. All rights reserved.

2014





*  96(<-193): Exergy analysis and parametric optimization of three power and fresh water cogeneration systems using refrigeration chillers

Three power and fresh water cogeneration systems that combine a GT
(gas turbine) power plant and a RO (reverse osmosis) desalination
system were compared based on the exergy viewpoint. In the first
system, the GT and RO systems were coupled mechanically to form a base
system. In the second and third systems, a VCR (vapor-compression
refrigeration) cycle and a single-effect AC(Water-LiBr) (water/lithium
bromide absorption chiller) were used, respectively, to cool the
compressor inlet air and preheat the RO intake seawater via waste heat
recovery in the VCR condenser and AC(Water-LiBr) absorber. A
parametric analysis-based exergy was conducted to evaluate the effects
of the key thermodynamic parameters including the compressor inlet air
temperature and the fuel-mass flow rate on the system exergy
efficiency. Parameter optimization was achieved using a GA (genetic
algorithm) to reach the maximum exergy efficiency, where the
thermodynamic improvement potentials of the systems were
identified. The optimum values of performance for the three
cogeneration systems were compared under the same conditions. The
results showed that the cogeneration system with the AC is the best
system among the three systems, since it can increase exergy and
energy efficiencies as well as net power generation by 3.79%, 4.21%,
and 38%, respectively, compared to the base system. (C) 2013 Elsevier
Ltd. All rights reserved.

2013



*  97(<-357): Multi-objective optimization of HVAC system with an evolutionary computation algorithm

A data-mining approach for the optimization of a HVAC (heating,
ventilation, and air conditioning) system is presented. A predictive
model of the HVAC system is derived by data-mining algorithms, using a
dataset collected from an experiment conducted at a research
facility. To minimize the energy while maintaining the corresponding
IAQ (indoor air quality) within a user-defined range, a
multi-objective optimization model is developed. The solutions of this
model are set points of the control system derived with an
evolutionary computation algorithm. The controllable input variables
supply air temperature and supply air duct static pressure set points
are generated to reduce the energy use. The results produced by the
evolutionary computation algorithm show that the control strategy
saves energy by optimizing operations of an HVAC system. (C) 2011
Elsevier Ltd. All rights reserved.

2011




*  98(<-470): Optimization of coal-fired boiler SCRs based on modified support vector machine models and genetic algorithms

An integrated combustion optimization approach is presented for the combined considering the trade offs in optimization of coal-fired boiler and selective catalyst reaction (SCR) system, to balance the unit thermal efficiency, SCR reagent consumption and NO, emissions. Field tests were performed at a 160 MW coal-fired unit to investigate the relationships between process controllable variables, and optimization targets and constraints. Based on the test data, a modified on-line support vector regression model was proposed for characteristic function approximation, in which the model parameters can be continuously adapted for changes in coal quality and other conditions of plant equipment. The optimization scheme was implemented by a genetic algorithm in two stages. Firstly, the multi-objective combustion optimization problem was solved to achieve an optimal Pareto front, which contains optimal solutions for lowest unit heat rate and lowest NO, emissions. Secondly, best operating settings for the boiler, and SCR system and air preheater were obtained for lowest operating cost under the constraints of NO, emissions limit and air preheater ammonium bisulfate deposition depth. (c) 2008 Elsevier Ltd. All rights reserved.

2009





*  99(<-413): Applying multiobjective RBFNNs optimization and feature selection to a mineral reduction problem

The Nickel reduction process is a complex task where many dynamic
optimization problems arises that, nowadays, requires a human operator
to take decisions based on his experience and intuition. In order to
help the operator to optimize the reduction process in terms of
maximum amount of mineral extracted and minimum energy consumption, a
control system integrated by several modules is being designed. One of
the modules has the task of predicting how much petroleum will be
burned in the ovens where the raw material is processed. This paper
proposes an algorithm to design Radial Basis Function Neural Networks
that will be able to predict accurately the amount of petroleum given
a set of input parameters. The algorithm is also able of identifying
the most relevant input parameters for the network so the
dimensionality reduction problem is ameliorated. Hence, this paper, as
it will be shown in the experiments section is able to apply the
synergy cif different Soft Computing techniques to the industrial
process obtaining satisfactory results. (C) 2009 Elsevier Ltd. All
rights reserved.

2010



* 100(<-637): Multi-criteria decision-making for optimization of product disassembly under multiple situations

With growing interest in recovering materials and subassemblies within consumer products at the end of their useful life, there has been an increasing interest in developing decision-making methodologies that determine how to maximize the environmental benefits of end-of-life (EOL) processing while minimizing costs under variable EOL situations. This paper describes a methodology to analyze how product designs and situational variables impact the Pareto set of optimal EOL strategies with the greatest environmental benefit for a given economic cost or profit. Since the determination of this Pareto set via enumeration of all disassembly sequences and EOL fates is prohibitively time-consuming even for relatively simple products, multi-objective genetic algorithms (GA) are utilized to rapidly approximate the Pareto set of optimal EOL trade-offs between cost and environmentally conscious actions. Such rapid calculations of the Pareto set are critical to better understand the influence of situational variables on how disassembly and recycling decisions change under different EOL-scenarios (e.g., under variable regulatory, infrastructure, or market situations). To illustrate the methodology, a case study involving the EOL treatment of a coffee maker is described. Impacts of situational variables on trade-offs between recovered energy and cost in Aachen, Germany, and in Ann Arbor, MI, are elucidated, and a means of presenting the results in the form of a multi-situational EOL strategy graph is described. The impact of the European Union Directive regarding Waste Electric and Electronic Equipment (WEEE) on EOL trade-offs between energy recovery and cost was also considered for both locations.

2003





* 101(<-207): Zone Refining of Tin: Optimization of Zone Length by a Genetic Algorithm

Zone refining comprises a number of techniques utilized to deal with
the rearrangement of soluble impurities or phases along a bar in order
to produce high-purity materials. The concentration curves can be
predicted for given values of segregation partition coefficient (k),
molten zone length, and a number of sequential zone passes. The
combination of such process parameters can result in many possible
experimental conditions, and the optimization by trial-and-error
methods is not suitable, even by numerical simulation due to
computational time consumption. The purpose of this work is to evolve
an interaction between a genetic algorithm (GA) and a predictive model
for impurity distribution, permitting the best zone length in each
pass to be determined in order to provide maximum purification,
minimum bar length waste and the lowest number of zone passes. The
proposed approach is validated against experimental results of zone
refining of tin, for impurities having opposite segregation behaviour,
i.e., k>1 and k<1.

2013



* 102(<-283): Optimization of End Milling Parameters under Minimum Quantity Lubrication Using Principal Component Analysis and Grey Relational Analysis

Machining is the major reliable practice in accomplishment of metal
cutting industries. The accelerated growing competition demands top
superior and large quantity with low cost products. Metal working
fluids have significant fragment of manufacturing cost and causes
ecological impacts and health problems. This work attempts to advance
a competent machining alignment with no ecological impacts. The
prediction of quality characteristics and enhancement of machining
field are consistently accepting great interest in machining sectors
to compress the accomplishment costs. In this paper, GA based ANN
prediction model proposes to envisage the quality characteristics of
surface roughness and tool wear. The comparison of predicted and
experimental values acknowledges the precision of the model. The end
milling experiments are conducted beneath minimum quantity
lubrication. This paper as well deals with the multiple objective
optimization with principal component analysis, grey relational
analysis and Taguchi method. ANOVA was carried out to determine each
parameter contribution percentage on quality characteristics. The
results show that cutting speed is the most influencing parameter
followed by feed velocity, lubricant flow rate and depth of cut. The
confirmation tests acknowledge that the proposed multiple-objective
methodology is able in determining optimum machining parameters for
minimum surface roughness and tool wear.

2012





* 103(<-294): Optimum design of run-flat tire insert rubber by genetic algorithm

A generalized multi-objective optimization method making use of
genetic algorithm (GA) is introduced, in order to simultaneously
improve the riding comfort and the durability of run-flat tire by
optimally tailoring the shape and stiffness of the sidewall insert
rubber. The sensitivity analysis invoking the CPU time-consuming
finite element analyses is replaced with the genetic evolution and the
fitness of each genome in the population is evaluated by utilizing the
response surfaces of objective functions approximated by ANN. It is
confirmed through the numerical experiment that a number of Pareto
solutions of the shape and stiffness of the sidewall insert rubber for
different combinations of weighting factors can be successfully
obtained. As well, the reliability of the Pareto solutions has been
justified from the comparison with the direct finite element
analysis. (C) 2011 Elsevier B.V. All rights reserved.

2012



* 108(<-382): Realization of Non Linear Controllers in Batch Reactor using GA and SVM

This paper presents the application of machine learning schemes,
namely SVM and GA, for realization of non linear control schemes and
optimization of Batch reactor. Batch reactor is an essential unit
operation in almost all batch-processing industries such as chemical
and pharmaceuticals. In this approach, the temperature profile of the
batch reactor is optimized using Genetic Algorithm (GA) with a view to
maximize the desired product and minimize the waste product as a
multi-objective function. Generic Model Control is implemented by
using SVM Estimator, and it includes the non-linear model of a process
to determine the control action. SVM estimator will predict the
current value of the heat release makes the control performance to be
more robust. The robustness performance of GMC has been
experienced. Other non linear control schemes, such as Direct Inverse
Control and Internal Model Control, are also implemented.

2011




* 110(<- 70): Multi-objective efficiency enhancement using workload spreading in an operational data center

The cooling systems of rapidly growing Data Centers (DCs) consume a
considerable amount of energy, which is one of the main concems in
designing and operating DCs. The main source of thermal inefficiency
in a typical air-cooled DC is hot air recirculation from outlets of
servers into their inlets, causing hot spots and leading to
performance reduction of the cooling system. In this study, a
thermally aware workload spreading method is proposed for reducing the
hot spots while the total allocated server workload is increased. The
core of this methodology lies in developing an appropriate thermal DC
model for the optimization process. Given the fact that utilizing a
high-fidelity thermal model of a DC is highly time consuming in the
optimization process, a three dimensional reduced order model of a
real DC is developed in this study. This model, whose boundary
conditions are determined based on measurement data of an operational
DC, is developed based on the potential flow theory updated with the
Rankine vortex to account for buoyancy and air recirculation effects
inside the DC. Before evaluating the proposed method, this model is
verified with a computational fluid dynamic (CFD) model simulated with
the same boundary conditions. The efficient load spreading method is
achieved by applying a multi-objective particle swarm optimization
(MOPSO) algorithm whose objectives are to minimize the hot spot
occurrences and to maximize the total workload allocated to
servers. In this case study, by applying the proposed method, the
Coefficient of Performance (COP) of the cooling system is increased by
17%, and the total allocated workload is increased by 10%. These
results demonstrate the effectiveness of the proposed method for
energy efficiency enhancement of DCs. Crown Copyright (C) 2014
Published by Elsevier Ltd. All rights reserved.

2015




* 111(<-426): Fuzzy TOPSIS approach for assessing thermal-energy storage in concentrated solar power (CSP) systems

The energy produced by thermal solar plants does not have to be
limited solely to hours of sunlight. It is possible to conceive a
storage system and it is possible to extend the production of heat
beyond the hours of full sunshine. The main aim of this paper is to
propose and test the validity and effectiveness of the proposed fuzzy
multi-criteria method (TOPSIS fuzzy) to compare different heat
transfer fluids (HTF) in order to investigate the feasibility of
utilizing a molten salt. The thermal processes involved in CSP will
not analyzed. The use of molten salt offers the potential to reduce
electricity production cost and to increase the energy performance in
an eco-compatible way. Salt is less expensive and more environmentally
benign than currently used HTFs but unfortunately the high freezing
point leads to significant O&M challenges and requires an innovative
freeze protection system. (C) 2009 Elsevier Ltd. All rights reserved.

2010




* 118(<-144): Multiobjective Optimization Design of Heating System in Electric Heating Rapid Thermal Cycling Mold for Yielding High Gloss Parts



2014






* 120(<-297): Multi-objective parametric optimization of powder mixed electro-discharge machining using response surface methodology and non-dominated sorting genetic algorithm

Powder mixed electro-discharge machining (EDM) is being widely used in
modern metal working industry for producing complex cavities in dies
and moulds which are otherwise difficult to create by conventional
machining route. It has been experimentally demonstrated that the
presence of suspended particle in dielectric fluid significantly
increases the surface finish and machining efficiency of EDM
process. Concentration of powder (silicon) in the dielectric fluid,
pulse on time, duty cycle, and peak current are taken as independent
variables on which the machining performance was analysed in terms of
material removal rate (MRR) and surface roughness (SR). Experiments
have been conducted on an EZNC fuzzy logic Die Sinking EDM machine
manufactured by Electronica Machine Tools Ltd. India. A copper
electrode having diameter of 25 mm is used to cut EN 31 steel for one
hour in each trial. Response surface methodology (RSM) is adopted to
study the effect of independent variables on responses and develop
predictive models. It is desired to obtain optimal parameter setting
that aims at decreasing surface roughness along with larger material
removal rate. Since the responses are conflicting in nature, it is
difficult to obtain a single combination of cutting parameters
satisfying both the objectives in any one solution. Therefore, it is
essential to explore the optimization landscape to generate the set of
dominant solutions. Non-sorted genetic algorithm (NSGA) has been
adopted to optimize the responses such that a set of mutually dominant
solutions are found over a wide range of machining parameters.

2012



* 121(<-312): Selection of EDM Process Parameters Using Biogeography-Based Optimization Algorithm

Amongst the nontraditional machining processes, electric discharge
machining (EDM) is considered to be one of the most important
processes for machining intricate and complex shapes in various
electrically conductive materials, including high-strength,
temperature-resistant (HSTR) alloys, especially in aeronautical and
automotive industries. For achieving the best performance of the EDM
process, it is imperative to carry out parametric design which
involves characterization of multiple process responses, such as
material removal rate, tool wear rate, surface finish and surface
integrity, heat affected zone, etc., with respect to different
machining parameters, like peak current, pulse-on time, duty factor,
gap voltage, and dielectric flushing pressure, followed by parametric
optimization of the process. This article focuses on the application
of the biogeography-based optimization (BBO) algorithm for single and
multiobjective optimization of the responses of two EDM processes. The
optimization performance of the BBO algorithm is compared with that of
other population-based algorithms, e. g., genetic algorithm (GA), ant
colony optimization (ACO), and artificial bee colony (ABC)
algorithm. It is observed that the BBO algorithm performs better than
the others with respect to the optimal process response values.

2012





* 126(<-293): Predicting torsional strength of RC beams by using Evolutionary Polynomial Regression

A new view for the analytical formulation of torsional ultimate
strength for reinforced concrete (RC) beams by experimental data is
explored by using a new hybrid regression method termed Evolutionary
Polynomial Regression (EPR). In the case of torsion in RC elements,
the poor assumptions in physical models often result into poor
agreement with experimental results. Nonetheless, existing models have
simple and compact mathematical expressions since they are used by
practitioners as building codes provisions. EPR combines the best
features of conventional numerical regression techniques with the
effectiveness of genetic programming for constructing symbolic
expressions of regression models. The EPR modeling paradigm allows to
figure out existing patterns in recorded data in terms of compact
mathematical expressions, according to the available physical
knowledge on the phenomenon (if any). The procedure output is
represented by different formulae to predict torsional strength of RC
beam. The multi-objective search paradigm used by EPR allows
developing a set of formulae showing different complexity of
mathematical expressions as resulting into different agreement with
experimental data. The efficiency of such approach is tested using
experimental data of 64 rectangular RC beams reported in technical
literature. The input parameters affecting the torsional strength were
selected as cross-sectional area of beams, cross-sectional area of
one-leg of closed stirrup, spacing of stirrups, area of longitudinal
reinforcement, yield strength of stirrup and longitudinal
reinforcement, concrete compressive strength. Those results are
finally compared with previous studies and existing building codes for
a complete comparison considering formulation complexity and
experimental data fitting. (C) 2011 Elsevier Ltd. All rights reserved.

2012




* 127(<-451): Combining support vector regression and cellular genetic algorithm for multi-objective optimization of coal-fired utility boilers

Support vector regression (SVR) was employed to establish mathematical
models for the NO, emissions and carbon burnout of a 300 MW coal-fired
utility boiler. Combined with the SVR models, the cellular genetic
algorithm for multi-objective optimization (MOCell) was used for
multi-objective optimization of the boiler combustion. Meanwhile, the
comparison between MOCell and the improved non-dominated sorting
genetic algorithm (NSGA-II) shows that MOCell has superior performance
to NSGA-II regarding the problem. The field experiments were carried
out to verify the accuracy of the results obtained by MOCell, the
results were in good agreement with the measurement data. The proposed
approach provides an effective tool for multi-objective optimization
of coal combustion performance, whose feasibility and validity are
experimental validated. A time period of less than 4 s was required
for a run of optimization under a PC system, which is suitable for the
online application. (c) 2009 Elsevier Ltd. All rights reserved.

2009





* 128(<- 35): Optimization of gear blank preforms based on a new R-GPLVM model utilizing GA-ELM

The determination of the key dimensions of gear blank preforms with
complicated geometries is a highly nonlinear optimization task. To
determine critical design dimensions, we propose a novel and efficient
dimensionality reduction (DR) model that adapts Gaussian process
regression (GPR) to construct a topological constraint between the
design latent variables (LVs) and the regression space. This procedure
is termed the regression-constrained Gaussian process latent variables
model (R-GPLVM), which overcomes GPLVM's drawback of ignoring the
regression constrains. To determine the appropriate sub-manifolds of
the high-dimensional sample space, we combine the maximum a posteriori
method with the scaled conjugate gradient (SCG) algorithm. This
procedure can estimate the coordinates of preform samples in the space
of LVs. Numerical experiments reveal that the R-GPLVM outperforms the
pure GPR in various dimensional spaces, when the proper
hyper-parameters and kernel functions are solved for. Results using an
extreme learning model (ELM) obtain a better prediction precision than
the back propagation method (BP), when the dimensions are reduced to
seven and a Gaussian kernel function is adopted. After the seven key
variables are screened out, the ELM model will be constructed with
realistic inputs and obtains improved prediction accuracy. However,
since the ELM has a problem with validity of the prediction, a genetic
algorithm (GA) is exploited to optimize the connection parameters
between each network layer to improve the reliability and
generalization. In terms of prediction accuracy for testing datasets,
GA has a better performance compared to the differential evolution
(DE) approach, which motivates the choice to use the genetic
algorithm-extreme learning model (GA-ELM). Moreover, GA-ELM is
employed to measure the aforementioned DR using engineering
criteria. In the end, to obtain the optimal geometry, a parallel
selection method of multi-objective optimization is proposed to obtain
the Pareto-optimal solution, while the maximum finisher forming force
(MFFF) and the maximum finisher die stress (MFDS) are both
minimized. Comparative analysis with other numerical models including
finite element model (FEM) simulation is conducted using the GA
optimized preform. Results show that the values of MFFF and MFDS
predicted by GA-ELM and R-GPLVM agree well with the experimental
results, which validates the feasibility of our proposed methods. (C)
2015 Elsevier B.V. All rights reserved.

2015




* 129(<-389): Optimization of Viscoelastic Systems Combining Robust Condensation and Metamodeling

The effective design of viscoelastic dampers as applied to real-world
complex engineering structures can be conveniently carried out by
using modern multiobjective numerical optimization techniques. The
large number of evaluations of the cost functions normally combined
with the typically high dimensions of finite element models of
industrial structures makes multiobjective optimization very costly,
sometimes unfeasible. Those difficulties motivate the study reported
in this paper, in which a strategy is proposed consisting in the use
of evolutionary algorithms specially adapted to multiobjective
optimization of viscoelastic systems, combined with robust
condensation and metamodeling. After the discussion of various
theoretical aspects, a numerical application is presented to
illustrate the use and demonstrate the effectiveness of the
methodology proposed for the optimal design of viscoelastic
constrained layers.

2010




* 130(<-430): Response surface methodology using Gaussian processes: Towards optimizing the trans-stilbene epoxidation over Co2+-NaX catalysts

Response surface methodology (RSM) relies on the design of experiments and empirical modelling techniques to find the optimum of a process when the underlying fundamental mechanism of the process is largely unknown. This paper proposes an iterative RSM framework, where Gaussian process (GP) regression models are applied for the approximation of the response surface. GP regression is flexible and capable of modelling complex functions, as opposed to the restrictive form of the polynomial models that are used in traditional RSM. As a result, GP models generally attain high accuracy of approximating the response surface, and thus provide great chance of identifying the optimum. In addition, GP is capable of providing both prediction mean and variance, the latter being a measure of the modelling uncertainty. Therefore, this uncertainty can be accounted for within the optimization problem, and thus the process optimal conditions are robust against the modelling uncertainty. The developed method is successfully applied to the optimization of trans-stilbene conversion in the epoxidation of trans-stilbene over cobalt ion-exchanged faujasite zeolites (Co2+-NaX) catalysts using molecular oxygen. (C) 2009 Elsevier B.V. All rights reserved.

2010



* 131(<-606): A hybrid numerical approach for multi-responses optimization of process parameters and catalyst compositions in CO2OCM process over CaO-MnO/CeO2 catalyst

A new hybrid numerical approach, using Weighted Sum of Squared
Objective Functions (WSSOF) algorithm, was developed for
multi-responses optimization of carbon dioxide oxidative coupling of
methane (CO2 OCM). The optimization was aimed to obtain optimal
process parameters and catalyst compositions with high catalytic
performances. The hybrid numerical approach combined the
single-response modeling and optimization using Response Surface
Methodology (RSM) and WSSOF technique of multi-responses
optimization. The hybrid algorithm resulted in Pareto-optimal
solutions and an additional criterion was proposed over the solutions
to obtain a final unique optimal solution. The simultaneous maximum
responses of C, selectivity and yield were obtained at the
corresponding optimal independent variables. The results of the
multi-response optimization could be used to facilitate in
recommending the suitable operating conditions and catalyst
compositions for the CO2 OCM process. (c) 2004 Elsevier B.V. All
rights reserved.

2005





* 135(<-616): Multiobjective generation dispatch through a neuro-fuzzy technique

The multiobjective generation dispatch in electric power system treats
economy and emission impact as competing objectives which requires
some form of conflict resolution to arrive at a solution. This paper
presents an integrated approach combining a fuzzy coordination method
and a radial basis function ANN along with a heuristic rule based
search algorithm to solve multiobjective generation dispatch
problem. The algorithm developed is simple to use and can effectively
obtain the well-coordinated optimal solution while allowing more
flexibility in operation. Adaptability of the performance indices
composed of fuel cost and emission level are measured by the
membership functions. Combining the adaptability indices a fuzzy
decision making (FDM) function is obtained and the two-objective
optimization is then solved by maximizing the FDM function. Then, a
radial basis function ANN is developed to reach a preliminary
schedule. Since, some practical constraints may be violated in the
preliminary schedule, a heuristic rule based search algorithm is
developed to reach a feasible best compromising generation schedule
which satisfies all practical constraints. The proposed neuro-fuzzy
technique has been applied to IEEE-14-bus and 30-bus test systems and
the results are presented to illustrate the performance and
applicability of the technique.

2004




* 136(<-183): Soft computing techniques in advancement of structural metals

Current trends in the progress of technology demand availability of
materials resources ahead of the advancing fronts of the application
areas. During the last couple of decades, significant progress has
been made in computational and experimental design of materials. Among
the potential computational techniques, soft computing stands in
distinction due to the inherent flexibility in capturing the
complexity of the problem in global scale. Since 1990s remarkable
success has been achieved in soft computing activities in different
facets of materials science and engineering. Extensive efforts have
been devoted in design of metals and alloys based on
composition-process-microstructure-property correlation. The present
review aims to address the contribution of soft computing in the field
of structural metals and alloys including processing and joining. The
critical issues concerning applicability of particular techniques in
specific materials problem have been particularly emphasised
encompassing the scope of integrating the gradual progress in
different techniques in hybrid and tandem framework to address greater
complexities in larger length and time scale. Attempt has also been
made to emphasise on the evolution of newer knowledge and materials
through soft computing activities. Finally, the potential of soft
computing techniques in futuristic design approaches has been
critically enumerated.

2013




* 140(<-384): Modeling and Optimal Design of Machining-Induced Residual Stresses in Aluminium Alloys Using a Fast Hierarchical Multiobjective Optimization Algorithm

The residual stresses induced during shaping and machining play an
important role in determining the integrity and durability of metal
components. An important issue of producing safety critical components
is to find the machining parameters that create compressive surface
stresses or to minimize tensile surface stresses. In this article, a
systematic data-driven fuzzy modeling methodology is proposed, which
allows constructing transparent fuzzy models considering both accuracy
and interpretability attributes of fuzzy systems. The new method
employs a hierarchical optimization structure to improve the modeling
efficiency, where two learning mechanisms cooperate together: the
Nondominated Sorting Genetic Algorithm II (NSGA-II) is used to improve
the model's structure, while the gradient descent method is used to
optimize the numerical parameters. This hybrid approach is then
successfully applied to the problem that concerns the prediction of
machining induced residual stresses in aerospace aluminium
alloys. Based on the developed reliable prediction models, NSGA-II is
further applied to the multiobjective optimal design of aluminium
alloys in a oreverse-engineeringo fashion. It is revealed that the
optimal machining regimes to minimize the residual stress and the
machining cost simultaneously can be successfully located.

2011



* 142(<-496): A systematic decision criterion for the elimination of useless overpasses

The Seoul Metropolitan Government (SMG) recently considered
eliminating some useless overpasses that had once played a significant
role in maintaining continuous traffic flow but soon lost their
original, positive function and became an environmental burden. SMG is
in the process of identifying such types of overpasses out of 22
installations. The aim of this study was to develop a definite
criterion that can be used to identify overpasses to be
eliminated. All of the 22 existing overpasses were investigated in
terms of functionality, structural stability and conflicts with other
sustainable policies of SMG. Functionality was interpreted based on
the traffic efficiency, environmental amenity and traffic safety. The
weights of these functionality sub-factors were derived from a
pair-wise comparison technique used in analytic hierarchy process
(AHIP) methodology. The remaining two aspects were not subdivided but
evaluated directly. Final judgments were made based an the three
aspects with the assistance of well-known classification methodologies
such as K-means and a support vector machine ISM). As a result, five
overpasses in Seoul were identified to be eliminable. (C) 2008
Elsevier Ltd. All rights reserved.

2008





* 147(<-402): A stochastic optimization approach for paper recycling reverse logistics network design under uncertainty

One of the most important objectives of a manufacturing firm is the
efficient design and operation of its supply chain to maximize
profit. Paper is an example of a valuable material that can be
recycled and recovered. Uncertainty is one of the characteristics of
the real world. The methods that cope with uncertainty help
researchers get realistic results. In this study, a two-stage
stochastic programing model is proposed to determine a long term
strategy including optimal facility locations and optimal flow amounts
for large scale reverse supply chain network design problem under
uncertainty. This network design problem includes optimal recycling
and collection center locations and optimal flow amounts between the
nodes in the multi-facility environment. Proposed model is suitable
for recycling/manufacturing type of systems in reverse supply
chain. All deterministic, stochastic models are mixed-integer
programing models and are solved by commercial software GAMS
21.6/CPLEX 9.0.

2010




* 152(<-155): Multi-Objective Genetic Algorithms and Genetic Programming Models for Minimizing Input Carbon Rates in a Blast Furnace Compared with a Conventional Analytic Approach

Data-driven models were constructed for the Productivity, CO2
emission, and Si content for an operational Blast furnace using
evolutionary approaches that involved two recent strategies based upon
bi-objective genetic Programming and neural nets evolving through
Genetic Algorithms. The models were utilized to compute the optimum
tradeoff between the level of CO2 emission and productivity at
different Si levels, using a Predator-Prey Genetic Algorithm, well
tested for computing the Pareto-optimality. The results were pitted
against some similar calculations performed with commercial softwares
and also compared with the results of thermodynamics-based analytical
models.

2014





* 158(<-394): A hybrid model using supporting vector machine and multi-objective genetic algorithm for processing parameters optimization in micro-EDM

In micro-electrical discharge machining (EDM), processing parameters
greatly affect processing efficiency and stability. However, the
complexity of micro-EDM makes it difficult to determine optimal
parameters for good processing performance. The important output
objectives are processing time (PT) and electrode wear (EW). Since
these parameters influence the output objectives in quite an opposite
way, it is not easy to find an optimized combination of these
processing parameters which make both PT and EW minimum. To solve this
problem, supporting vector machine is adopted to establish a micro-EDM
process model based on the orthogonal test. A new multi-objective
optimization genetic algorithm (GA) based on the idea of non-dominated
sorting is proposed to optimize the processing
parameters. Experimental results demonstrate that the proposed
multi-objective GA method is precise and effective in obtaining
Pareto-optimal solutions of parameter settings. The optimized
parameter combinations can greatly reduce PT while making EW
relatively small. Therefore, the proposed method is suitable for
parameter optimization of micro-EDM and can also enhance the
efficiency and stability of the process.

2010





* 159(<-487): A Study on Uncertainty-Complexity Tradeoffs for Dynamic Nonlinear Sensor Compensation

In this paper, we focus on the design of reduced-complexity sensor
compensation modules based on learning-from-examples techniques. A
multiobjective optimization design framework is proposed, where system
complexity and compensation uncertainty are considered as two
conflicting costs to be jointly minimized. In addition, suitable
statistical techniques are applied to cope with the variability in the
uncertainty estimation arising from the limited availability of data
at design time. Numerical simulations are provided on a set of
synthetic models to show the validity of the proposed methodology.

2009




* 171(<-469): Incorporating prior model into Gaussian processes regression for WEDM process modeling

Sufficient sampling is usually time-consuming and expensive but also
is indispensable for supporting high precise data-driven modeling of
wire-cut electrical discharge machining (WEDM) process. Considering
the natural way to describe the behavior of a WEDM process by IF-THEN
rules drawn from the field experts, engineering knowledge and
experimental work. in this paper, the fuzzy logic model is chosen as
prior knowledge to leverage the predictive performance. Focusing on
the fusion between rough fuzzy system and very scarce noisy samples, a
simple but effective re-sampling algorithm based on piecewise
relational transfer interpolation is presented and it is integrated
with Gaussian processes regression (GPR) for WEDM process
modeling. First, by using re-sampling algorithm encoded derivative
regularization, the prior model is translated into a pseudo training
dataset, and then the dataset is trained by the Gaussian processes. An
empirical study on two benchmark datasets intuitively demonstrates the
feasibility and effectiveness of this approach. Experiments on
high-speed WEDM (DK7725B) are conducted for validation of nonlinear
relationship between the design variables (i.e., workpiece thickness,
peak current, ontime and off-time) and the responses (i.e., material
removal rate and surface roughness). The experimental result shows
that combining very rough fuzzy prior model with training examples
still significantly improves the predictive performance of WEDM
process modeling. even with very limited training dataset. That is,
given the generalized prior model, the samples needed by GPR model
could be reduced greatly meanwhile keeping precise. (c) 2008 Elsevier
Ltd. All rights reserved.

2009





* 181(<-466): Parameter identification of a non-associative elastoplastic constitutive model using ANN and multi-objective optimization

This paper deals with the identification of material parameters using
a hybrid method of multi-objective optimization. This approach was
used in a previous work to identify the Hill'48 criterion under the
associative normality assumption and the Voce law hardening parameters
of the Stainless Steel AISI 304. In this work, we apply the proposed
method in order to identify the orthotropic criterion of Hill'48 under
the non-associative normality assumption. The two models are compared
and analysed using several experimental tests.

2009




* 213(<-167): Spatial modelling of site suitability assessment for hospitals using geographical information system-based multicriteria approach at Qazvin city, Iran

Due to the population growth and continuous migration of people from
rural areas to urban areas, it is important to identify the suitable
locations for future development in order to find suitable sites for
various kinds of facilities such as schools, hospital and fire
stations for new and existing urban areas. Site suitability modelling
is a complex process involving various kinds of objectives and
issues. Such a complex process includes spatial analysis, use of
several decision support tools such as high-spatial resolution
remotely sensed data, geographical information system (GIS) and multi
criteria analysis (MCA) such as analytical hierarchy process (AHP),
and in some cases, prediction techniques like cellular automata (CA)
or artificial neural networks (ANN). This paper presents a comparison
between the results of AHP and the ordinary least square (OLS)
evaluation model, based on various criteria, to select suitable sites
for new hospitals in Qazvin city, Iran. Based on the obtained results,
proximity to populated areas (0.3) and distance to air polluted areas
(0.23-0.26) were the two highest important criteria with high weight
value. The results show that these two techniques not only have
similarity in size (in m(2)) for each suitability class but they also
have similarity in spatial distribution of each class in the entire
study area. Based on calculations of both techniques, 1-2%, 25%,
40-43%, 16-20% and 14% of study areas are assigned as 'not suitable',
'less suitable', 'moderately suitable', 'suitable' and 'most suitable'
areas for construction of new hospitals. Results revealed that a 75%
similarity was found in the distribution of suitability classes in
Qazvin city using both techniques. Nineteen per cent (19%) of the
study area are assigned as 'suitable' and ` most suitable' by both
methods, so these areas can be considered as safe or secure areas for
clinical purposes. Moreover, almost all (99.8%) suitable areas are
located in district 3, because of its higher population, less numbers
of existing hospitals and large numbers of barren land plots of
acceptable size.

2014




* 217(<-449): Multi-objective turbomachinery optimization using a gradient-enhanced multi-layer perceptron

Response Surface models (RSMs) have found widespread use to reduce the
overall Computational cost of turbomachinery blading design
optimization. Recent developments have seen the successful use of
gradient information alongside sampled response values in building
accurate response surfaces. This paper describes the use of gradients
to enhance the performance of the RSM provided by a multi-layer
perceptron. Gradient information is included in the perceptron by
modifying the error function such that the perceptron is trained to
fit the gradients as well as the response values. As a consequence,
the back-propagation scheme that assists the training is also
changed. The paper formulates the gradient-enhanced multi-layer
perceptron using algebraic notation, with an emphasis on the ease of
use and efficiency of computer code implementation. To illustrate the
benefit of using gradient information, the enhanced neural network
model is used in a multi-objective transonic fan blade optimization
exercise of engineering relevance. Copyright (C) 2008 John Wiley &
Sons, Ltd.

2009




* 245(<-553): A genetic algorithms based multi-objective neural net applied to noisy blast furnace data

A genetic algorithms based multi-objective optimization technique was
utilized in the training process of a feed forward neural network,
using noisy data from an industrial iron blast furnace. The number of
nodes in the hidden layer, the architecture of the lower part of the
network, as well as the weights used in them were kept as variables,
and a Pareto front was effectively constructed by minimizing the
training error along with the network size. A predator-prey algorithm
efficiently performed the optimization task and several important
trends were observed. (C) 2005 Elsevier B.V. All rights reserved.

2007




* 259(<-252): Optimization and experimental test of a miniature permanent magnet structure for a microfluidic magnetic resonance chip

We propose a general global optimal algorithm to optimize the
miniature permanent magnet structure of a micro magnetic resonance
chip (mu NMR-chip). For this purpose, we analyze the sensitivity of
the permanent magnet structure to the design variables and determine
the optimization variables. After this, radial basis function neural
networks (RBFNNs) are constructed to model the objective functions,
and the nondominated sorting genetic algorithm II (NSGA II) is
improved by introducing a different weighting factor for each
objective function in calculating the crowding distance. Combining the
RBFNN with the improved NSGA II optimizes the miniature permanent
magnet structure. Through comparison, the optimization solutions are
proven effective. Finally, the optimized permanent magnet structure is
manufactured and tested experimentally. After optimization, the volume
of the permanent magnet block is reduced by 39%, and the permanent
magnet becomes easier to manufacture.

2013




* 280(<-448): Multi-objective simultaneous prediction of waterborne coating properties

Multi-objective simultaneous prediction of waterborne coating
properties was studied by the neural network combined with
programming. The conditions of network with one input layer, three
hidden layers and one output layer were confirmed. The monomers mass
of BA, MMA, St and pigments mass of TiO(2) and CaCO(3) were used as
input data. Four properties, which were hardness, adhesion, impact
resistance and reflectivity, were used as network output. After
discussing the hidden layer neurons, learn rate and the number of
hidden layers, the best net parameters were confirmed. The results of
experiment show that multi-hidden layers was advantageous to improve
the accuracy of multi-objective simultaneous prediction. 36 kinds of
coating formulations were used as the training subset and 9 acrylate
waterborne coatings were used as testing subset in order to predict
the performance. The forecast error of hardness was 8.02% and
reflectivity was 0.16%. Both forecast accuracy of adhesion and impact
resistance were 100%.

2009




* 285(<-341): Neuro-simulation modeling of chemical flooding

Chemical flooding has proved to enhance oil recovery of reservoirs
considerably. Development strategies of this method are more efficient
when they consider both aspects of operation (recovery factor, RF) and
economics (net present value, NPV). In this study, a multi-layer
perceptron (MLP) neural network is developed for modeling of chemical
flooding using surfactant and polymer via prediction of both RF and
NPV in a unique model. The modeling algorithm is divided into three
processes: training, generalization, and operation. In training
process, the initial structure of the network is trained, and then the
architecture of the trained network is optimized for reduction of
prediction errors in generalization process. Furthermore, the optimum
structure is compared with other methods like Radial Basis Function
(RBF) neural network, quadratic and multi-objective regressions. The
optimum architecture of the network contains one hidden layer with 8
neurons and training function of Bayesian regularization. In operation
process, sensitivity analysis is studied for evaluating of effective
parameters (inputs) on the performance of chemical flooding. The error
is always less than 5% during the implementation of all processes. The
results demonstrate that neuro-simulation of chemical flooding is
reliable, inexpensive, fast in computational effort, and capable in
accurate prediction of both RF and NPV in one model. (C) 2011 Elsevier
B.V. All rights reserved.

2011




* 292(<-641): Nonlinear identification of aircraft gas-turbine dynamics

Identification results for the shaft-speed dynamics of an aircraft gas
turbine, under normal operation, are presented. As it has been found
that the dynamics vary with the operating point, nonlinear models are
employed. Two different approaches are considered: NARX models, and
neural network models, namely multilayer perceptrons, radial basis
function networks and B-spline networks. A special attention is given
to genetic programming, in a multiobjective fashion, to determine the
structure of NARMAX and B-spline models. (C) 2003 Elsevier B.V. All
rights reserved.

2003



* 294(<-340): Nonlinear Modeling Method Applied to Prediction of Hot Metal Silicon in the Ironmaking Blast Furnace

Feedforward neural networks have been established as versatile tools for nonlinear black-box modeling, but in many data-mining tasks the choice of relevant inputs and network complexity still constitute major challenges. Statistical tests for detecting relations between inputs and outputs proposed in the literature are largely based on the theory for linear systems, and laborious retraining combined with the risk of getting stuck in local minima make the application of exhaustive search through all possible network configurations impossible but for toy problems. This paper proposes a systematic method to tackle the problem where an output shall be estimated on the basis of a (large) set of potential inputs. Feedforward neural networks of multilayer perceptron type are used in the three-stage approach: First, starting from sufficiently large networks, an efficient pruning method is applied to detect potential model candidates. Next, the best results of the pruning runs are extracted by forming a Pareto-frontier, with the contradictory objectives of minimizing network complexity and estimation error. The networks on this frontier are considered to contain promising hidden nodes with their specific connections to relevant input variables. These hidden nodes are therefore optimally combined by mixed-integer linear programming to form a final set of neural network models, from which the user can select a model of suitable complexity. The modeling method is applied on an illustrative test example as well as on a complex modeling problem in the metallurgical industry, i.e., prediction of the silicon content of hot metal produced in a blast furnace. It is demonstrated to find relevant inputs and to yield parsimonious sparsely connected neural models of the output.

2011




* 307(<-417): Multiobjective scheduling for semiconductor manufacturing plants

Scheduling of semiconductor wafer manufacturing system is identified
as a complex problem, involving multiple and conflicting objectives
(minimization of facility average utilization, minimization of waiting
time and storage, for instance) to simultaneously satisfy. In this
study, we propose an efficient approach based on an artificial neural
network technique embedded into a multiobjective genetic algorithm for
multi-decision scheduling problems in a semiconductor wafer
fabrication environment. (C) 2010 Elsevier Ltd All rights reserved

2010




* 308(<-447): A Neural Network and a Genetic Algorithm for Multiobjective Scheduling of Semiconductor Manufacturing Plants

Scheduling of semiconductor wafer fabrication system is identified as a complex problem, involving multiple objectives to be satisfied simultaneously (maximization of workstation utilization and minimization of waiting time and storage, for instance). In this study, we propose a methodology based oil an artificial neural network technique, for computing the various objective functions, embedded into a multiobjective genetic algorithm for multidecision scheduling problems in a semiconductor wafer fabrication environment. A discrete event simulator, developed and validated in our previous works, serves here to feed the neural network. Six criteria related to both equipment (facility average utilization) and products (average cycle time (ACT), standard deviation of ACT, average waiting time, work in process, and total storage) are chosen as significant performance indexes of the workshop. The optimization variables are the time between campaigns and the release time of batches into the plant. An industrial size example is taken as a test bench to validate the approach.

2009


* 309(<-679): Genetic neuro-scheduler: A new approach for job shop scheduling

In this paper, a hybrid approach between two new techniques, genetic
algorithms and artificial neural network is described for generating
job shop schedules in a discrete manufacturing environment based on
nonlinear multiobjective function. Genetic algorithm (GA) is used as
an effective search technique for finding an optimal schedule via
population of gene strings which represent alternative feasible
schedules. GA propagates new population of genes through number of
cycles called generations by implementing natural genetic
mechanism. Specifically, gene strings should have a structure that
imposes the most common restrictive constraint; a precedence
constraint. The other technique is an artificial neural network that
performs multiobjective schedule evaluation. The intention is to
establish an effective model that maps a complex set of scheduling
criteria (i.e. flowtime, lateness) to appropriate values provided by
experienced schedulers. The proposed approach is prototyped and tested
on four different job shop scheduling problems based on problem size,
namely; small, medium, large, and a sample problem provided by a
company. The comparative results indicate that the proposed approach
is consistently better than those of heuristic algorithms used
extensively in industry.

1995


* 310(<-686): GENETIC NEURO-SCHEDULER FOR JOB-SHOP SCHEDULING

This paper describes a hybrid approach between two new techniques, Genetic Algorithms and Artificial Neural Networks, for generating Job Shop Schedules (JSS) in a discrete manufacturing environment based on non-linear multi-criteria objective function. Genetic Algorithm (GA) is used as a search technique for an optimal schedule via a uniform randomly generated population of gene strings which represent alternative feasible schedules. GA propagates this specific gene population through a number of cycles or generations by implementing natural genetic mechanism ( i.e. reproduction operator and crossover operator). It is important to design an appropriate format of genes for JSS problems. Specifically, gene strings should have a structure that imposes the most common restrictive constraint; a precedence constraint. The other is an Artificial Neural Network, which uses its highly connected-neuron network to perform as a multi-criteria evaluator. The basic idea is a neural network evaluator which maps a complex set of scheduling criteria (i.e. flowtime, lateness) to evaluate values provided by experienced experts. Once, the network is fully trained, it will be used as an evaluator to access the fitness or performance of those simulated gene strings. The proposed approach was prototyped and implemented on JSS problems based on different model sizes; namely small, medium, and large. The results are compared to the Shortest Processing Time heuristic used extensively in industry.

1993




* 311(<-617): Process optimisation of transfer moulding for electronic packages using artificial neural networks and multiobjective optimisation techniques

Transfer moulding is the most common process for the encapsulation of
electronic packages in semiconductor manufacturing. Quality of the
moulding is affected by a large number of mould design parameters and
process parameters. Currently, the parameters setting is performed by
experienced engineers in a trial and error manner and often the
optimal setting can not be obtained. In the face of global
competition, the current practice is inadequate. In this research, a
process optimisation system for transfer moulding of electronic
packages is described which involves design of experiments (DOE)
techniques, artificial neural networks (ANNs), multiple regression
analysis and the minimax method. The system is aimed to determine the
optimal mould design parameters and process parameter settings of
transfer moulding of electronic packages for multiobjective
problem. Implementation of the optimisation system has demonstrated
that the time for the determination of optimal mould design parameters
and process parameters setting can be greatly reduced and the
parameters setting recommended by the system can contribute to the
good quality of moulded packages without relying on experienced
engineers.

2004


* 312(<-624): Intelligent process design system for the transfer moulding of electronic packages

Currently, mould design and the setting of the process parameters of transfer moulding for electronic packages are done manually in a trial-and-error manner. The effectiveness of the setting of parameters is largely dependent on the experience of engineers. The paper describes an intelligent process design system for transfer moulding of electronic packages that is used to determine optimal mould design parameters and the setting of the process parameters mainly based on case-based reasoning, artificial neural networks and a multiobjective optimization scheme. The system consists of two modules: a case-based reasoning module and a process optimization module. The former module is used to determine initial mould design parameters and the setting of the process parameters while the latter module is used to determine optimal mould design parameters and the setting of the process parameters. Implementation of the intelligent system has demonstrated that the time for the determination of optimal mould design parameters and the setting of the process parameters can be greatly reduced, and the setting of parameters recommended by the system can contribute to the good quality of moulded packages.

2004





* 314(<-480): Noise Reduction in a Non-Homogenous Ground Penetrating Radar Problem by Multiobjective Neural Networks

This paper applies artificial neural networks (ANNs) trained with a
multiobjective algorithm to preprocess the ground penetrating radar
data obtained from a finite-difference time-domain (FDTD) model. This
preprocessing aims at improving the target's reflected wave
signal-to-noise ratio (SNR). Once trained, the NN behaves as an
adaptive filter which minimizes the cross-validation error. Results
considering both white and colored Gaussian noise, with many different
SNR, are presented and they show the effectiveness of the proposed
approach.

2009




* 318(<-386): Artificial neural network-based resistance spot welding quality assessment system

On-line quality assessment has become one of the most critical
requirements for improving the efficiency and the autonomy of
automatic resistance spot welding (RSW) processes. An accurate and
efficient model to perform non-destructive quality estimation is an
essential part of the assessment process. This paper presents a
structured and systematic approach developed to design an effective
ANN-based model for on-line quality assessment in RSW. The proposed
approach examines welding parameters and conditions known to have an
influence on weld quality, and builds a quality estimation model step
by step. The modeling procedure begins by examining, through a
structured experimental design, the effect of welding parameters
(welding time, welding current, electrode force and sheet metal
thickness) and welding conditions represented by typical
characteristics of the dynamic resistance curves on multiple welding
quality indicators (indentation depth, nugget diameter and nugget
penetration) and by analyzing their interactions and their sensitivity
to the variation of the dynamic process conditions. Using these
results and by combining an efficient modeling planning method, neural
network paradigm, multi-criteria optimization and various statistical
tools, the identification of the model form and the variables to be
included in the model is achieved by executing a systematic model
optimization procedure. The results demonstrate that the proposed
approach can lead to a general ANN-based model able to accurately and
reliably provide an appropriate assessment of the weld quality under
diverse and variable welding conditions.

2011




* 320(<-409): Artificial neural networks for machining processes surface roughness modeling

In recent years, several papers on machining processes have focused on the use of artificial neural networks for modeling surface roughness. Even in such a specific niche of engineering literature, the papers differ considerably in terms of how they define network architectures and validate results, as well as in their training algorithms, error measures, and the like. Furthermore, a perusal of the individual papers leaves a researcher without a clear, sweeping view of what the field's cutting edge is. Hence, this work reviews a number of these papers, providing a summary and analysis of the findings. Based on recommendations made by scholars of neurocomputing and statistics, the review includes a set of comparison criteria as well as assesses how the research findings were validated. This work also identifies trends in the literature and highlights their main differences. Ultimately, this work points to underexplored issues for future research and shows ways to improve how the results are validated.

2010




* 325(<-529): Springback Compensation of Sheet Metal Bending Process Based on DOE & ANN

Nowadays, the trend to a lightweight design accelerates the use of advanced high strength steel (AHSS) in automotive industry. Springback phenomena is a hot issue in the sheet metal forming, especially bending process using AHSS. Several analytical methods for that have been proposed in recent years. Each of method has their advantages and disadvantages. There are only a few optimal solutions which can minimize the two objectives simultaneously. In this study, an effective method optimized the multi objective value. The method by the design of experiments(DOE) and artificial neural network(ANN) was presented to compensate springback of bending parts. This method was applied to L and V bending process. The effective method could be optimized to multiple object. It was confirmed that the proposed method was more efficient than traditional manual FEA procedure and the trial and error approach for springback compensation.\

2008



* 327(<-323): Modelling, optimization and decision making techniques in designing of functional clothing

Functional clothing are actually engineered textiles as they require
to meet the stringent performance characteristics rather than the
aesthetic properties. Therefore, the trial and error approach of
product design does not seem to be a viable way for functional
clothing. It needs more potent approaches of modelling, optimization
and decision making so that the design and functional requirements of
clothing can be met with acceptable tolerance. This paper provides a
brief outline of various techniques of modelling, optimization and
decision making intended for designing of functional clothing. In the
modelling part, regression and artificial neural network approaches
have been discussed with the examples of thermal property and water
repellency modelling. Subsequently, linear programming and genetic
algorithm techniques have been invoked in the optimization
part. Optimization of ultraviolet radiation protective clothing is
taken up as a case study. Finally, multi-criteria decision making
techniques have been explained with the hypothetical example of
selection of best body armour vest for defense applications.

2011



* 328(<-489): On the modeling of car passenger ferryship design parameters with respect to selected sea-keeping qualities and additional resistance in waves

This paper presents the modeling of car passenger ferryship design parameters with respect to such design criteria as selected sea-keeping qualities and additional resistance in waves. In the first part of the investigations approximations of selected statistical parameters of design criteria of ferryship were elaborated with respect to ship design parameters. The approximation functions were obtained with the use of artificial neural networks. In the second part of the investigations design solutions were searched for by applying the single- and multi-criterial optimization methods. The multi-criterial optimization was performed by using Pareto method. Such approach made it possible to present solutions in such form as to allow decision makers (shipowner, designer) to select solutions the most favourable in each individual case.

2009



* 330(<-281): Learning and training techniques in fuzzy control for energy efficiency in buildings

A novel procedure for learning Fuzzy Controllers (FC) is proposed that concerns with energy efficiency issues in distributing electrical energy to heaters in an electrical energy heating system. Energy rationalization together with temperature control can significantly improve energy efficiency, by efficiently controlling electrical heating systems and electrical energy consumption. The novel procedure, which improves the training process, is designed to train the FC, as well as to run the control algorithm and to carry out energy distribution. Firstly, the dynamic thermal performance of different variables is mathematically modelled for each specific building type and climate zone. Secondly, an exploratory projection pursuit method is used to extract the relevant features. Finally, a supervised dynamic neural network model and identification techniques are applied to FC learning and training. The FC rule-set and parameter-set learning process is a multi-objective problem that minimizes both the indoor temperature error and the energy deficit in the house. The reliability of the proposed procedure is validated for a city in a winter zone in Spain.

2012



* 331(<-359): A New and Efficient Intelligent Collaboration Scheme for Fashion Design

Technology-mediated collaboration process has been extensively studied for over a decade. Most applications with collaboration concepts reported in the literature focus on enhancing efficiency and effectiveness of the decision-making processes in objective and well-structured workflows. However, relatively few previous studies have investigated the applications of collaboration schemes to problems with subjective and unstructured nature. In this paper, we explore a new intelligent collaboration scheme for fashion design which, by nature, relies heavily on human judgment and creativity. Techniques such as multicriteria decision making, fuzzy logic, and artificial neural network (ANN) models are employed. Industrial data sets are used for the analysis. Our experimental results suggest that the proposed scheme exhibits significant improvement over the traditional method in terms of the time-cost effectiveness, and a company interview with design professionals has confirmed its effectiveness and significance.

2011





* 333(<- 46): Implementation and testing of a soft computing based model predictive control on an industrial controller

This work presents a real time testing approach of an Intelligent Multiobjective Nonlinear-Model Predictive Control Strategy (iMO-NMPC). The goal is the testing and analysis of the feasibility and reliability of some Soft Computing (SC) techniques running on a real time industrial controller. In this predictive control strategy, a Multiobjective Genetic Algorithm is used together with a Recurrent Artificial Neural Network in order to obtain the control action at each sampling time. The entire development process, from the numeric simulation of the control scheme to its implementation and testing on a PC-based industrial controller, is also presented in this paper. The computational time requirements are discussed as well. The obtained results show that the SC techniques can be considered also to tackle highly nonlinear and coupled complex control problems in real time, thus optimising and enhancing the response of the control loop. Therefore this work is a contribution to spread the SC techniques in on-line control applications, where currently they are relegated mainly to be used off-line, as is the case of optimal tuning of control strategies. (C) 2014 Elsevier B.V. All rights reserved.

2015



* 334(<-139): Framework for Creating Digital Representations of Structural Components Using Computational Intelligence Techniques

A framework for creating a digital representation of physical
structural components is investigated. A model updating scheme used
with an artificial neural network to map updating parameters to the
error observed between simulated experimental data and an analytical
model of a turbine-engine fan blade. The simulated experimental
airfoil has as-manufactured geometric deviations from the nominal,
design-intent geometry on which the analytical model is based. The
manufacturing geometric deviations are reduced through principal
component analysis, where the scores of the principal components are
the unknown updating parameters. A range of acceptable scores is used
to devise a design of computer experiments that provides training and
testing data for the neural network. This training data is composed of
principal component scores as inputs. The outputs are the calculated
errors between the analytical and experimental predictions of modal
properties and frequency-response functions. Minimizing these errors
will result in an updated analytical model that has predictions closer
to the simulated experimental data. This minimization process is done
through the use of two multiobjective evolutionary algorithms. The
goal is to determine if the updating process can identify the
principal components used in simulating the experiment data.

2014





* 341(<-302): An integration methodology based on fuzzy inference systems and neural approaches for multi-stage supply-chains

This paper proposes a methodology for supply chain (SC) integration
from customers to suppliers through warehouses, retailers, and plants
via both adaptive network based fuzzy inference system and artificial
neural networks approaches. The methodology presented provides this
integration by finding the requested supplier capacities using the
demand and order lead time information across the whole SC in an
uncertain environment. The SC structure is investigated stage by
stage. The sensitivity analysis is made by comparing the obtained
results with the traditional statistical techniques. A company serving
in durable consumer goods industry that produces consumer electronics
in Istanbul, Turkey was examined to demonstrate the applicability of
the proposed methodology. (C) 2011 Elsevier Ltd. All rights reserved.

2012



* 342(<-463): Entropy-based optimal sensor networks for structural health monitoring of a cable-stayed bridge

The sudden collapse of Interstate 35 Bridge in Minneapolis gave a wake-up call to US municipalities to re-evaluate aging bridges. In this situation, structural health monitoring (SUM) technology can provide the essential help needed for monitoring and maintaining the nation's infrastructure. Monitoring long span bridges such as cable-stayed bridges effectively requires the use of a large number of sensors. In this article, we introduce a probabilistic approach to identify optimal locations of sensors to enhance damage detection. Probability distribution functions are established using an artificial neural network traced using a priori knowledge of damage locations. The optimal number of sensors is identified using multi-objective optimization that simultaneously considers information entropy and sensor cost-objective functions. Luling Bridge, a cable-stayed bridge over the Mississippi River, is selected as a case study to demonstrate the efficiency of the proposed approach.

2009





* 348(<-607): An evolutionary artificial neural networks approach for BF hot metal silicon content prediction

This paper presents an evolutionary artificial neural network (EANN)
to the prediction of the BF hot metal silicon content. The pareto
differential evolution (PDE) algorithm is used to optimize the
connection weights and the network's architecture (number of hidden
nodes) simultaneously to improve the prediction precision. The
application results show that the prediction of hot metal silicon
content is successful. Data, used in this paper, were collected from
No. I BF at Laiwu Iron and Steel Group Co..

2005




* 351(<-559): Genetic algorithms in optimization of strength and ductility of low-carbon steels

A comparative study between the conventional goal attainment strategy
and an evolutionary approach using a genetic algorithm has been
conducted for the multiobjective optimization of the strength and
ductility of low-carbon ferrite-pearlite steels. The optimization is
based upon the composition and microstructural relations of the
mechanical properties suggested earlier through regression
analyses. After finding that a genetic algorithm is more suitable for
such a problem, Pareto fronts have been developed which give a range
of strength and ductility useful in alloy design. An effort has been
made to optimize the strength ductility balance of
thermomechanically-processed high-strength multiphase steels. The
objective functions are developed from empirical relations using
regression and neural network modeling, which have the capacity to
correlate high number of compositional and process variables, and
works better than the conventional regression analyses.

2007




* 355(<-535): Multi-objective evolutionary optimization of subsonic airfoils by meta-modelling and evolution control

The current work concerns the application of multi-objective evolutionary optimization by approximation function to aerodynamic design. A new general technique, named evolution control (EC), is used in order to manage the on-line enriching of correct solutions database, which is the basis of the learning procedure for the approximators. Substantially, this approach provides that the database, initially quite small and enabling a very inaccurate approximation, should be integrated during the optimization. Such integration is done by means of some choice criteria, allowing deciding which individuals of the current population should be verified. The technique showed being efficacious and very efficient for the considered problem, whose dimensionality are 5. Even if general principle of EC is valid independently from the kind of adopted approximator, this last strongly affects the application. Obtained results are utilized to show how the adoption of artificial neural networks and kriging can differently influence the whole optimization process. Moreover, first results, achieved after reformulating the same problem with seven parameters, support the idea of the performance of the method scale well with dimensionality.

2007




* 359(<-552): Evolutionary artificial neural network for selecting flexible manufacturing systems under disparate level-of-satisfaction of decision maker

This paper proposes the application of Meta-Learning Evolutionary
Artificial Neural Network (MLEANN) in selecting the best flexible
manufacturing systems (FMS) from a group of candidate
FMSs. Multi-criteria decision-making (MCDM) methodology using an
improved S-shaped membership function has been developed for finding
out the "best candidate FMS alternative" from a set of
candidate-FMSS. The MCDM model trade-offs among various parameters,
viz., design parameters, economic considerations, etc., affecting the
FMS selection process under multiple, conflicting-in-nature criteria
environment. The selection of FMS is made according to the error
output of the results found from the proposed MCDM model.

2007



* 360(<-584): Meta-learning evolutionary artificial neural network for selecting flexible manufacturing systems

This paper proposes the application of Meta-Learning Evolutionary
Artificial Neural Network (MLEANN) in selecting flexible manufacturing
systems (FMS) from a group of candidate FMS's. First, multi-criteria
decision-making (MCDM) methodology using an improved S-shaped
membership function has been developed for finding out the 'best
candidate FMS alternative' from a set of candidate-FMSs. The MCDM
model trade-offs among various parameters, namely, design parameters,
economic considerations, etc., affecting the FMS selection process in
multi-criteria decision-making environment. Genetic algorithm is used
to evolve the architecture and weights of the proposed neural network
method. Further, a back-propagation (BP) algorithm is used as the
local search algorithm. The selection of FMS is made according to the
error output of the results found from the MCDM model.

2006






* 361(<- 95): Fuzzy reliability analysis of repairable industrial systems using soft-computing based hybridized techniques

The purpose of the present study is to analyze the fuzzy reliability
of a repairable industrial system utilizing historical vague,
imprecise and uncertain data which reflects its components' failure
and repair pattern. Soft-computing based two different hybridized
techniques named as Genetic Algorithms Based Lambda-Tau (GABLT) and
Neural Network and Genetic Algorithms Based Lambda-Tau (NGABLT) along
with a traditional Fuzzy Lambda-Tau (FLT) technique are used to
evaluate some important reliability indices of the system in the form
of fuzzy membership functions. As a case study, all the three
techniques are applied to analyse the fuzzy reliability of the washing
system in a paper mill and results are compared. Sensitivity analysis
has also been performed to analyze the effect of variation of
different reliability parameters on system performance. The analysis
can help maintenance personnel to understand and plan suitable
maintenance strategy to improve the overall performance of the
system. Based on results some important suggestions are given for
future course of action in maintenance planning. (C) 2014 Elsevier
B.V. All rights reserved.

2014



* 362(<- 71): Multi-criteria Optimization of Hole Geometry for the Laser Trepanning of the Titanium Alloy Ti-6Al-4V

Titanium and its alloys are extensively used materials in aerospace
industry due to their remarkable metallurgical and mechanical
characteristics; however, superior mechanical properties, poor thermal
conductivity and higher chemical affinity of titanium alloys make them
difficult-to-cut by conventional machining methods. The present
research investigates the possibility of machining quality improvement
in Ti-6Al-4V by applying the laser trepan drilling. The drilled hole
quality in terms of hole taper, and circularity at top and bottom
sides have been considered for multi-criteria optimization. Authors
have applied a new hybrid approach of modelling and multi-objective
optimization of hole geometry in laser trepan drilling of difficult to
cut Ti-alloy. The hole quality comprising taper and circularities are
represented by a common performance index in mathematical form using
artificial neural network modelling coupled with grey-entropy
measurement technique. Optimization of the performance index function
using genetic algorithms show considerable improvements in hole
quality. The parametric effect show that high chemical reactivity and
low thermal conductivity of Ti-alloy play important role in
deteriorating the hole geometry.

2015



* 363(<- 98): Modeling and optimization of turning duplex stainless steels

The attractive combination of high mechanical strength, good corrosion resistance and relatively low cost has contributed to making duplex stainless steels (DSSs) one of the fastest growing groups of stainless steels. As the importance of DSSs is increasing, practical information about their successful machining is expected to be crucial. To address this industrial need, standard EN 1.4462 and super EN 1.4410 DSSs are machined under constant cutting speed multi-pass facing operations. A systematic approach which employs different modeling and optimization tools under a three phase investigation scheme has been adopted. In phase I, the effect of design variables such as cutting parameters, cutting fluids and axial length of cuts are investigated using the D-Optimal method. The mathematical models for performance characteristics such as; percentage increase in radial cutting force (%F-r), effective cutting power (P-e), maximum tool flank wear (VBmax) and chip volume ratio (R) are developed using response surface methodology (RSM). The adequacy of derived models for each cutting scenario is checked using analysis of variance (ANOVA). Parametric meta-heuristic optimization using Cuckoo search (CS) algorithm is then performed to determine the optimum design variable set for each performance. In the phase II, comprehensive experiment-based production cost and production rate models are developed. To overcome the conflict between the desire of minimizing the production cost and maximizing the production rate, compromise solutions are suggested using Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). The alternatives are ranked according to their relative closeness to the ideal solution. In the phase III, expert systems based on fuzzy rule modeling approach are adopted to derive measures of machining operational sustainability called operational sustainability index (OSI). Artificial neural network (ANN) based models are developed to study the effect of design variables on computed OSIs. Cuckoo search neural network systems (CSNNS) are finally utilized to constrainedly optimize the cutting process per each cutting scenario. The most appropriate cutting setup to ensure successful turning of standard EN 1.4462 and super EN 1.4410 for each scenario is selected in accordance with conditions which give the maximum OSI. (C) 2014 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.

2014




* 367(<-663): A systemic self-modelling method and its application to material design and optimization

A self-modelling system for material research has been developed based
on discriminant analysis, artificial neural networks, classification
mapping and genetic algorithms. It provides systemic methodologies for
nonlinear multivariate modelling and multi-objective optimizing. It is
designed to unveil connotative information from a limited experimental
data set and gives qualitative, quantitative and geometry models of
the object to be researched. In addition, optimized research schemes
can be derived from these models by genetic algorithms and
classification mapping. The technique is suitable for subjects that
have some original study results but for the following reasons there
are difficulties in doing further research. (i) The object researched
has too many controlling factors and is too complex to analyse. (ii)
The object is controlled by some unexplainable mechanisms and is
difficult to analyse. (iii) The mathematical expression has strong
nonlinearity and is difficult to resolve strictly.

2001


* 368(<-476): Multi-objective optimization of material selection for sustainable products: Artificial neural networks and genetic algorithm approach

Material properties and selection are very important in product design. To get more sustainable products. not only the technical and economic factors, but also the environmental factors should be considered. To satisfy the requirements, evaluation indicators of materials are presented. Environmental impacts were Calculated by the Life Cycle Assessment method (LCA method). An integration of artificial neural networks (ANN) with genetic algorithms (GAs) is proposed to optimize the multi-objectives of material selection, it was validated by an example that the system can select suitable materials to develop sustainable products. (C) 2008 Elsevier Ltd. All rights reserved.

2009





* 372(<-433): Multi-objective scheduling of dynamic job shop using variable neighborhood search

Dynamic job shop scheduling that considers random job arrivals and machine breakdowns is studied in this paper. Considering an event driven policy rescheduling, is triggered in response to dynamic events by variable neighborhood search (VNS). A trained artificial neural network (ANN) updates parameters of VNS at any rescheduling point. Also, a multi-objective performance measure is applied as objective function that consists of makespan and tardiness. The proposed method is compared with some common dispatching rules that have widely used in the literature for dynamic job shop scheduling problem. Results illustrate the high effectiveness and efficiency of the proposed method in a variety of shop floor conditions. (C) 2009 Published by Elsevier Ltd.

2010




* 378(<-182): Application of computational intelligence techniques for load shedding in power systems: A review

Recent blackouts around the world question the reliability of
conventional and adaptive load shedding techniques in avoiding such
power outages. To address this issue, reliable techniques are required
to provide fast and accurate load shedding to prevent collapse in the
power system. Computational intelligence techniques, due to their
robustness and flexibility in dealing with complex non-linear systems,
could be an option in addressing this problem. Computational
intelligence includes techniques like artificial neural networks,
genetic algorithms, fuzzy logic control, adaptive neuro-fuzzy
inference system, and particle swarm optimization. Research in these
techniques is being undertaken in order to discover means for more
efficient and reliable load shedding. This paper provides an overview
of these techniques as applied to load shedding in a power
system. This paper also compares the advantages of computational
intelligence techniques over conventional load shedding
techniques. Finally, this paper discusses the limitation of
computational intelligence techniques, which restricts their usage in
load shedding in real time. (C) 2013 Elsevier Ltd. All rights
reserved.

2013



* 380(<-368): Machine scheduling in custom furniture industry through neuro-evolutionary hybridization

Machine scheduling is a critical problem in industries where products are custom-designed. The wide range of products, the lack of previous experiences in manufacturing, and the several conflicting criteria used to evaluate the quality of the schedules define a huge search space. Furthermore, production complexity and human influence in each manufacturing step make time estimations difficult to obtain thus reducing accuracy of schedules. The solution described in this paper combines evolutionary computing and neural networks to reduce the impact of (i) the huge search space that the multi-objective optimization must deal with and (ii) the inherent problem of computing the processing times in a domain like custom manufacturing. Our hybrid approach obtains near optimal schedules through the Non-dominated Sorting Genetic Algorithm II (NSGA-II) combined with time estimations based on multilayer perceptron neural networks. (C) 2010 Elsevier B. V. All rights reserved.

2011






* 381(<-541): Multi-objective particle swarm optimization hybrid algorithm: An application on industrial cracking furnace

In this paper, a new multi-objective particle swarm optimization (MOPSO) procedure, based on the Pareto dominance hybrid algorithm, is proposed and applied in a naphtha industrial cracking furnace for the first time. Pareto dominance is incorporated into particle swarm optimization (PSO). Our algorithm takes the Pareto set as a repository of particles that is later used by other particles to guide their own flight. In addition, an MOPSO and artificial neural network (ANN) hybrid model is applied in the operation optimization of a naphtha industrial cracking furnace. Therein, sensitivity analysis is investigated and taken as the basis on which decision variables of multi-objective problem base. From both theoretical computation and practical application, the validity and reliability of proposed algorithm are verified by two test functions studied, and actual application example of the optimization of operation parameter of cracking process. Moreover, the yields of ethylene and propylene are improved.

2007





* 382(<-598): Global optimization of a feature-based process sequence using GA and ANN techniques

Operation sequencing has been a key area of research and development
for computer-aided process planning (CAPP). An optimal process
sequence could largely increase the efficiency and decrease the cost
of production. Genetic algorithms (GAs) are a technique for seeking to
`breed' good solutions to complex problems by survival of the
fittest. Some attempts using GAs have been made on operation
sequencing optimization, but few systems have intended to provide a
globally optimized fittest function definition. In addition, most of
the systems have a lack of adaptability or have an inability to
learn. This paper presents an optimization strategy for process
sequencing based on multi-objective fittness:minimum manufacturing
cost, shortest manufacturing time and best satisfaction of
manufacturing sequence rules. A hybrid approach is proposed to
incorporate a genetic algorithm, neural network and analytical
hierarchical process (AHP) for process sequencing. After a brief study
of the current research, relevant issues of process planning are
described. A globally optimized fittness function is then defined
including the evaluation of manufacturing rules using AHP, calculation
of cost and time and determination of relative weights using neural
network techniques. The proposed GA-based process sequencing, the
implementation and test results are discussed. Finally, conclusions
and future work are summarized.

2005



* 383(<-633): An intelligent simulation method based on artificial neural network for container yard operation

This paper presents an intelligent simulation method for regulation of
container yard operation on container terminal. This method includes
the functions of system status evaluation, operation rule and stack
height regulation, and operation scheduling. In order to realize
optimal operation regulation, a control architecture based on fuzzy
artificial neural network is established. The regulation process
includes two phases: prediction phase forecasts coming container
quantity; inference phase makes decision on operation rule and stack
height. The operation scheduling is a fuzzy multi-objective
programming problem with operation criteria such as minimum ship
waiting time and operation time. The algorithm combining genetic
algorithm with simulation is developed. A case study is presented to
verify the validity and usefulness of the method in simulation
environment.

2004



* 384(<-237): Optimization of the Activated Sludge Process

This paper presents a multiobjective model for optimization of the activated sludge process (ASP) in a wastewater-treatment plant (WWTP). To minimize the energy consumption of the activated sludge process and maximize the quality of the effluent, three different objective functions are modeled [i.e., the airflow rate, the carbonaceous biochemical oxygen demand (CBOD) of the effluent, and the total suspended solids (TSS) of the effluent]. These models are developed using a multilayer perceptron (MLP) neural network based on industrial data. Dissolved oxygen (DO) is the controlled variable in these objectives. A multiobjective model that included these objectives is solved with a multiobjective particle swarm optimization (MOPSO) algorithm. Computation results are reported for three trade-offs between energy savings and the quality of the effluent. A 15% reduction in airflow can be achieved by optimal settings of dissolved oxygen, provided that energy savings take precedence over the quality of the effluent. DOI: 10.1061/(ASCE)EY.1943-7897.0000092. (C) 2013 American Society of Civil Engineers.

2013





* 387(<- 63): Thermochromic sensor design based on Fe(II) spin crossover/polymers hybrid materials and artificial neural networks as a tool in modelling

This article explores the use of multi-objective evolutionary machine learning techniques to find the minimum number of sensors from a pull of 6 sensors as well as the minimum number of analytical signals belonging to each selected sensor for the design of an optimal colourimetric temperature sensor. The analytical information was obtained with a calibrated neural network that provides the best temperature estimation with respect to the selected colourimetric sensor responses from a previously developed sensor array. The sensor array was developed by embedding the linear spin crossover material [Fe-(NH(2)trz)(3)](BF4)(2) into polymers with different polarity, offering different thermochromic responses related to different morphologies of the spin crossover particles when embedded in each polymer. The different thermochromic responses are tracked by the green component of the RGB colour space and the a* from CIEL*a*b* obtained with a conventional photographic digital camera. These two colour signals are used as analytical parameters for the subsequent computer processing and model calibration. The use of multi-objective optimization techniques for neural network calibration demonstrated that only 3 signals coming from 3 sensors of the 6 studied are sufficient to provide optimal temperature estimation. The optimized selection was the green channel from polyurethane hydrogel D6 and PVC prepared in THF and a* from PMMA prepared in toluene. (C) 2014 Elsevier B.V. All rights reserved.

2015





* 388(<-104): Neuro-genetic multi-objective optimization and computer-aided design of pantoprazole molecularly imprinted polypyrrole sensor

A molecularly imprinted polymer (MIP) of pantoprazole (PNZ) was prepared through electropolymerization of pyrrole on a functionalized multi-walled carbon nanotube modified pencil graphite electrode. The preparation of MIP and quantitative measurements were performed by cyclic voltammetry and differential pulse voltammetry (DPV), respectively. Several important parameters controlling the performance of polypyrrole film. The factors, i.e. pH of buffer solution, cyclic voltammetric scan rate in polymerization step, number of cyclic voltammetric scans, monomer and template concentrations in prepolymerization mixture, nanotube concentration in functionalized multi-walled carbon nanotubes-coating step, uptake time after MIP preparation and uptake step stirring rate were expected to affect MIP preparation and voltammetric measurements. The optimization of parameters was performed using Plackett-Burman design, central composite design, artificial neural network and genetic algorithm. The Pareto plot showed that effects of monomer concentration and pH are most important to the process. The best MIP to NIP response ratio was obtained 17.4. The selection of monomer was performed computationally using ab initio calculations. The calibration curve demonstrated linearity over a concentration range of 5-700 mu M with a correlation coefficient (r) of 0.9980. The detection limit of PNZ was obtained 3.75 x 10(-7) M. The minimum and maximum recovery (%) through the spiking 0.1-0.4 mM PNZ to a biological and some pharmaceutical matrices were obtained 95.9% (human blood serum) and 106% (PNZ tablet), respectively. (C) 2014 Elsevier B.V. All rights reserved.

2014





* 389(<-372): A SELF ORGANIZING MAP BASED HYBRID MULTI-OBJECTIVE OPTIMIZATION OF WATER DISTRIBUTION NETWORKS

Water Distribution Networks (WDNs) are an essential infrastructure of
every civilization. In the past decades, there has been a lot of work
on the optimization of WDNs. This paper presents a hybrid NSGA-II for
multi-objective optimization of combinatorial WDN design, utilizing
the SOM network as a tool to find the genotypic or phenotypic
similarities. SOM is a versatile unsupervised Artificial Neural
Network (ANN) that can be used to extract the similarities and find
the related vectors with the use of a proper similarity measure. The
proposed method, SOM-NSGA-II, derives subpopulations or virtual
islands for inbreeding similar individuals to speed up the convergence
process of the optimization. The cross-over operation between similar
individuals of the subpopulations at the constraint dominated region
of the solution space showed a faster convergence and a wider
Pareto-front for the test problems considered. An added advantage of
the method is the application of genotypic sorting of the population
by SOM for visual representation of the structure of the Pareto
front. The resulted maps showed the extent of variation of the
decision variables and their relative importance. This method may be
utilized to speed up optimization of large scale WDNs and as an
important visual aid for decision makers and designers of WDNs.

2011





* 393(<-681): IN-PROCESS REGRESSIONS AND ADAPTIVE MULTICRITERIA NEURAL NETWORKS FOR MONITORING AND SUPERVISING MACHINING OPERATIONS

The authors develop a monitoring and supervising system for machining
operations using in-process regressions (for monitoring) and adaptive
feedforward artificial neural networks (for supervising).  The system
is designed for: (1) in-process tool life measurement and prediction;
(2) supervision of machining operations in terms of the best machining
setup; and (3) catastrophic tool failure monitoring.  The monitoring
system predicts tool life by using different sensors for gathering
information based on a regression model that allows for the variations
between tools and different machine setups.  The regression model
makes its prediction by using the history of other tools and combining
it with the information obtained about the tool under consideration.
The supervision system identifies the best parameters for the machine
setup problem within the framework of multiple criteria decision
making.  The decision maker (operator) considers several criteria,
such as cutting quality, production rate and tool life.  To make the
optimal decision with several criteria, an adaptive feedforward
artificial neural network is used to assess the decision maker's
preferences.  The authors' neural network approach learns from the
decision maker's complex behavior and hence, in automatic mode, can
make decisions for the decision maker. The approach is not
computationally demanding, and experiments demonstrate that its
predictions are accurate.

1995



* 394(<- 65): Decision support for management of urban transport projects

The planning phase within the urban-transport project management is a complex process from both the management and techno-economic aspects. The focus of this research is on decision-making processes related to the planning phase during management of urbanroad infrastructure projects. The proposed concept is based on multicriteria methods and Artificial Neural Networks. The decision-support concept presented in this paper is tested on the road infrastructure of the city of Split, and it shows how urban road infrastructure planning can be improved.

2015



* 396(<-232): Optimization of mechanical property and shape recovery behavior of Ti-(similar to 49 at.%) Ni alloy using artificial neural network and genetic algorithm

Multi-objective genetic algorithm based searching is used for designing the process schedule of Ti-(similar to 49 at.%) Ni alloy, to achieve optimum mechanical property and shape recovery behavior. Artificial neural network technique based data driven models are developed to empirically describe the relationship between the processing conditions and the properties. The models are used as objective functions for the optimization process. The optimization search found to be helpful to design the decision space variables for the improvement in shape recovery behavior without sacrificing the mechanical properties of the alloy. The Pareto solutions have been used as the guideline to find the process schedules, which is validated by suitable experimentation. (C) 2012 Elsevier Ltd. All rights reserved.

2013




* 397(<-509): An integrated multi-objective immune algorithm for optimizing the wire bonding process of integrated circuits

Optimization of the wire bonding process of an integrated circuit (IC) is a multi-objective optimization problem (MOOP). In this research, an integrated multi-objective immune algorithm (MOIA) that combines an artificial immune algorithm (IA) with an artificial neural network (ANN) and a generalized Pareto-based scale-independent fitness function (GPSIFF) is developed to find the optimal process parameters for the first bond of an IC wire bonding. The back-propagation ANN is used to establish the nonlinear multivariate relationships between the wire boning parameters and the multi-responses, and is applied to generate the multiple response values for each antibody generated by the IA. The GPSIFF is then used to evaluate the affinity for each antibody and to find the non-dominated solutions. The "Error Ratio" is then applied to measure the convergence of the integrated approach. The "Spread Metric" is used to measure the diversity of the proposed approach. Implementation results show that the integrated MOIA approach does generate the Pareto-optimal solutions for the decision maker, and the Pareto-optimal solutions have good convergence and diversity performance.

2008


* 398(<-521): Design of electroceramic materials using artificial neural networks and multiobjective evolutionary algorithms

We describe the computational design of electroceramic materials with optimal permittivity for application as electronic components. Given the difficulty of large-scale manufacture and characterization of these materials, including the theoretical prediction of their materials properties by conventional means, our approach is based on a recently established database containing composition and property information for a wide range of ceramic compounds. The electroceramic materials composition-function relationship is encapsulated by an artificial neural network which is used as one of the objectives in a multiobjective evolutionary algorithm. Evolutionary algorithms are stochastic optimization techniques which we employ to search for optimal materials based on chemical composition. The other objectives optimized include the reliability of the neural network prediction and the overall electrostatic charge of the material. The evolutionary algorithm searches for materials which simultaneously have high relative permittivity, minimum overall charge, and Good prediction reliability. We find that we are able to predict a range of new electroceramic materials with varying degrees of reliability. In some cases the materials are similar to those contained in the database; in others, completely new materials are predicted.

2008



* 399(<-574): Approach to optimization of cutting conditions by using artificial neural networks

Optimum selection of cutting conditions importantly contribute to the
increase of productivity and the reduction of costs, therefore utmost
attention is paid to this problem in this contribution. In this paper,
a neural network-based approach to complex optimization of cutting
parameters is proposed. It describes the multi-objective technique of
optimization of cutting conditions by means of the neural networks
taking into consideration the technological, economic and
organizational limitations. To reach higher precision of the predicted
results. a neural optimization algorithm is developed and presented to
ensure simple, fast and efficient optimization of all important
turning parameters. The approach is suitable for fast determination of
optimum cutting parameters during machining. where there is not enough
time for deep analysis. To demonstrate the procedure and performance
of the neural network approach, an illustrative example is discussed
in detail. (c) 2005 Elsevier B.V. All rights reserved.

2006





* 401(<-674): Evaluation of factors and current approaches related to computerized design of tillage tools: a review

The objectives of this paper are to evaluate the factors that are
involved in the tillage process, and to explore the potential
approaches for the computer-aided design of tillage tools. An overview
related to the dynamic effect on the performance of tillage operations
has been conducted. Compared with the analytical methods, the finite
element method (FEM) has some advantages for the computerized design
of tillage tools. The artificial neural networks (ANN) may be useful
for the integrated evaluation of tillage performance with
multi-objectives. ANN can be employed for simulation of a dynamic
constitutive model and identification of soil conditions for
agricultural soils. The integral approach of ANN analysis with FEM is
found to be promising for optimizing design of tillage tools. (C) 1998
ISTVS. All rights reserved.

1998




* 402(<-682): COMPARING BP AND ART-II NEURAL-NETWORK CLASSIFIERS FOR FACILITY LOCATION

This paper compares the performance of Artificial Neural Networks
(ANNs) as classifiers in the facility location domain. The ART II
(Adaptive Resonance Theory) and BP (Back Propagation) paradigms are
used as exemplars of ANNs developed using supervised and unsupervised
learning. Their performances are compared with that obtained using a
linear multi-attribute utility model (MAUM) to classify the 48 states
in the continental U.S.A. based on location profiles developed from
government publications. In this paper, the models are used to
classify the U.S. states based on their suitability for accommodating
new manufacturing facilities. For this data set, the BP ANN model
displayed robust performance and showed better convergence with the
MAUM.

1995




* 403(<-267): Application of a linearly decreasing weight particle swarm to optimize the process conditions of al matrix nanocomposites

In this paper, SiC nanoparticles were added into the commercial
casting Al-Si aluminum alloy to fabricate metal matrix nanocomposites
(MMNCs) with uniform reinforcement distribution. Experimental results
revealed that the presence of nano-SiC reinforcement led to
significant improvement in hardness and UTS while the ductility of the
aluminum matrix is retained. An integrated optimization approach using
an artificial neural network and a modified particle swarm is proposed
to solve a process parameter design problem in casting. The artificial
neural network is used to obtain the relationships between decision
variables and the performance measures of interest, while the particle
swarm is used to perform the optimization with multiple
objectives. The results showed that the particle swarm is an effective
method for solving multi-objective optimization problems, and that an
integrated approach using an artificial neural network and a modified
particle swarm can be used to solve complex process parameter design
problems.

2012


* 404(<-353): Optimization of tile manufacturing process using particle swarm optimization

In this paper, an integrated optimization approach using an artificial neural network and a bidirectional particle swarm is proposed. The artificial neural network is used to obtain the relationships between decision variables and the performance measures of interest, while the bidirectional particle swarm is used to perform the optimization with multiple objectives. Finally, the proposed approach is used to solve a process parameter design problem in cement roof-tile manufacturing. The results showed that the bidirectional particle swarm is an effective method for solving multi-objective optimization problems, and that an integrated approach using an artificial neural network and a bidirectional particle swarm can be used to solve complex process parameter design problems. (C) 2011 Elsevier B.V. All rights reserved.

2011





* 420(<-158): Optimal spatial allocation of water resources based on Pareto ant colony algorithm

The spatial allocation of water resources is optimised using the
multi-objective functions and multi-constrained conditions of the
Pareto ant colony algorithm (PACA). The objective function is the
highest benefit to the economy, society and the environment, while the
constraints include water supply, demand and quality. The PACA is
improved by limiting local pheromone scope and dynamically updating
global pheromone levels. Since both strategies guide the ant towards
borders of high-pheromone concentration, the new approach enhances the
global search capability and convergence speed. Programming, database
management and interface tools are then integrated into geographic
information systems (GIS) software. The study area is located in
Zhenping County, Henan Province, China, and water resource data are
obtained using remote sensing (RS) and GIS technology. The improved
PACA is solved in the GIS environment. Optimal spatial allocation
schemes are obtained for surface, ground and transferred water and the
model yields optimal spatial benefit schemes of water resources,
embracing economic, social and ecological benefits. The results of
improved PACA are superior to those of other intelligent optimisation
algorithms, including the ant colony algorithm, multi-objective
genetic algorithm and back-propagation artificial neural
network. Therefore, the integration of RS, GIS and PACA can
effectively optimise the large-scale, multi-objective allocation of
water resources. The model also enhances the global search capability,
convergence speed and result precision, and can potentially solve
other optimal spatial problems with multi-objective functions.

2014



* 421(<- 39): A New Hybrid Artificial Intelligence Approach to Predicting Global Thermal Comfort of Stretch Knitted Fabrics

Today numerous consumers consider thermal comfort to be one of the
most significant attributes when purchasing textile and apparel
products, so there is a need to develop a model able to simulate
objectively the consumers' perception. The global thermal comfort of
stretch knitted fabrics is a:multi-criteria phenomenon that requires
the satisfaction of several properties at the same time. In this
paper, we used the desirability functions to evaluate the satisfaction
degree of global thermal comfort. Statistical method was used to
investigate the interrelationship among knit thermo-physical
properties, and group them into factors. Two models of artificial
neural network (general and special) have been set up to predict the
global thermal comfort from structural parameters (inputs) of knitted
fabrics made from pure yam cotton (cellulose) and viscose (regenerated
cellulose) fibers and plated knitted with elasthane (Lycra) fibers. A
virtual leave one out approach dealing with over fitting phenomenon
and allowing the selection of the optimal neural network architecture
was used. By combining the strengths of statistics and fuzzy logic
(data reduction and information summation) also a neural network
(self-learning ability), hybrid model was developed to simulate the
consumer thermal comfort perception. After that, ANN model is
inverted. With a required output value and some input parameters it is
possible to calculate the unknown optimum input parameter. Finally,
this forecasting can help industrials to anticipate the consumer's
taste. Thus, they can adjust the knitting production parameter to
reach the desired global thermal comfort to satisfy this consumer.

2015


* 422(<-172): An evolutionary geometric primitive for automatic design synthesis of functional shapes: The case of airfoils

A novel self-adaptive geometric primitive for functional geometric shape synthesis is presented. This novel geometric primitive, for CAD use, is specifically designed to reproduce geometric shapes with functional requirements, such as the aerodynamic and hydrodynamic ones, once the functional parameters are furnished. It produces a typical CAD representation of a functional profile: a set of Bezier curves. The proposed primitive follows a generate-and-test approach and takes advantage of the use of a properly designed artificial neural network (BNN). It combines the properties of a geometric primitive and the capability to manage the engineering knowledge in a specific field of application. The proposed evolutionary primitive is applied to a real engineering application: the automatic synthesis of airfoils. Some examples are simulated in order to test the effectiveness of the proposed method. The results obtained by an original prototypal software are presented and critically discussed. (C) 2013 Elsevier Ltd. All rights reserved.

2014


* 423(<-292): Multi-objective shape optimization of transonic airfoil sections using swarm intelligence and surrogate models

In this article, the optimization problem of designing transonic airfoil sections is solved using a framework based on a multi-objective optimizer and surrogate models for the objective functions and constraints. The computed Pareto-optimal set includes solutions that provide a trade-off between maximizing the lift-to-drag ratio during cruise and minimizing the trailing edge noise during the aircraft's approach to landing. The optimization problem was solved using a recently developed multi-objective optimizer, which is based on swarm intelligence. Additional computational intelligence tools, e.g., artificial neural networks, were utilized to create surrogate models of the objective functions and constraints. The results demonstrate the effectiveness and efficiency of the proposed optimization framework when applied to simulation-based engineering design optimization problems.

2012


* 424(<-313): Multi-point robust design optimization of wind turbine airfoil under geometric uncertainty

Modern wind turbine airfoil designs are increasingly emphasizing low sensitivity to the leading edge roughness in addition to good aerodynamic performances under varying wind conditions. In this study, a multi-point robust design optimization method has been systematically established for the wind turbine airfoil. The objective is to not only maximize the lift-to-drag ratio and lift coefficient at and near the design point, but also minimize their sensitivity to the leading edge roughness associated with the geometry profile uncertainty. The geometry parameterization of the airfoil is conducted by the Bezier curves. In the robust optimization, the multi-objective genetic algorithm, Monte Carlo simulation technique, and artificial neural network model are used. The results show that both the robust Pareto optimal (RPO) and deterministic Pareto optimal (DPO) airfoils have higher lift-to-drag ratio and lift coefficient than the original design FX 63-137 at and near the design point. A smaller maximum camber and larger radii near the leading edge help the RPO airfoil outperform both the DPO and original ones in terms of low sensitivity to the leading edge roughness. This study may be useful to the future development of the wind turbine airfoils with higher efficiency and reliability.

2012




* 425(<-393): SIPAR_ID: Freeware for surface irrigation parameter identification

SIPAR_ID is a software based on a robust multiobjective inverse
modeling technique for estimating field values of infiltration and
roughness parameters of a surface irrigation event under both steady
and variable inflow conditions. Its simulation engine is quite
flexible and accurate thanks to a hybrid model that combines a
volume-balance model with artificial neural networks. SIPAR_ID also
provides an estimate of the uncertainty and sensitivity of the
identified parameters. (C) 2008 Elsevier Ltd. All rights reserved.

2010




* 427(<-520): Stochastic sampling design for water distribution model calibration

A novel approach to determine optimal sampling locations under
parameter uncertainty in a water distribution system (WDS) for the
purpose of its hydraulic model calibration is presented. The problem
is formulated as a multi-objective optimisation problem under
calibration parameter uncertainty The objectives are to maximise the
calibrated model accuracy and to minimise the number of sampling
devices as a surrogate of sampling design cost. Model accuracy is
defined as the average of normalised traces of model prediction
covariance matrices, each of which is constructed from a randomly
generated sample of calibration parameter values. To resolve the
computational time issue, the optimisation problem is solved using a
multi-objective genetic algorithm and adaptive neural networks
(MOGA-ANN). The verification of results is done by comparison of the
optimal sampling locations obtained using the MOGA-ANN model to the
ones obtained using the Monte Carlo Simulation (MCS) method. In the
MCS method, an equivalent deterministic sampling design optimisation
problem is solved for a number of randomly generated calibration model
parameter samples. The results show that significant computational
savings can be achieved by using MOGA-ANN compared to the MCS model or
the GA model based on all full fitness evaluations without significant
decrease in the final solution accuracy.

2008



* 428(<-222): Research of traffic flow multi-objectives intelligent control method for junction network

Junction network is a special type of roadwork pattern that scatters
and distributes around the specific zone of metropolitan and in that
contains different grade and functional roads of arterial road, urban
freeway and expressway. Intelligent control is new development where
the control problem is to find the combination of control measures
that result for the best road performance and control
effectiveness. The problems of multi-objective coordinated metering
and evaluation for local ramp is considered. This paper discusses the
optimal coordination of mainline and ramp, a modified ramp latency
model is posed using the method of queuing theory, and a ramp control
with better mechanism compare to artificial neural network using
radical based function-support vector machine algorithm is
designed. With in-situ traffic flow data of Beijing ring and radial
freeway during high-density period, three known and the designed novel
methodologies are compared, the intensive simulations show the
effectiveness of the proposed approach, particularly at the aspect of
minimize reduplicated waiting time for junction network. using these
methodologies demonstrates the comparative control efficiency and
accuracy.

2013




* 431(<-493): Optimization of composition of heat-treated chromium white cast iron casting by phosphate graphite mold

In present work. the difference among orthogonal design, Fuzzy optimum
design and artificial neural network ANN was performed on the basis of
the optimization of chemical composition of chromium white cast
iron. It is found that Fuzzy optimum design is suitable for
multi-objective comprehensive evaluation. and the optimum composition
of white cast iron is Cr 4%, Si 3.5%, Mn 3% and Cu 1% in the
orthogonal array. On the other hand. the orthogonal analysis is
suitable for analyzing the effect of each factor on the performances
and obtaining the theoretical optimum combination of each factor for
the performances and the optimum theoretical performances,
respectively. Moreover, the prediction and simulation results show
that the RBFANN not only can be used to establish the model with high
accuracy for the orthogonal test but also outperforms the traditional
orthogonal analysis method. Therefore, the combination of three
methods can more effectively deal with the optimization of chemical
composition of materials. (C) 2007 Elsevier B.V. All rights reserved.

2008



* 440(<-639): Neural networks and reinforcement learning in control of water systems

In dynamic real-time control (RTC) of regional water systems, a
multicriteria optimization problem has to be solved to determine the
optimal control strategy. Nonlinear and/or dynamic programming based
on simulation models can be used to find the solution, an approach
being used in the Aquarius decision support system (DSS) developed in
The Netherlands. However, the computation time required for complex
models is often prohibitively long, and therefore such a model cannot
be applied in RTC of water systems. In this study, Aquarius DSS is
chosen as a reference model for building a controller using machine
learning techniques such as artificial neural networks (ANN) and
reinforcement learning (RL), where RL is used to decrease the error of
the ANN-based component. The model was tested with complex water
systems in The Netherlands, and very good results were obtained. The
general conclusion is that a controller, which has learned to
replicate the optimal control strategy, can be used in RTC operations.

2003




* 442(<-141): Blade Tip Shape Optimization for Enhanced Turbine Aerothermal Performance

In high-speed, unshrouded turbines, tip leakage flows generate large aerodynamic losses and intense unsteady thermal loads over the rotor blade tip and casing. The stage-loading and rotational speeds are steadily increased to achieve higher turbine efficiency, and hence, the overtip leakage flow may exceed the transonic regime. However, conventional blade tip geometries are not designed to cope with supersonic tip flow velocities. A great potential lies in the modification and optimization of the blade tip shape as a means to control the tip leakage flow aerodynamics, limit the entropy production in the overtip gap, manage the heat-load distribution over the blade tip, and improve the turbine efficiency at high-stage loading coefficients. The present paper develops an optimization strategy to produce a set of blade tip profiles with enhanced aerothermal performance for a number of tip gap flow conditions. The tip clearance flow was numerically simulated through two-dimensional compressible Reynolds-averaged Navier-Stokes (RANS) calculations that reproduce an idealized overtip flow along streamlines. A multiobjective optimization tool, based on differential evolution combined with surrogate models (artificial neural networks), was used to obtain optimized 2D tip profiles with reduced aerodynamic losses and minimum heat transfer variations and mean levels over the blade tip and casing. Optimized tip shapes were obtained for relevant tip gap flow conditions in terms of blade thickness to tip gap height ratios (between 5 and 25) and blade pressure loads (from subsonic to supersonic tip leakage flow regimes), imposing fixed inlet conditions. We demonstrated that tip geometries that perform superior in subsonic conditions are not optimal for supersonic tip gap flows. Prime tip profiles exist, depending on the tip flow conditions. The numerical study yielded a deeper insight on the physics of tip leakage flows of unshrouded rotors with arbitrary tip shapes, providing the necessary knowledge to guide the design and optimization strategy of a full blade tip surface in a real 3D turbine environment.

2014



* 443(<-166): Robust design of Mars entry micro-probe with evidence theory and multi-fidelity strategies

Purpose - A multi-disciplinary robust design optimization method for
micro Mars entry probe (no more than 0.8m in diameter) is
proposed. The purpose of this paper is to design a Mars entry probe,
not only the geometric configuration, but the trajectory and thermal
protection system (TPS). In the design optimization, the uncertainties
of atmospheric and aerodynamic parameters are taken into account. The
probability distribution information of the uncertainties are supposed
to be unknown in the design. To ensure accuracy levels, time-consuming
numerical models are coupled in the optimization. Multi-fidelity
approach is designed for model management to balance the computational
cost and accuracy. Design/methodology/approach - Uncertainties which
cannot defined by usual Gaussian probability distribution are modeled
with degree of belief, and optimized through with multiple-objective
optimization method. The optimization objectives are set to be the
thermal performance of the probe TPS and the corresponding belief
values. Robust Pareto front is obtained by an improved multi-objective
density estimator algorithm. Multi-fidelity management is performed
with an Artificial Neural Network (ANN) surrogate model. Analytical
model is used first, and then with the improvement of accuracy, rather
complex numerical models are activated. ANN updates the database
during the optimization, and makes the solutions finally converge to a
high-level accuracy. Findings - The optimization method provides a way
for conducting complex design optimization involving multi-discipline
and multi-fidelity models. Uncertainty effects are analyzed and
optimized through multi-disciplinary robust design. Because of the
micro size, and uncertain impacts of aerodynamic and atmospheric
parameters, simulation results show the performance trade-off by the
uncertainties. Therefore an effective robust design is necessary for
micro entry probe, particularly when details of model parameter are
not available. Originality/value - The optimization is performed
through a new developed multi-objective density estimator
algorithm. Affinity propagation algorithm partitions adaptively the
samples by passing and analyzing messages between data points. Local
principle component techniques are employed to resample and reproduce
new individuals in each cluster. A strategy similar to NSGA-II selects
data with better performance, and converges to the Pareto
front. Models with different fidelity levels are incorporated in the
multi-disciplinary design via ANN surrogate model. Database of
aerodynamic coefficients is updated in an online manner. The
computational time is greatly reduced while keeping nearly the same
accuracy level of high fidelity model.

2014





* 445(<-420): Simulation-Optimization Modeling: A Survey and Potential Application in Reservoir Systems Operation

This paper presents a survey of simulation and optimization modeling
approaches used in reservoir systems operation problems. Optimization
methods have been proved of much importance when used with simulation
modeling and the two approaches when combined give the best
results. The main objective of this review article is to discuss
simulation, optimization and combined simulation-optimization modeling
approach and to provide an overview of their applications reported in
literature. In addition to classical optimization techniques,
application and scope of computational intelligence techniques, such
as, evolutionary computations, fuzzy set theory and artificial neural
networks, in reservoir system operation studies are
reviewed. Conclusions and suggestive remarks based on this survey are
outlined, which could be helpful for future research and for system
managers to decide appropriate methodology for application to their
systems.

2010




* 446(<-434): A multi-criteria adaptive control scheme based on neural networks and fuzzy inference for DRC manufacturing systems

Manufacturing systems are uncertain and dynamic systems, hence, they
require real-time scheduling to adapt to changing manufacturing
conditions. Current real-time scheduling approaches have been devised
mainly for machine-only constrained systems, in which the shop
capacity is constrained only by machine capacity, rather than for dual
resource constrained (DRC) systems, in which the shop capacity is
constrained by machine and worker capacity. In particular, there is no
study on DRC system scheduling in which the 'where' and 'when' worker
assignment rules, basic features of DRC systems, are altered in
real-time (dynamically selected) to respond to new manufacturing
conditions. Besides, multi-criteria DRC system scheduling has not yet
been addressed extensively. Also, there has been little research on
the interactions of dynamically selected job dispatching, worker
assignment and job routing rules, which have a significant impact on
DRC system performance. This paper proposes a multi-criteria real-time
scheduling methodology for DRC systems to address these issues, and
investigates these interactions. The methodology employs artificial
neural networks as meta-models to reduce computational complexity and
a fuzzy inference system to cope with multiple performance
criteria. Various simulation experiments demonstrate that the
methodology provides satisfactory results for real-time DRC systems
scheduling.

2010



* 447(<-571): Prediction of building's temperature using neural networks models

The use of artificial neural networks in various applications related
with energy management in buildings has been increasing significantly
over the recent years. In this paper the design of inside air
temperature predictive neural network models, to be used for
predictive control of air-conditioned systems, is discussed. The use
of multi-objective genetic algorithms for designing off-line radial
basis function neural network models is detailed. The performance of
these data-driven models is compared, favourably, with a multi-node
physically based model. Climate and environmental data from a
secondary school located in the south of Portugal, collected by a
remote data acquisition system, are used to generate the models. By
using a sliding window adaptive methodology, the good results obtained
off-line are extended throughout the whole year. The use of long-range
predictive models for air-conditioning systems control is
demonstrated, in simulations, achieving a good temperature regulation
with important energy savings. (C) 2005 Elsevier B.V. All rights
reserved.

2006




* 449(<-490): Decision support system for optimal reservoir operation modeling within sediment deposition control

Suspended sediment deals with surface runoff moving toward watershed
affects reservoir sustainability due to the reduction of storage
capacity. The purpose of this study is to introduce a reservoir
operation model aimed at minimizing sediment deposition and maximizing
energy production expected to obtain optimal decision policy for both
objectives. The reservoir sediment-control operation model is
formulated by using Non-Linear Programming with an iterative procedure
based on a multi-objective measurement in order to achieve optimal
decision policy that is established in association with the
development of a relationship between stream inflow and sediment rate
by utilizing the Artificial Neural Network. Trade off evaluation is
introduced to generate a strategy for controlling sediment deposition
at same level of target ratio while producing hydroelectric
energy. The case study is carried out at the Sanmenxia Reservoir in
China where redesign and reconstruction have been
accomplished. However, this model deals only with the original design
and focuses on a wet year operation. This study will also observe a
five-year operation period to show the accumulation of sediment due to
the impact of reservoir storage capacity.

2009



* 454(<-651): Calibration of stochastic cellular automata: the application to rural-urban land conversions

Despite the recognition of cellular automata (CA) as a flexible and
powerful tool for urban growth simulation, the calibration of CA had
been largely heuristic until recent efforts to incorporate
multi-criteria evaluation and artificial neural network into rule
definition. This study developed a stochastic CA model, which derives
its initial probability of simulation from observed sequential land
use data. Furthermore, this initial probability is updated dynamically
through local rules based on the strength of neighbourhood
development. Consequentially the integration of global (static) and
local (dynamic) factors produces more realistic simulation
results. The procedure of calibrated CA can be applied in other
contexts with minimum modification. In this study we applied the
procedure to simulate rural-urban land conversions in the city of
Guangzhou, China. Moreover, the study suggests the need to examine the
result of CA through spatial, tabular and structural validation.

2002



* 461(<-102): Multi-objective optimization for building retrofit: A model using genetic algorithm and artificial neural network and an application

Retrofitting of existing buildings offers significant opportunities
for improving occupants' comfort and well-being, reducing global
energy consumption and greenhouse gas emissions. This is being
considered as one of the main approaches to achieve sustainability in
the built environment at relatively low cost and high uptake
rates. Although a wide range of retrofit technologies is readily
available, methods to identify the most suitable set of retrofit
actions for particular projects are still a major technical and
methodological challenge. This paper presents a multi-objective
optimization model using genetic algorithm (GA) and artificial neural
network (ANN) to quantitatively assess technology choices in a
building retrofit project. This model combines the rapidity of
evaluation of ANNs with the optimization power of GAs. A school
building is used as a case study to demonstrate the practicability of
the proposed approach and highlight potential problems that may
arise. The study starts with the individual optimization of objective
functions focusing on building's characteristics and performance:
energy consumption, retrofit cost, and thermal discomfort hours. Then
a multi-objective optimization model is developed to study the
interaction between these conflicting objectives and assess their
trade-offs. (C) 2014 Elsevier B.V. All rights reserved.

2014


* 462(<-118): Expert energy management of a micro-grid considering wind energy uncertainty

Recently, the use of wind generation has rapidly increased in
micro-grids. Due to the fluctuation of wind power, it is difficult to
schedule wind turbines (WTs) with other distributed energy resources
(DERs). In this paper, we propose an expert energy management system
(EEMS) for optimal operation of WTs and other DERs in an
interconnected micro-grid. The main purpose of the proposed EEMS is to
find the optimal set points of DERs and storage devices, in such a way
that the total operation cost and the net emission are simultaneously
minimized. The EEMS consists of wind power forecasting module, smart
energy storage system (ESS) module and optimization module. For
optimal scheduling of WTs, the power forecasting module determines the
possible available capacity of wind generation in the micro-grid. To
do this, first, an artificial neural network (ANN) is used to forecast
wind speed. Then, the obtaining results are used considering
forecasting uncertainty by the probabilistic concept of confidence
interval. To reduce the fluctuations of wind power generation and
improve the micro-grid performances, a smart energy storage system
(ESS) module is used. For optimal management of the ESS, the
comprehensive mathematical model with practical constraints is
extracted. Finally, an efficient modified Bacterial Foraging
Optimization (MBFO) module is proposed to solve the multi-objective
problem. An interactive fuzzy satisfying method is also used to
simulate the trade-off between the conflicting objectives (cost and
emission). To evaluate the proposed algorithm, the EEMS is applied to
a typical micro-grid which consists of various DERs, smart ESS and
electrical loads. The results show that the EEMS can effectively
coordinate the power generation of DERs and ESS with respect to
economic and environmental considerations. (C) 2014 Elsevier Ltd. All
rights reserved.

2014


* 463(<-131): Intelligent design of waste heat recovery systems using thermoelectric generators and optimization tools

Optimal design of thermal systems that effectively use energy
resources is one of the foremost challenges that researchers almost
confront. Until now, several researches have been made to enhance the
performance of major thermal systems. In this investigation, the
authors try to make a conceptual design to maximize the electricity
demand of Damavand power plant as the biggest thermal system in Middle
East sited in Iran. The idea of designing is laid behind applying a
number of thermoelectric modules within the condenser in order to
recover the waste heat of the thermal systems. Besides, the authors
have developed some intelligent tools to elaborate on the performance
of their proposed model. Firstly, an artificial neural network has
been utilized to estimate the potential power generation of the
thermoelectric modules. At the second step, computational fluid
dynamic solver, FLUENT is used to determine the variation of the
temperature through the length of the thermoelectric module
assembly. Based on the gained results, an intelligent multi-objective
optimization algorithm called Pareto based mutable smart bee is
developed to optimize the properties of the thermoelectric component.

2014


* 465(<-233): Multiobjective Intelligent Energy Management for a Microgrid

In this paper, a generalized formulation for intelligent energy
management of a microgrid is proposed using artificial intelligence
techniques jointly with linear-programming-based multiobjective
optimization. The proposed multiobjective intelligent energy
management aims to minimize the operation cost and the environmental
impact of a microgrid, taking into account its preoperational
variables as future availability of renewable energies and load demand
(LD). An artificial neural network ensemble is developed to predict
24-h-ahead photovoltaic generation and 1-h-ahead wind power generation
and LD. The proposed machine learning is characterized by enhanced
learning model and generalization capability. The efficiency of the
microgrid operation strongly depends on the battery scheduling
process, which cannot be achieved through conventional optimization
formulation. In this paper, a fuzzy logic expert system is used for
battery scheduling. The proposed approach can handle uncertainties
regarding to the fuzzy environment of the overall microgrid operation
and the uncertainty related to the forecasted parameters. The results
show considerable minimization on operation cost and emission level
compared to literature microgrid energy management approaches based on
opportunity charging and Heuristic Flowchart (HF) battery management.

2013



* 466(<-298): Load forecasting framework of electricity consumptions for an Intelligent Energy Management System in the user-side

This work presents an electricity consumption-forecasting framework configured automatically and based on an Adaptative Neural Network Inference System (ANFIS). This framework is aimed to be implemented in industrial plants, such as automotive factories, with the objective of giving support to an Intelligent Energy Management System (IEMS). The forecasting purpose is to support the decision-making (i.e. scheduling workdays, on-off production lines, shift power loads to avoid load peaks, etc.) to optimize and improve economical, environmental and electrical key performance indicators. The base structure algorithm, the ANFIS algorithm, was configured by means of a Multi Objective Genetic Algorithm (MOGA), with the aim of getting an automatic-configuration system modelling. This system was implemented in an independent section of an automotive factory, which was selected for the high randomness of its main loads. The time resolution for forecasting was the quarter hour. Under these challenging conditions, the autonomous configuration, system learning and prognosis were tested with success. (C) 2011 Elsevier Ltd. All rights reserved.

2012


* 467(<-309): Analyzing, controlling, and optimizing Damavand power plant operating parameters using a synchronous parallel shuffling self-organized Pareto strategy and neural network: a survey

In recent decades, analyzing and optimizing thermal systems have become of great interest to researchers. Recently, the engineers concentrated on variant concepts of artificial intelligence such as machine learning, simulation, fuzzy logic, game theory, and evolutionary computing to deal with complicated barriers and obstacles. Artificial intelligence and expert system techniques play an important role for surveying and controlling mechanical systems such as power plants and reservoirs. This is because of their interdisciplinary applications and versatile servicing potential in mathematical modeling of industrial systems. In this article, a new method called synchronous parallel shuffling self-organized Pareto strategy algorithm is presented which synthesizes different artificial techniques, nominally evolutionary computing, swarm intelligence techniques, and time adaptive self-organizing map that apply simultaneously incorporating with a stochastic data sharing behavior. Thereafter, it is applied to verify the optimum operating parameter of Damavand power plant as the biggest constructed power plant in Middle East with the potential of producing about 2300MW electricity sited in Tehran, capital of Iran, as a multi-objective, multi-modal complex problem. It is also proved that implementing the governing equations of power plant leads to a multi-objective problem where some of these objectives are non-linear, non-convex, and multi-modal with different type of real-life engineering constraints. The results confirm the acceptable performance of proposed technique in optimizing the operating parameters of Damavand power plant.

2012




* 470(<- 48): Hydraulic optimization of a double-channel pump's impeller based on multi-objective genetic algorithm

Computational fluid dynamics (CFD) can give a lot of potentially very
useful information for hydraulic optimization design of pumps,
however, it cannot directly state what kind of modification should be
made to improve such hydrodynamic performance. In this paper, a more
convenient and effective approach is proposed by combined using of
CFD, multi-objective genetic algorithm (MOGA) and artificial neural
networks (ANN) for a double-channel pump's impeller, with maximum head
and efficiency set as optimization objectives, four key geometrical
parameters including inlet diameter, outlet diameter, exit width and
midline wrap angle chosen as optimization parameters. Firstly, a
multi-fidelity fitness assignment system in which fitness of impellers
serving as training and comparison samples for ANN is evaluated by
CFD, meanwhile fitness of impellers generated by MOGA is evaluated by
ANN, is established and dramatically reduces the computational
expense. Then, a modified MOGA optimization process, in which
selection is performed independently in two sub-populations according
to two optimization objectives, crossover and mutation is performed
afterword in the merged population, is developed to ensure the global
optimal solution to be found. Finally, Pareto optimal frontier is
found after 500 steps of iterations, and two optimal design schemes
are chosen according to the design requirements. The preliminary and
optimal design schemes are compared, and the comparing results show
that hydraulic performances of both pumps 1 and 2 are improved, with
the head and efficiency of pump 1 increased by 5.7% and 5.2%,
respectively in the design working conditions, meanwhile shaft power
decreased in all working conditions, the head and efficiency of pump 2
increased by 11.7% and 5.9%, respectively while shaft power increased
by 5.5%. Inner flow field analyses also show that the backflow
phenomenon significantly diminishes at the entrance of the optimal
impellers 1 and 2, both the area of vortex and intensity of vortex
decreases in the whole flow channel. This paper provides a promising
tool to solve the hydraulic optimization problem of pumps' impellers.

2015



* 471(<-218): Robust Design of a Reentry Unmanned Space Vehicle by Multifidelity Evolution Control

This paper addresses the preliminary robust design of a small/medium-scale reentry unmanned space vehicle. A hybrid optimization technique is proposed that couples an evolutionary multi-objective algorithm with a direct-transcription method for optimal control problems. Uncertainties on the aerodynamic forces and vehicle mass are integrated in the design process, and the hybrid algorithm searches for geometries that 1) minimize the mean value of the maximum heat flux, 2) maximize the mean value of the maximum achievable distance, and 3) minimize the variance of the maximum heat flux. The evolutionary part handles the system-design parameters of the vehicle and the uncertain functions, while the direct-transcription method generates optimal control profiles for the reentry trajectory of each individual of the population. During the optimization process, artificial neural networks are used to approximate the aerodynamic forces required by the direct-transcription method. The artificial neural networks are trained and updated by means of a multifidelity, evolution-control approach.

2013


* 472(<-223): Three-Dimensional Design and Optimization of a Transonic Rotor in Axial Flow Compressors

This paper presents a 3-D optimization of a moderately loaded transonic compressor rotor by means of a multiobjective optimization system. The latter makes use of a differential evolutionary algorithm in combination with an Artificial Neural Network and a 3D Navier-Stokes solver. Operating it on a cluster of 30 processors enabled the evaluation of the off-design performance and the exploration of a large design space composed of the camber line and spanwise distribution of sweep and chord length. Objectives were an increase of efficiency at unchanged stall margin by controlling the shock waves and off-design performance curve. First designs of single blade rows allowed a better understanding of the impact of the different design parameters. Forward sweep with unchanged camber improved the peak efficiency by only 0.3% with the same stall margin. Backward sweep with an optimized S shaped camber line improved the efficiency by 0.6% at unchanged stall margin. It is explained how the camber line control can introduce the same effect as forward sweep and compensate the expected negative effects of backward sweep. The best results (0.7% increase in efficiency and unchanged stall margin) have been obtained by a stage optimization that allows also a spanwise redistribution of the rotor flow and an increase of loading by extra flow turning. The latter compensates the loading shift induced by the backward sweep in order to reduce the inlet Mach number at the downstream stator hub.

2013


* 473(<-247): A Study on the Design Method to Optimize an Impeller of Centrifugal Compressor

A numerical study was conducted to improve the performance of an impeller of centrifugal compressor. Nine design variables were chosen with constraints. Only meridional contours and blade profile were adjusted. ANN (Artificial Neural Net) was adopted as a main optimization algorithm with PSO (Particle Swarm Optimization) in order to reduce the optimization time. At first, ANN was learned and trained with the design variable sets which were obtained using DOE (Design of Experiment). This ANN was continuously improved its accuracy for each generation of which population was one hundred. New design variable set in each generation was selected using a non-gradient based method of PSO in order to obtain the global optimized result. After 7th generation, the prediction difference of efficiency and pressure ratio between ANN and CFD was less than 0.6%. From more than 1,200 design variable sets, a pareto of efficiency versus pressure ratio was obtained and an optimized result was selected based on the multi-objective function. On this optimized impeller, the efficiency and pressure ratio were improved by 1% and 9.3%, respectively.

2013


* 474(<-310): Optimal Design of a Centrifugal Compressor Impeller Using Evolutionary Algorithms

An optimization study was conducted on a centrifugal compressor. Eight design variables were chosen from the control points for the Bezier curves which widely influenced the geometric variation; four design variables were selected to optimize the flow passage between the hub and the shroud, and other four design variables were used to improve the performance of the impeller blade. As an optimization algorithm, an artificial neural network (ANN) was adopted. Initially, the design of experiments was applied to set up the initial data space of the ANN, which was improved during the optimization process using a genetic algorithm. If a result of the ANN reached a higher level, that result was re-calculated by computational fluid dynamics (CFD) and was applied to develop a new ANN. The prediction difference between the ANN and CFD was consequently less than 1% after the 6th generation. Using this optimization technique, the computational time for the optimization was greatly reduced and the accuracy of the optimization algorithm was increased. The efficiency was improved by 1.4% without losing the pressure ratio, and Pareto-optimal solutions of the efficiency versus the pressure ratio were obtained through the 21st generation.

2012


* 475(<-331): Robust multi-fidelity design of a micro re-entry unmanned space vehicle

This article addresses the preliminary robust design of a small-scale re-entry unmanned space vehicle by means of a hybrid optimization technique. The approach, developed in this article, closely couples an evolutionary multi-objective algorithm with a direct transcription method for optimal control problems. The evolutionary part handles the shape parameters of the vehicle and the uncertain objective functions, while the direct transcription method generates an optimal control profile for the re-entry trajectory. Uncertainties on the aerodynamic forces and characteristics of the thermal protection material are incorporated into the vehicle model, and a Monte-Carlo sampling procedure is used to compute relevant statistical characteristics of the maximum heat flux and internal temperature. Then, the hybrid algorithm searches for geometries that minimize the mean value of the maximum heat flux, the mean value of the maximum internal temperature, and the weighted sum of their variance: the evolutionary part handles the shape parameters of the vehicle and the uncertain functions, while the direct transcription method generates the optimal control profile for the re-entry trajectory of each individual of the population. During the optimization process, artificial neural networks are utilized to approximate the aerodynamic forces required by the optimal control solver. The artificial neural networks are trained and updated by means of a multi-fidelity approach: initially a low-fidelity analytical model, fitted on a waverider type of vehicle, is used to train the neural networks, and through the evolution a mix of analytical and computational fluid dynamic, high-fidelity computations are used to update it. The data obtained by the high-fidelity model progressively become the main source of updates for the neural networks till, near the end of the optimization process, the influence of the data obtained by the analytical model is practically nullified. On the basis of preliminary results, the adopted technique is able to predict achievable performance of the small spacecraft and the requirements in terms of thermal protection materials.

2011



* 477(<-147): Optimizing Rotor Blades with Approximate British Experimental Rotor Programme Tips

This work presents a framework for the optimization of certain aspects
of a British Experimental Rotor Programme-like rotor blade in hover
and forward flight so that maximum performance can be obtained from
the blade. The proposed method employs a high-fidelity, efficient
computational fluid dynamics technique that uses the harmonic balance
method in conjunction with artificial neural networks as metamodels,
and genetic algorithms for optimization. The approach has been
previously demonstrated for the optimization of blade twist in hover
and the optimization of rotor sections in forward flight, transonic
aerofoils design, wing and rotor tip planforms. In this paper, a
parameterization technique was devised for the British Experimental
Rotor Programme-like rotor tip and its parameters were optimized for a
forward flight case. A specific objective function was created using
the initial computational fluid dynamics data and the metamodel was
used for evaluating the objective function during the optimization
using the genetic algorithms. The objective function was adapted to
improve forward flight performance in terms of pitching moment and
torque. The obtained results suggest optima in agreement with
engineering intuition but provide precise information about the shape
of the final lifting surface and its performance. The main
computational cost was associated with the population of the genetic
algorithms database necessary for the metamodel, especially because a
full factorial method was used. The computational time of the
optimization process itself, after the database has been obtained, is
relatively insignificant. Therefore, the computational time was
reduced with the use of the harmonic balance method as opposed to the
time marching method. The novelty in this paper is
two-fold. Optimization methods so far have used simple aerodynamic
models employing direct "calls" to the aerodynamic models within the
optimization loops. Here, the optimization has been decoupled from the
computational fluid dynamics data allowing the use of higher-fidelity
computational fluid dynamics methods based on Navier-Stokes
computational fluid dynamics. This allows a more realistic approach
for more complex geometries such as the British Experimental Rotor
Programme tip. In addition, the harmonic balance method has been used
in the optimization process.

2014


* 478(<-181): Multi-disciplinary design optimization and performance evaluation of a single stage transonic axial compressor

The multidisciplinary design optimization method, which integrates aerodynamic performance and structural stability, was utilized in the development of a single-stage transonic axial compressor. An approximation model was created using artificial neural network for global optimization within given ranges of variables and several design constraints. The genetic algorithm was used for the exploration of the Pareto front to find the maximum objective function value. The final design was chosen after a second stage gradient-based optimization process to improve the accuracy of the optimization. To validate the design procedure, numerical simulations and compressor tests were carried out to evaluate the aerodynamic performance and safety factor of the optimized compressor. Comparison between numerical optimal results and experimental data are well matched. The optimum shape of the compressor blade is obtained and compared to the baseline design. The proposed optimization framework improves the aerodynamic efficiency and the safety factor.

2013


* 479(<-107): Multi-objective optimization of groove casing treatment in a transonic compressor

The paper presents a multi-objective optimization of circumferential casing grooves geometries for the NASA Rotor 37 transonic compressor. The depth normalized by the tip clearance and the width normalized by the tip chord are selected as the design variables. The stall margin and peak efficiency are used as the objective functions. The Latin Hypercube Sampling technique was used to select the sample points in the design space. Based on the numerical results of the sample points, the radial basis function network model of the artificial neural network was constructed. The NSGA-II multi-objective evolutionary algorithm is then employed to search for Pareto-optimal solutions. The leave-one-out cross validation method was also used to evaluate the precision of the radial basis function network model. The results of the optimization show the present method can be effectively used for the design of circumferential casing grooves to take account of the stall margin and efficiency. From the Pareto-optimal solutions, two groove configurations are selected and the internal flow fields are compared with the smooth casing. The effect mechanism of the circumferential casing grooves on the performance of the transonic compressor is discussed by the analysis of the flow in the blade tip region.

2014


* 480(<-367): Reliability-based design optimization of axial compressor using uncertainty model for stall margin

Reliability-based design optimization (RBDO) of the NASA stage 37 axial compressor is performed using an uncertainty model for stall margin in order to guarantee stable operation of the compressor. The main characteristics of RBDO for the axial compressor are summarized as follows: First, the values of mass flow rate and pressure ratio in stall margin calculation are defined as statistical models with normal distributioa for consideration of the uncertainty in stall margin. Second, Monte Carlo Simulation is used in the RBDO process to calculate failure probability of stall margin accurately. Third, an approximation model that is constructed by an artificial neural network is adopted to reduce the time cost Of RBDO. The present method is applied to the NASA stage 37 compressor to improve the reliability of stall margin with both maximized efficiency and minimized weight. The RBDO result is compared with the deterministic optimization (DO) result which does not include an uncertainty model. In the DO case, stall margin is slightly higher than the reference value of the required constraint, but the probability of stall is 43%. This is unacceptable risk for an aircraft engine, which requires absolutely stable operation in flight. However, stall margin obtained in RBDO is 2.7% higher than the reference value, and the probability of success increases to 95% with the improved efficiency and weight. Therefore, RBDO of the axial compressor for aircraft engine can be a reliable design optimization method through consideration of unexpected disturbance of the flow conditions.

2011


* 481(<- 36): Dynamic modeling of exergy efficiency of turboprop engine components using hybrid genetic algorithm-artificial neural networks

Genetic algorithm is utilized to design the optimum initial value of parameters and topology of the artificial neural network which is trained by applying the improved backpropagation algorithm using momentum factor so as to minimize the spent time and effort. In this study, a comprehensive dynamic modeling of turboprop engine components plant is accomplished using hybrid GA (genetic algorithm) ANN (artificial neural networks) strategy. The turboprop engine is equipped with main components such as compressor, combustor, gas turbine and power turbine. Newly derived GA-ANN model takes into account five independent engine variables (i.e., torque, power, gas generator speed, engine mass air flow and fuel flow). These dynamic variables are used as inputs of the ANN while exergy efficiencies of the components are considered as the output parameter of the network. The results show that the hybridization with the genetic algorithm has improved the accuracy even further compared to the trial-and-error case, and the estimated values of exergy efficiencies of the components obtained by the derived model provide a close fit with the reference data. (C) 2015 Elsevier Ltd. All rights reserved.

2015


* 482(<-265): An airfoil optimization technique for wind turbines

Optimization algorithms coupled with computational fluid dynamics are used for wind turbines airfoils design. This differs from the traditional aerospace design process since the lift-to-drag ratio is the most important parameter and the angle of attack is large. Computational fluid dynamics simulations are performed with the incompressible Reynolds-averaged Navier-Stokes equations in steady state using a one equation turbulence model. A detailed validation of the simulations is presented and a computational domain larger than suggested in literature is shown to be necessary. Different approaches to parallelization of the computational code are addressed. Single and multiobjective genetic algorithms are employed and artificial neural networks are used as a surrogate model. The use of artificial neural networks is shown to reduce computational time by almost 50%. (C) 2011 Elsevier Inc. All rights reserved.

2012


* 483(<-443): A grid-enabled asynchronous metamodel-assisted evolutionary algorithm for aerodynamic optimization

A Grid-enabled asynchronous metamodel-assisted evolutionary algorithm is presented and assessed on a number of aerodynamic shape optimization problems. An efficient way of implementing surrogate evaluation models or metamodels (artificial neural networks) in the context of an asynchronous evolutionary algorithm is proposed. The use of metamodels relies on the inexact pre-evaluation technique already successfully applied to synchronous (i.e. generation-based) evolutionary algorithms, which needs to be revisited so as to efficiently cooperate with the asynchronous search method. The so-created asynchronous metamodel-assisted evolutionary algorithm is further enabled for Grid Computing. The Grid deployment of the algorithm relies on three middleware layers: GridWay, Globus Toolkit and Condor. Single- and multi-objective CFD-based designs of isolated airfoils and compressor cascades are handled using the proposed algorithm and the gain in CPU cost is demonstrated.

2009


* 484(<- 80): A Pareto optimal multi-objective optimization for a horizontal axis wind turbine blade airfoil sections utilizing exergy analysis and neural networks

In this study a multi-objective genetic algorithm is utilized to obtain a Pareto optimal set of solutions for geometrical characteristics of airfoil sections for 10-meter blades of a horizontal axis wind turbine. The performance of the airfoil sections during the process of energy conversion is evaluated deploying a 20 incompressible unsteady CFD solver and the second law analysis. Artificial neural networks are trained employing CFD obtained data sets to represent objective functions in an algorithm which implements exergetic performance and integrity characteristics as optimization objectives. The results show that utilizing the second law approach along with Pareto optimality concept leads to a set of precise solutions which represent minimum energy waste, maximum efficiency, and topmost stability. Furthermore, enhanced rotor performance coefficients are observed through a BEM study which compares conventional designs with the second law obtained configurations. Exergy analysis is believed to be an efficient tool in the optimal design of wind turbine blades with the capability of determining the amount of lost opportunities to do useful work (C) 2014 Elsevier Ltd. All rights reserved.

2015



* 485(<-151): Optimization of a novel carbon dioxide cogeneration system using artificial neural network and multi-objective genetic algorithm

In this research study, a combined cycle based on the Brayton power
cycle and the ejector expansion refrigeration cycle is proposed. The
proposed cycle can provide heating, cooling and power
simultaneously. One of the benefits of such a system is to be driven
by low temperature heat sources and using CO2 as working fluid. In
order to enhance the understanding of the current work, a
comprehensive parametric study and exergy analysis are conducted to
determine the effects of the thermodynamic parameters on the system
performance and the exergy destruction rate in the components. The
suggested cycle can save the energy around 46% in comparison with a
system producing cooling, power and hot water separately. On the other
hand, to optimize a system to meet the load requirement, the surface
area of the heat exchangers is determined and optimized. The results
of this section can be used when a compact system is also an objective
function. Along with a comprehensive parametric study and exergy
analysis, a complete optimization study is carried out using a
multi-objective evolutionary based genetic algorithm considering two
different objective functions, heat exchangers size (to be minimized)
and exergy efficiency (to be maximized). The Pareto front of the
optimization problem and a correlation between exergy efficiency and
total heat exchangers length is presented in order to predict the
trend of optimized points. The suggested system can be a promising
combined system for buildings and outland regions. (C) 2013 Elsevier
Ltd. All rights reserved.

2014



* 486(<-  2): Optimization of heat transfer enhancement of nanofluid in a channel with winglet vortex generator

A numerical simulation was performed to investigate the effect of the
nanoparticles, vortex generator and combination of them on heat
transfer and fluid flow characteristics in a rectangular channel using
Commercial Computational Fluid Dynamics Code FLUENT. Euler-Lagrangian
approach was utilized in the numerical modeling of the nanofluids. The
fluid was considered as a continuous phase and the heat and flow
fields were analyzed by solving Navier-Stokes and energy equations and
the nanoparticles were simulated as a discrete phase in a Lagrangian
frame. The thermal and hydraulic performances of the channel were
investigated at different nanoparticle concentration, shape and angle
of attack of the vortex generator. According to the results, the
Nusselt number increases by raising the nanoparticles concentration
and adding nanoparticles is more effective than placing VG from
thermal point of view in the range of study. Using a combination of
these heat transfer enhancers maximizes the thermal performance of the
channel as well as the flow resistance. So, in order to achieve the
best thermal-hydraulic performance, combination of the computational
fluid dynamics analyses, artificial neural networks and
multi-objective genetic algorithm was used to determine the optimal
values of these parameters. Finally, a set of optimal solutions as
well as the best shape, angle of attack of the VG and nanoparticles
volume fraction was obtained. (C) 2015 Elsevier Ltd. All rights
reserved.

2015


* 487(<- 49): Optimization of shape and angle of attack of winglet vortex generator in a rectangular channel for heat transfer enhancement

A three dimensional numerical simulation was performed to study the effect of the shape and the angle of attack of the winglet vortex generator on the heat transfer and fluid flow characteristics in a rectangular heat sink. The mass, momentum and energy equations were solved using finite volume method by considering the steady state, laminar and incompressible fluid flow. The average and local Nusselt number and pressure drop were investigated in the presence of the vortex generators at different shapes and angles of attack. To achieve a maximum heat transfer enhancement and a minimum pressure drop, the optimal values of these parameters were calculated using the Pareto optimal strategy. For this purpose, computational fluid dynamics analyses, multi-objective genetic algorithm and artificial neural networks were combined together and used in the optimization process. Finally, the optimal values of these parameters were presented. (C) 2015 Elsevier Ltd. All rights reserved.

2015


* 488(<-115): Modelling and optimisation of laser shock peening using an integrated simulated annealing-based method

Laser shock peening is an innovative surface treatment technique, which has been successfully applied to improve fatigue performance of metallic components. Laser shock peening improves the surface morphology and microstructure of the material. In this paper, three Nd3+:YAG laser process parameters (voltage, focus position and pulse duration) are varied in an experiment, in order to determine the optimal process parameters that could simultaneously meet the specifications for seven correlated responses of processed Nimonic 263 sheets. The modelling and optimisation of a process were performed using the advanced, problem-independent method. First, responses are expressed using Taguchi's quality loss function, followed by the application of multivariate statistical methods to uncorrelate and synthesise them into a single performance measure. Then, artificial neural networks are used to build the process model, and simulated annealing was utilised to find the optimal process parameters setting in a global continual space of solutions. Effectiveness of the proposed method in the development of a robust laser shock peening was proved in comparison to several commonly used approaches from the literature, resulting in the highest process performance measure, the most favourable response values and the corresponding process parameters optimum. Besides the improved surface characteristics, the optimised laser shock peening (LSP) showed an improvement in terms of microhardness and formation of favourable microstructural transformations.

2014


* 489(<-358): Multi-response design of Nd:YAG laser drilling of Ni-based superalloy sheets using Taguchi's quality loss function, multivariate statistical methods and artificial intelligence

This paper presents a hybrid design strategy for the determination of the optimum laser drilling parameters which simultaneously meets the requirements for seven quality characteristics (responses) of the holes produced during pulsed Nd:YAG laser drilling of a thin sheet of nickel-based superalloy Nimonic 263. The process was designed using two approaches based on the experimental data. In the first approach, the quality losses of seven correlated responses were uncorrelated into a set of components using the principal component analysis; then the grey relational analysis was applied to synthesise components into a synthetic performance measure. Since this approach considered only parameter values used in the experiment, the second approach was developed to find the global optimal parameters solution using an artificial neural network to model the relation between parameters and a synthetic performance measure, and a genetic algorithm to perform a search for the global optimum in a continual multidimensional space. The analysis of the application indicated that the proposed approaches gave a better result, in terms of the optimal parameter settings that yield the maximal synthetic performance measure, than several commonly used methods for multi-response process parameters design. The results demonstrated that the robust Nd:YAG laser drilling of Ni-based superalloy sheets was designed with respect to the requirements for seven quality characteristics of the drilled holes, by using the proposed strategy.

2011


* 490(<-137): A Novel Design Approach to X-Band Minkowski Reflectarray Antennas using the Full-Wave EM Simulation-based Complete Neural Model with a Hybrid GA-NM Algorithm

In this work, a novel multi-objective design optimization procedure is presented for the Minkowski Reflectarray RA s using a complete 3-D CST Microwave Studio MWS- based Multilayer Perceptron Neural Network MLP NN model including the substrate constant Er with a hybrid Genetic GA and Nelder-Mead NM algorithm. The MLP NN model provides an accurate and fast model and establishes the reflection phase of a unit Minkowski RA element as a continuous function within the input domain including the substrate 1 <= epsilon(r) <= 6; 0.5 mm <= h <= 3 mm in the frequency between 8 GHz <= f <= 12 GHz. This design procedure enables a designer to obtain not only the most optimum Minkowski RA design all throughout the X-band, at the same time the optimum Minkowski RAs on the selected substrates. Moreover a design of a fully optimized X-band 15 x 15 Minkowski RA antenna is given as a worked example together with the tolerance analysis and its performance is also compared with those of the optimized RA s on the selected traditional substrates. Finally it may be concluded that the presented robust and systematic multi-objective design procedure is conveniently applied to the Microstrip Reflectarray RAs constructed from the advanced patches.

2014


* 491(<-152): Multi-objective aero acoustic optimization of rear end in a simplified car model by using hybrid Robust Parameter Design, Artificial Neural Networks and Genetic Algorithm methods

In this paper, optimization of rear end of a simplified car model is performed considering aerodynamic and acoustic objectives. Slant angle, rear box angle, boat tail angle, and rear box length are considered as main variables of the rear end. For numerical simulation of flow around the model and studying aerodynamic noise, realizable turbulent model and broad band noise model are used, respectively. Simulation results are validated by the experimental results reported in the literature. To reduce number of simulations to reach optimum values of parameters, Taguchi method has been used. The results of Taguchi are in good agreement with simulation results. Then, the results of Taguchi have been used to obtain a relation between parameters and objectives employing Artificial Neural Networks. Optimization of the model has been conducted by the Neural Network and Multi Objective Genetic Algorithm methods. Finally, flow around the optimized model has been studied by numerical simulation and results have been reported. (C) 2013 Elsevier Ltd. All rights reserved.

2014


* 492(<-355): A hybrid of back propagation neural network and genetic algorithm for optimization of injection molding process parameters

This paper presents a hybrid optimization method for optimizing the process parameters during plastic injection molding (PIM). This proposed method combines a back propagation (BP) neural network method with an intelligence global optimization algorithm, i.e. genetic algorithm (GA). A multi-objective optimization model is established to optimize the process parameters during PIM on the basis of the finite element simulation software Moldflow, Orthogonal experiment method, BP neural network as well as Genetic algorithm. Optimization goals and design variables (process parameters during PIM) are specified by the requirement of manufacture. A BP artificial neural network model is developed to obtain the mathematical relationship between the optimization goals and process parameters. Genetic algorithm is applied to optimize the process parameters that would result in optimal solution of the optimization goals. A case study of a plastic article is presented. Warpage as well as clamp force during PIM are investigated as the optimization objectives. Mold temperature, melt temperature, packing pressure, packing time and cooling time are considered to be the design variables. The case study demonstrates that the proposed optimization method can adjust the process parameters accurately and effectively to satisfy the demand of real manufacture. (C) 2011 Elsevier Ltd. All rights reserved.

2011


* 493(<-287): Process optimization of the serial-parallel hybrid polishing machine tool based on artificial neural network and genetic algorithm

A kind of serial-parallel hybrid polishing machine tool based on the elastic polishing theory is developed and applied to finish mould surface with using bound abrasives. It mainly consists of parallel mechanism of three dimensional moving platform, serial rotational mechanism of two degrees of freedom and the elastic polishing tool system. The active compliant control and passive conformity of polishing tool are provided by a pneumatic servo system and a spring, respectively. Considering the contradiction between the machining quality and efficiency, the optimization model of process parameters is found according to different machining requirements, namely single objective optimization and multi-objective optimization, which provide a choice of parameters as a basis for the operators in practice. Many polishing experiments are conducted to collect the data samples. The genetic algorithm integrated with artificial neural network is used for researching for the optimal process parameters in term of the various optimization objectives. This research also lays the foundation for further establishing polishing expert system.

2012


* 494(<-291): Multi-objective process parameter optimization for energy saving in injection molding process

This paper deals with a multi-objective parameter optimization framework for energy saving in injection molding process. It combines an experimental design by Taguchi's method, a process analysis by analysis of variance (ANOVA), a process modeling algorithm by artificial neural network (ANN), and a multi-objective parameter optimization algorithm by genetic algorithm (GA)-based lexicographic method. Local and global Pareto analyses show the trade-off between product quality and energy consumption. The implementation of the proposed framework can reduce the energy consumption significantly in laboratory scale tests, and at the same time, the product quality can meet the pre-determined requirements.

2012




* 496(<-564): Single and multi objective optimization for injection molding using numerical simulation with surrogate models and genetic algorithms

The objective of this study is to develop an integrated computer-aided engineering (CAE) optimization system that can quickly and intelligently determine the optimal process conditions for injection molding. This study employs support vector regression (SVR) to establish the surrogate model based on executions of three-dimensional (3D) simulation for a selected dataset using the latin hypercube sampling (LHS) technique. Once the surrogate model can satisfactorily capture the characteristics of simulations with much less computing resources, a hybrid optimization genetic algorithm (GA) or a multi-objective optimization GA is then used to evaluate the surrogate model to search the global optimal solutions for the single or multiple objectives, respectively. The performance and capabilities of other surrogate modeling approaches, such as polynomial regression (PR) and artificial neural network (ANN), are also investigated in terms of accuracy, robustness, efficiency, and requirements for training samples. Experimental validations and applications of this work for process optimization of a special box mold and a precision optical lens are presented.

2006





* 497(<- 30): An evolutionary approach with surrogate models and network science concepts to design optical networks

Physical topology design of optical networks is frequently
accomplished by using evolutionary approaches. However, fitness
evaluation for this type of problems is time consuming and the overall
optimization process presents a huge execution time. In this paper we
propose a new method that uses a multi-objective evolutionary approach
to handle the design of all-optical networks. We focused on the
simultaneous optimization of the network topology and the device
specifications in order to minimize both the capital expenditure of
the network and the network performance. Our method uses surrogate
models to accelerate the fitness evaluation and a novel network
generative model based on preferential attachment to generate the
seeds for the evolutionary process. Our approach can provide high
quality solutions with a very small execution time when compared to
the previous approaches. In order to assess our proposal we performed
a set of simulations aiming to analyze the convergence ability and the
diversity of the generated solutions for scenarios considering uniform
and non-uniform traffic matrices. From our results, we obtained an
evolutionary approach that presents better solutions than previous
proposals for all analyzed scenarios. Our proposal presents an
execution time that is up to 84% and 88% lower than the execution time
needed by the previous approaches for uniform and non-uniform traffic,
respectively. (C) 2015 Elsevier Ltd. All rights reserved.

2015



* 498(<-648): Optimization of a microwave amplifier using neural performance data sheets with genetic algorithms

In this work, the neural performance data sheets of the transistor are employed to determine the feasible design target space in the optimization of a microwave amplifier. In order to obtain these data sheets the ANN model of the active device is utilized to approximate the small-signal [S] and noise [N] parameter functions in the operation domain. Inputting of these characterization parameters into the performance characterization of the device results in the triplet of gain G(T), noise F, and input VSWR V-i and its source (Z(S)) and load (Z(L)) termination functions in the operation domain, from which the neural performance data sheets can be obtained. The genetic algorithms with the binary (BGA) and decimal (CPGA) numbers are utilized in the multi-objective optimization process for the global minimum of the objective function which is expressed as a function only gain of a matching circuit, in the negative exponential form to ensure the rapid convergence. Here optimization of a microwave amplifier with the Pi - type matching circuits is given as a worked example and its resulted performance ingredients are compared with the design targets.

2003


* 499(<-671): Optimising a car chassis

This paper presents a method to optimise a car chassis fitted either with passive or active suspensions. Provided that a full vehicle model is available for accurate simulations of many different driving situations (J-turn, lane-change, power-on/power-off on even/rough, dry/wet roads), the method allows to tune the parameters of the chassis system (suspension elastokinematics, stiffnesses, dampings, actuator gains, tyre pressures...) in order to achieve the desired dynamic behaviour of the car in all of the considered driving situations. According with the Global Approximation approach, the original physical car model is substituted by another purely numerical mathematical model (backpropagating Artificial Neural Network). This reduces the simulation time dramatically and enables the optimisation process to come to successful results. The computation of the Pareto-optimal set is performed by using Generic Algorithms. The method is validated by optimising the parameters of the suspension system of an actual car.

1999




* 504(<-656): Integrated genetic algorithm-artificial neural network strategy for modeling important multiphase-flow characteristics

Numerous investigations have shown that artificial neural networks
(ANNs) can be successful for correlating experimental data sets of
macroscopic multiphase-flow characteristics, e.g., holdup, pressure
drop, and interfacial mass transfer. The approach proved its worth
especially when rigorous fluid mechanics treatment based on the
solution of first-principle equations is not tractable. One perennial
obstacle facing correlations is the choice of a low-dimensionality
input vector containing the most expressive dimensionless independent
variables allowing the best correlation of the dependent output
variable. Because no clue is known in advance, one has recourse to a
laborious, often inefficient, and nonsystematic trial-and-error
procedure to identify from a broad reservoir of possible candidates,
the most relevant combination of ANN input dimensionless
variables. The combinatorial nature of the problem renders the
determination of the best combination, especially for multiphase
flows, computationally difficult because of the large scale of the
search space of combinations. A methodology is devised in this work to
cope with this computational complexity by illustrating the potential
of genetic algorithms (GAs) to efficiently identify the elite ANN
input combination required for the prediction of desired
characteristics. The multiobjective function to be minimized is a
composite criterion that includes ANN prediction errors on both
learning and generalization data sets, as well as a penalty function
that embeds phenomenological rules accounting for ANN model likelihood
and adherence to behavior dictated by the process physics. The proof
of concept of the integrated GA-ANN methodology was illustrated using
a comprehensive database of experimental total liquid holdup for
countercurrent gas-liquid flows in randomly packed towers for
extracting the best liquid hold-up correlation.

2002




* 506(<-101): Electrode design optimization of lithium secondary batteries to enhance adhesion and deformation capabilities

Safety, performance and lifetime of LSB (lithium secondary batteries)
are affected by the adhesion of the active material to the electrode
substance, and to the electrode deformation and the spring back limit
in the electrode manufacturing process. This study explores the
optimization process using decision tree analysis, an ANN (artificial
neural network), and a multi-objective genetic algorithm. In the
electrode design optimization, the objectives are to maximize the
adhesion and to minimize the electrode deformation subjected to the
allowable limit on the spring-back. Experimental data for use in
design analysis and optimization is obtained via a measurement
test. The decision tree analysis is first performed to extract major,
effective parameters sensitive to adhesion force, electrode
deformation and spring-back. The ANN-based approximate meta-models are
then established for function approximations. The ANN-based causality
analysis is further explored to determine dominant design variables
for each of three design requirements for the optimization. A
multi-objective optimization is finally conducted using ANN-based
approximate meta-models. An optimized solution obtained from the
numerical optimization process is compared with experimental data to
verify the actual performance of the LSB in terms of physical and
electro-chemical properties. (C) 2014 Elsevier Ltd. All rights
reserved.

2014



* 507(<-422): Multiobjective optimization of building design using TRNSYS simulations, genetic algorithm, and Artificial Neural Network

Building optimization involving multiple objectives is generally an
extremely time-consuming process. The GAINN approach presented in this
study first uses a simulation-based Artificial Neural Network (ANN) to
characterize building behaviour, and then combines this ANN with a
multiobjective Genetic Algorithm (NSGA-II) for optimization. The
methodology has been used in the current study for the optimization of
thermal comfort and energy consumption in a residential house. Results
of ANN training and validation are first discussed. Two optimizations
were then conducted taking variables from HVAC system settings,
thermostat programming, and passive solar design. By integrating ANN
into optimization the total simulation time was considerably reduced
compared to classical optimization methodology. Results of the
optimizations showed significant reduction in terms of energy
consumption as well as improvement in thermal comfort. Finally, thanks
to the multiobjective approach, dozens of potential designs were
revealed, with a wide range of trade-offs between thermal comfort and
energy consumption. (C) 2009 Elsevier Ltd. All rights reserved.

2010



* 508(<-501): Multidisciplinary design and optimization methodologies in electronics packaging: State-of-the-art review

Electronics packaging design is a process that requires optimized solutions based on multidisciplinary design trade-offs, which usually have complex relationships among multiple design variables. Required numerical analyses combining electrical, thermal, and thermomechanical, among others, have made the multidisciplinary design and optimization process more challenging because of their time-intensive modeling and computation. In this paper, a state-of-the-art review of recent multidisciplinary design and optimization methodologies in electronics packaging is presented. The reported methodologies are divided into three groups: (1) integrated multidisciplinary computer aided design (CAD) environment, (2) semi-automated design optimization techniques, and (3) automated component placement techniques. In the first group, multidisciplinary design and optimization are carried out using interactive CAD environment software. The electronics packaging designer inputs data and makes decisions, while the CAD software provides a comprehensive multidisciplinary modeling and simulation environment. In the second group, using semi-automated design optimization methodologies, various objectives are optimized simultaneously mainly based on package configurations (dimensions), material properties, and operating conditions. In the third group, optimal placement of heat generating components is performed automatically based on multiple requirements. In recent years, methodologies using (1) detailed numerical analysis models directly connected to optimization algorithms, (2) design of experiments (DoE), and (3) artificial neural networks (ANNs) have been proposed as new trends in this field. These methodologies have led to significant improvement in design optimization capabilities, while they require intensive computational effort. Advantages as well as disadvantages of these methods are discussed.

2008




* 510(<-415): Design, analysis, and stiffness optimization of a three degree of freedom parallel manipulator

This paper proposed a novel three degree of freedom (DOF) parallel
manipulator two translations and one rotation. The mobility study and
inverse kinematic analysis are conducted, and a CAD model is presented
showing the design features. The optimization techniques based on
artificial intelligence approaches are investigated to improve the
system stiffness of the proposed 3-DOF parallel manipulator. Genetic
algorithms and artificial neural networks are implemented as the
intelligent optimization methods for the stiffness synthesis. The mean
value and the standard deviation of the global stiffness distribution
are proposed as the design indices. Both the single objective and
multi-objective optimization issues are addressed. The effectiveness
of this methodology is validated with Matlab.

2010


* 511(<-419): Design optimization of a spatial six degree-of-freedom parallel manipulator based on artificial intelligence approaches

Optimizing the system stiffness and dexterity of parallel manipulators by adjusting the geometrical parameters can be a difficult and time-consuming endeavor, especially when the variables are diverse and the objective functions are excessively complex. However, optimization techniques that are based on artificial intelligence approaches can be an effective solution for addressing this issue. Accordingly, this paper describes the implementation of genetic algorithms and artificial neural networks as an intelligent optimization tool for the dimensional synthesis of the spatial six degree-of-freedom (DOF) parallel manipulator. The objective functions of system stiffness and dexterity are derived according to kinematic analysis of the parallel mechanism. In particular, the neural network-based standard backpropagation learning algorithm and the Levenberg-Marquardt algorithm are utilized to approximate the analytical solutions of system stiffness and dexterity. Subsequently, genetic algorithms are derived from the objective functions described by the trained neural networks, which model various performance solutions. The multi-objective optimization (MOO) of performance indices is established by searching the Pareto-optimal frontier sets in the solution space. Consequently, the effectiveness of this method is validated by simulation. (C) 2009 Elsevier Ltd. All rights reserved.

2010


* 512(<-185): A combustor liner cooling system design methodology based on a fluid/structure approach

The paper presents a multi-disciplinary multi-objective design optimization methodology of a combustion chamber effusion cooling system. The optimizer drives an Artificial Neural Network and a Manufacturing Time Model in a repeated analysis scheme in order to increase the combustor liner LCF life and to reduce the liner cooling system manufacturing time, simultaneously. The ANN is trained with a set of fluid/structure/lifing simulations arranged in a three-levels dataset based on a properly designed DOE approach. The CFD simulations are carried out with a reliable and robust in-house developed three-dimensional high resolution reactive viscous flow solver, accounting for conjugate heat transfer approach; the liner structural analysis is performed with a standard FEM code while the liner life assessments are obtained through an in-house developed software operating on the temperature/stress fields. Results demonstrate the validity of the overall approach in a five-dimensional state space with truly moderate computational costs. (C) 2013 Elsevier Ltd. All rights reserved.

2013




* 513(<-431): A multi-objective design optimization strategy as applied to pre-mixed pre-vaporized injection systems for low emission combustors

This paper presents a multi-objective optimization procedure as
applied to the design of the injection system of a Lean Pre-mixed
Pre-vaporized combustion chamber. The optimizer drives an Artificial
Neural Network in a repeated analysis scheme in order to
simultaneously reduce NOX and CO pollutant emissions. The ANN is
trained with a few three-dimensional high resolution reactive viscous
flow simulations, carried out with a reliable and robust CFD
code. Results, obtained in a four-dimensional state space, demonstrate
the validity of the overall procedure with truly moderate
computational costs.

2010


* 514(<-133): Simulation, analysis and optimal design of fuel tank of a locomotive

In this paper, fuel tank of the locomotive ER 24 has been
studied. Firstly the behavior of fuel and air during the braking time
has been investigated by using a two-phase model. Then, the
distribution of pressure on the surface of baffles caused by sloshing
has been extracted. Also, the fuel tank has been modeled and analyzed
using Finite Element Method (FEM) considering loading conditions
suggested by the DIN EN 12663 standard and real boundary
conditions. In each loading condition, high stressed areas have been
identified. By comparing the distribution of pressure caused by
sloshing phenomena and suggested loading conditions, optimization of
the tank has been taken into consideration. Moreover, internal baffles
have been investigated and by modifying their geometric properties,
search of the design space has been done to reach the optimal
tank. Then, in order to reduce the mass and manufacturing cost of the
fuel tank, Non-dominated Sorting Genetic Algorithm (NSGA-II) and
Artificial Neural Networks (ANNs) have been employed. It is shown that
compared to the primary design, the optimized fuel tank not only
provides the safety conditions, but also reduces mass and
manufacturing cost by %39 and %73, respectively.

2014



* 515(<-162): Application of imperialist competitive algorithm and neural networks to optimise the volume fraction of three-parameter functionally graded beams

This paper deals with optimisation of three-parameter power-law
distribution of functionally graded (FG) beam. The main goal of the
optimisation problem is to determine the optimum volume fraction
relation for maximising the first natural frequency of FG beam. Since
the search space is large, the optimisation processes become very
complicated and too time consuming. Thus, a novel meta-heuristic
called imperialist competitive algorithm (ICA), which is a
socio-politically motivated global search strategy is applied to find
the optimal solution. Applying the proposed algorithm to some of
benchmark cost functions, it shows its ability in dealing with
different types of optimisation problems. A proper and accurate
artificial neural network (ANN) is trained by training data sets
obtained from generalised differential quadrature method and then is
applied as the objective function in ICA. The ANN improves the speed
of optimisation process by a considerable amount by reproducing the
fundamental natural frequency of the structure. The performance of ICA
is evaluated in comparison with other nature-inspired technique
genetic algorithm. Comparison shows the success of combination of ANN
and ICA for design of material profile of beam. Finally the optimised
material profile for the optimisation problem is presented.

2014




* 516(<-199): Hybridization of multi-objective evolutionary algorithms and artificial neural networks for optimizing the performance of electrical drives

Performance optimization of electrical drives implies a lot of degrees of freedom in the variation of design parameters, which in turn makes the process overly complex and sometimes impossible to handle for classical analytical optimization approaches. This, and the fact that multiple non-independent design parameter have to be optimized synchronously, makes a soft computing approach based on multi-objective evolutionary algorithms (MOEAs) a feasible alternative. In this paper, we describe the application of the well known Non-dominated Sorting Genetic Algorithm II (NSGA-II) in order to obtain high-quality Pareto-optimal solutions for three optimization scenarios. The nature of these scenarios requires the usage of fitness evaluation functions that rely on very time-intensive finite element (FE) simulations. The key and novel aspect of our optimization procedure is the on-the-fly automated creation of highly accurate and stable surrogate fitness functions based on artificial neural networks (ANNs). We employ these surrogate fitness functions in the middle and end parts of the NSGA-II run (-> hybridization) in order to significantly reduce the very high computational effort required by the optimization process. The results show that by using this hybrid optimization procedure, the computation time of a single optimization run can be reduced by 46-72% while achieving Pareto-optimal solution sets with similar, or even slightly better, quality as those obtained when conducting NSGA-II runs that use FE simulations over the whole run-time of the optimization process. (C) 2013 Elsevier Ltd. All rights reserved.

2013



* 517(<-200): Decomposition-based multi-objective differential evolution particle swarm optimization for the design of a tubular permanent magnet linear synchronous motor

This article proposes a decomposition-based multi-objective differential evolution particle swarm optimization (DMDEPSO) algorithm for the design of a tubular permanent magnet linear synchronous motor (TPMLSM) which takes into account multiple conflicting objectives. In the optimization process, the objectives are evaluated by an artificial neural network response surface (ANNRS), which is trained by the samples of the TPMSLM whose performances are calculated by finite element analysis (FEA). DMDEPSO which hybridizes differential evolution (DE) and particle swarm optimization (PSO) together, first decomposes the multi-objective optimization problem into a number of single-objective optimization subproblems, each of which is associated with a Pareto optimal solution, and then optimizes these subproblems simultaneously. PSO updates the position of each particle (solution) according to the best information about itself and its neighbourhood. If any particle stagnates continuously, DE relocates its position by using two different particles randomly selected from the whole swarm. Finally, based on the DMDEPSO, optimization is gradually carried out to maximize the thrust of TPMLSM and minimize the ripple, permanent magnet volume, and winding volume simultaneously. The result shows that the optimized TPMLSM meets or exceeds the performance requirements. In addition, comparisons with chosen algorithms illustrate the effectiveness of DMDEPSO to find the Pareto optimal solutions for the TPMLSM optimization problem.

2013




* 520(<-391): Optimum Design of Tubular Permanent-Magnet Motors for Thrust Characteristics Improvement by Combined Taguchi-Neural Network Approach

Although tubular permanent-magnet motors have advantages such as remarkable force capability and high efficiency due to lack of end winding, they suffer from high thrust force ripple. This paper presents the use of Taguchi method and artificial neural network (ANN) for shape optimization of axially magnetized tubular linear permanent-magnet (TLPM) motors. A multiobjective design optimization is presented to improve force ripple, developed thrust, and permanent-magnet volume simultaneously. The iron pole-piece slotting technique is used and its design parameters are optimized to minimize the motor's force pulsation. To obtain optimal configuration using this technique, four design variables are selected and their approximate optimum values are determined by the Taguchi method using analysis of means (ANOM). In the next step, two more influential parameters are selected by analysis of variance (ANOVA) and their accurate optimum values are obtained by a trained ANN. Finite-element analysis (FEA) is used to appraise the performance of the motor in different experiments of the Taguchi method and for training the ANN. The results show that force pulsation of the optimized motor is greatly reduced while there is small drop in the motor thrust.

2010



* 521(<-545): A neural networks inversion-based algorithm for multiobjective design of a high-field superconducting dipole magnet

In this paper, an original algorithm to solve multiobjective design problems, which makes use of a neural network (NN) inversion method, is presented. The proposed approach allows us to explore the solutions directly in the objectives space, rather than in the parameters space, with a great saving of computation time in the reconstruction of the Pareto front. A multilayer perceptron NN is first trained to solve the analysis design problem. The inversion of the neural model allows us to obtain the design parameters, starting from the desired requirements on all the conflicting multiple objectives. The performance of the method is demonstrated by its application to the design of a high-field superconducting dipole magnet, where a tradeoff between the superconductors volumes is required in order to obtain a prescribed magnetic field value in the dipole axis.

2007


* 522(<-  1): Minimum-weight design for three dimensional woven composite stiffened panels using neural networks and genetic algorithms

The paper describes a modeling strategy for multi-scale analysis and optimization of stiffened panels, made of three-dimensional woven composites. Artificial neural network techniques are utilized to generate an approximate response for the optimum structural design in order to increase efficiency and applicability. The artificial neural networks are integrated with genetic algorithms to optimize mixed discrete-continuous design variables for the three dimensional woven composite structures. The proposed procedure is then applied to the multi-objective optimal design of a stiffened panel subject to buckling and post-buckling requirements. (C) 2015 Elsevier Ltd. All rights reserved.

2015




* 524(<-177): Multi-objective optimization of a building envelope for thermal performance using genetic algorithms and artificial neural network

The objective of this paper is to present a method to optimize the
equivalent thermophysical properties of the external walls (thermal
conductivity k(wall) and volumetric specific heat (rho c)(wall)) of a
dwelling in order to improve its thermal efficiency. Classical
optimization involves several dynamic yearly thermal simulations,
which are commonly quite time consuming. To reduce the computational
requirements, we have adopted a methodology that couples an artificial
neural network and the genetic algorithm NSGA-H. This optimization
technique has been applied to a dwelling for two French climates,
Nancy (continental) and Nice (Mediterranean). We have chosen to
characterize the energy performance of the dwelling with two criteria,
which are the optimization targets: the annual energy consumption
Q(TOT) and the summer comfort degree I-sum. First, using a design of
experiments, we have quantified and analyzed the impact of the
variables k(wall) and (rho c)(wall) on the objectives Q(TOT) and
I-sum, depending on the climate. Then, the optimal Pareto fronts
obtained from the optimization are presented and analyzed. The optimal
solutions are compared to those from mono-objective optimization by
using an aggregative method and a constraint problem in GenOpt. The
comparison clearly shows the importance of performing multi-objective
optimization. (C) 2013 Elsevier B.V. All rights reserved.

2013




* 526(<-186): Generalized evolutionary optimum design of fiber-reinforced tire belt structure

This paper deals with the multi-objective optimization of tire
reinforcement structures such as the tread belt and the carcass
path. The multi-objective functions are defined in terms of the
discrete-type design variables and approximated by artificial neutral
network, and the sensitivity analyses of these functions are replaced
with the iterative genetic evolution. The multi-objective optimization
algorithm introduced in this paper is not only highly
CPU-time-efficient but it can also be applicable to other
multi-objective optimization problems in which the objective function,
the design variables and the constraints are not continuous but
discrete. Through the illustrative numerical experiments, the
fiber-reinforced tire belt structure is optimally tailored. The
proposed multi-objective optimization algorithm is not limited to the
tire reinforcement structure, but it can be applicable to the
generalized multi-objective structural optimization problems in
various engineering applications.

2013



* 528(<-272): Pareto-optimal analysis of Zn-coated Fe in the presence of dislocations using genetic algorithms

To design a coating that will absorb maximum energy prior to failure
with minimum deformation, the shearing process of polycrystalline Zn
coated Fe is simulated in the presence of dislocations, using
molecular dynamics. The results fed to an Evolutionary Neural Network
generated the meta-models of objective functions required in the
subsequent Pareto-optimization task using a Multi-objective Genetic
Algorithm. Similar calculations conducted for single crystals, and
also in the absence of dislocations, are compared and analyzed. (C)
2012 Elsevier B. V. All rights reserved.

2012




* 529(<-450): Analyzing Fe-Zn system using molecular dynamics, evolutionary neural nets and multi-objective genetic algorithms

Failure behavior of Zn coated Fe is simulated through molecular
dynamics (MD) and the energy absorbed at the onset of failure along
with the corresponding strain of the Zn lattice are computed for
different levels of applied shear rate. temperature and
thickness. Data-driven models are constructed by feeding the MD
results to an evolutionary neural network. The outputs of these neural
networks are utilized to carry out a multi-objective optimization
through genetic algorithms, where the best possible tradeoffs between
two conflicting requirements, minimum deformation and maximum energy
absorption at the onset of failure, are determined by constructing a
Pareto frontier. (C) 2009 Elsevier B.V. All rights reserved.

2009



* 530(<-582): Multiobjective RBFNNs designer for function approximation: An application for mineral reduction

Radial Basis Function Neural Networks (RBFNNs) are well known because,
among other applications, they present a good performance when
approximating functions. The function approximation problem arises in
the construction of a control system to optimize the process of the
mineral reduction. In order to regulate the temperature of the ovens
and other parameters, it is necessary a module to predict the final
concentration of mineral that will be obtained from the source
materials. This module can be formed by an RBFNN that predicts the
output and by the algorithm that designs the RBFNN dynamically as more
data is obtained. The design of RBFNNs is a very complex task where
many parameters have to be determined, therefore, a genetic algorithm
that determines all of them has been developed. This algorithm
provides satisfactory results since the networks it generates are able
to predict quite precisely the final concentration of mineral.

2006


* 531(<-149): Optimal design of floating substructures for spar-type wind turbine systems

The platform and floating structure of spar type offshore wind turbine systems should be designed in order for the 6-DOF motions to be minimized, considering diverse loading environments such as the ocean wave, wind, and current conditions. The objective of this study is to optimally design the platform and substructure of a 3MW spar type wind turbine system with the maximum postural stability in 6-DOF motions as well as the minimum material cost. Therefore, design variables of the platform and substructure were first determined and then optimized by a hydrodynamic analysis. For the hydrodynamic analysis, the body weight of the system was considered, and the ocean wave conditions were quantified to the wave forces using the Morison's equation. Moreover, the minimal number of computation analysis models was generated by the Design of Experiments (DOE), and the design variables of the platform and substructure were finally optimized by using a genetic algorithm with a neural network approximation.

2014


* 532(<-441): Optimization of Vertical Roller Mill by Using Artificial Neural Networks

The vertical roller mill is important for machine grinding and mixing various crude materials in the process of producing Portland cement. A vertical roller mill is subjected to cyclic bending stress because of the roller load. Because of the cyclic bending stress, only 4?106 ?8?106 cycles are achieved instead of 4?107 cycles. The stress also causes fractures at the edge of grinding path of the outer roller. The expenses incurred in repairing the grinding path amounts to 30% of the total maintenance cost. Therefore, it is desirable to redesign the vertical roller mill in order to reduce the expenses incurred in repairing the roller. In this study, artificial neural networks (ANNs) were applied in order to solve the multiobjective optimization problem for vertical roller mills by using the function approximation ability of ANNs. To learn and generalize ANNs, the maximum and minimum stresses were estimated from the results of the finite-element analysis of a vertical roller mill. Thus, ANNs could be applied to solve the multiobjective optimization problem.

2010





* 533(<-452): Probabilistic Evaluation of Optimal Location of Surge Arresters on EHV and UHV Networks Due to Switching and Lightning Surges

Switching surges are of primary importance in insulation coordination
of extremely high voltage and ultra-high voltage networks. However, in
regions of high lightning activity or high ground resistance
insulation design, preferably, should be based on the risk of failure
caused by lightning and switching surges and the probability of line
outage, a combination of lightning and switching flashover rates
(SSFOR). This paper describes an effective installation of
transmission line arresters (TLAs) to obtain a better protection
scheme (i.e., minimizing global risk to the network). As a
consequence, protection costs are reduced in accordance with the costs
of elements actually protected and the number of TLAs utilized. In
order to accomplish this, a probabilistic method for calculating the
lightning related failure and an artificial neural network for
estimating the SSFOR are presented. A multicriteria optimization
method based on a genetic algorithm is also developed to determine the
optimum location of TLAs.

2009



* 534(<-497): ANN for multi-objective optimal reactive compensation of a power system with wind generators

In this paper, we develop a method aimed to impose an acceptable
voltages profile and to reduce active losses of an electrical supply
network including wind generators in real time. These tasks are
ensured by acting on capacitor and reactor banks implemented in the
load nodes. To solve this problem, we minimize multi-objective
functions associated to the active losses and the compensation devices
cost under constraints imposed on the voltages and the reactive
productions of the various banks. The minimization procedure was
realised by the use of evolutionary algorithms. After a training
phase, a neural model has the capacity to provide a good estimation of
the voltages, the reactive productions and the losses for actual
curves of the load and the wind speed, in real time. (C) 2008 Elsevier
B.V. All rights reserved.

2008





* 536(<-311): Influence of hollow glass microspheres on the mechanical and physical properties and cost of particle reinforced polymer composites

The goal of the study was to find a cost-effective composition of a
particle reinforced composite that is light in weight but has
sufficient mechanical properties. The matrix of the particulate
composite is unsaturated polyester resin that is reinforced with
alumina trihydrate particles. Part of the alumina trihydrate
proportion was replaced with hollow glass microspheres to reduce
weight and save costs. In order to find out the influence of the light
filler on the physical and mechanical properties of composites,
materials with different percentages of the light filler were
prepared. Test specimens were cut from moulded sheets that were
fabricated with vacuum assisted extruder. Tensile strength,
indentation hardness measured with a Barcol impressor, and density
were determined. Based on the experimental data a multi-criteria
optimization problem was formulated and solved to find the optimal
design of the material. Artificial neural networks and a hybrid
genetic algorithm were used. The optimal solution is given as a Pareto
curve to represent the distinction between the density and selected
mechanical properties of the composite material. The composite
material filled with 6% hollow glass microspheres showed 3% loss in
the tensile strength and 26% loss in the surface hardness compared to
the composition without the filler. The weight decreased by 13%
compared with the initial composition. The addition of hollow glass
microspheres did not lower the net value of the material, it increased
7%.

2012


* 537(<- 77): Multi-Criteria Design Optimization of Ultra Large Diameter Permanent Magnet Generator

This paper presents a novel design optimization procedure for an ultra large diameter permanent magnet generator. As the machine features unorthodox electromagnetic and mechanical layouts, basic principles for determining structural loads together with material quantities for cost estimation are described. Finite element modelling with beam elements is used for retrieving stresses and deformations of the novel carrier structure. Mathematical system response model of the generator is created with artificial neural networks, while genetic algorithm with gradient method is utilized for determining the optimal solutions. Input dataset for the model build-up is constructed with a help of a full factorial experimental method. Achieved results are utilized for describing the relationship between the structural response and efficiency values of the generator. As the design of the machine has to fulfil contradicting technical and economical requirements, Pareto optimality concept is employed. As an example, a set of optimal solutions is determined for the particular case.

2015


* 538(<-527): A method for optimal design of automotive body assembly using multi-material construction

This paper proposes a new method for designing lightweight automotive body assemblies using multi-material construction with low cost penalty. Current constructions of automotive structures are based on single types of materials, e.g., steel or aluminium. The principle of the multi-material construction concept is that proper materials are selected for their intended functions. The design problem is formulated as a multi-objective nonlinear mathematical programming problem involving both discrete and continuous variables. The discrete variables are the material types and continuous variables are the thicknesses of the panels. This problem is then solved using a multi-objective genetic algorithm. An artificial neural network is employed to approximate the constraint functions and reduce the number of finite element runs. The proposed method is illustrated through a case study of lightweight design of an automotive door assembly. (c) 2007 Elsevier Ltd. All rights reserved.

2008


* 539(<- 56): Topographical optimisation of single-storey non-domestic steel framed buildings using photovoltaic panels for net-zero carbon impact

A methodology is presented that combines a multi-objective evolutionary algorithm and artificial neural networks to optimise single-storey steel commercial buildings for net-zero carbon impact. Both symmetric and asymmetric geometries are considered in conjunction with regulated, unregulated and embodied carbon. Offsetting is achieved through photovoltaic (PV) panels integrated into the roof. Asymmetric geometries can increase the south facing surface area and consequently allow for improved PV energy production. An exemplar carbon and energy breakdown of a retail unit located in Belfast UK with a south facing PV roof is considered. It was found in most cases that regulated energy offsetting can be achieved with symmetric geometries. However, asymmetric geometries were necessary to account for the unregulated and embodied carbon. For buildings where the volume is large due to high eaves, carbon offsetting became increasingly more difficult, and not possible in certain cases. The use of asymmetric geometries was found to allow for lower embodied energy structures with similar carbon performance to symmetrical structures. (C) 2014 Elsevier Ltd. All rights reserved.

2015


* 540(<-119): Free vibration analysis of an adhesively bonded functionally graded double containment cantilever joint

In this study, Genetic Algorithms (GAs) combined with the proposed neural networks were implemented to the free vibration analysis of an adhesively bonded double containment cantilever joint with a functionally graded plate. The proposed neural networks were trained and tested based on a limited number of data including the natural frequencies and modal strain energies calculated using the finite element method. GA evaluates a value generated iteratively by an objective function and this value is calculated by the finite element method. The iteration process restricts us apparently to use directly the finite element method in our multi-objective optimisation problem in which the natural frequency is maximised and the corresponding modal strain energy is minimised. The proposed neural networks were used accurately to predict the natural frequencies and modal strain energies instead of calculating directly them by using the finite element method. Consequently, the computation time and efforts were reduced considerably. The adhesive joint was observed to tend vertical bending modes and torsional modes. Therefore, the multi-objective optimisation problem was limited to only the first mode which appeared as a bending mode. The effects of the geometrical dimensions and the material composition variation through the plate thickness were investigated. As the material composition of the horizontal plate becomes ceramic rich, both natural frequency and modal strain energy of the adhesive joint increased regularly. The plate length and plate thickness were more effective geometrical design parameters whereas the support length and thickness were less effective. However, the adhesive thickness had a small effect on the optimal design of the adhesive joint as far as the natural frequencies and modal strain energies are concerned. The distributions of optimal solutions were also presented for the adhesive joints with fundamental joint lengths and material compositions in reference to their natural frequencies and corresponding modal strain energies.

2014




* 541(<-533): Multi-objective stacking sequence optimization of laminated cylindrical panels using a genetic algorithm and neural networks

A imulti-objective optimization strategy for optimal stacking sequence
of laminated cylindrical panels is presented, with respect to the
first natural frequency and critical buckling load, using the weighted
summation method. To improve the speed of the optimization process,
artificial neural networks are used to reproduce the behavior of the
structure both in free vibration and buckling conditions. Based on
first order shear deformation theory of laminated shells, a finite
element code, capable of evaluating the first natural frequency and
buckling load, is prepared of which the outputs are used for training
and testing the developed neural networks. In order to find the
optimal solution. a genetic algorithm is implemented. Verifications
are made for both finite element code results and utilization of
neural networks in the optimization process. With the purpose of
illustrating the optimization process, numerical results are presented
for a symmetric angle-ply six layer cylindrical panel. (c) 2006
Elsevier Ltd. All rights reserved.

2007


* 542(<-164): Dynamic Response and Optimal Design of Curved Metallic Sandwich Panels under Blast Loading

It is important to understand the effect of curvature on the blast response of curved structures so as to seek the optimal configurations of such structures with improved blast resistance. In this study, the dynamic response and protective performance of a type of curved metallic sandwich panel subjected to air blast loading were examined using LS-DYNA. The numerical methods were validated using experimental data in the literature. The curved panel consisted of an aluminum alloy outer face and a rolled homogeneous armour (RHA) steel inner face in addition to a closed-cell aluminum foam core. The results showed that the configuration of a "soft" outer face and a "hard" inner face worked well for the curved sandwich panel against air blast loading in terms of maximum deflection (MaxD) and energy absorption. The panel curvature was found to have a monotonic effect on the specific energy absorption (SEA) and a nonmonotonic effect on the MaxD of the panel. Based on artificial neural network (ANN) metamodels, multiobjective optimization designs of the panel were carried out. The optimization results revealed the trade-off relationships between the blast-resistant and the lightweight objectives and showed the great use of Pareto front in such design circumstances.

2014


* 543(<-184): Blast resistance and multi-objective optimization of aluminum foam-cored sandwich panels

In this work, a group of metallic aluminum foam-cored sandwich panels (AFSPs) were used as vehicle armor against blast loading. The dynamic responses of the AFSPs with various combinations of face-sheet materials were analyzed using LS-DYNA. It was found that the AFSP with an aluminum (AA2024 T3) front face and a Rolled Homogeneous Armor (RHA) steel back face (labeled T3-AF-RHA) outperformed the other panel configurations in terms of maximum back face deflection (MaxD) and areal specific energy absorption (ASEA). It was also found that boundary conditions and the standoff distance (SoD) between an explosive and a target surface both have a remarkable influence on the blast response of the AFSPs. Using artificial neural network (ANN) approximation models, multi-objective design optimization (MDO) of the T3-AF-RHA panel was performed both with and without variations in blast load intensity. The optimization results show that the two objectives of MaxD minimization and ASEA maximization conflict with each other and that the optimal designs must be identified in a Pareto sense. Moreover, the Pareto curves obtained are different for varied blast impulse levels. Consequently, it is concluded that loading variation should be considered when designing such sandwich armors to achieve more robust blast-resistant performance. (C) 2013 Elsevier Ltd. All rights reserved.

2013





* 544(<-515): Identification of Constitutive Parameters using Hybrid ANN multi-objective optimization procedure

This paper deals with the identification of material parameters for an
elastoplastic behaviour model with isotropic hardening using several
experimental tests at the same time. But, these tests are generally
inhomogeneous and finite element simulations are necessary for their
analysis. Therefore an inverse analysis is carried out and the
identification problem is converted into a multi-objective
optimization where prohibitive computing time is required. We propose
in this work a hybrid approach where Artificial Neural Networks (ANN)
are trained by finite element results. Then, the multi objective
procedure calls the ANN function in place of the finite element
code. The proposed approach is exemplified on the identification of
non-associative Hill'48 criterion and Voce parameters model of the
Stainless Steel AISI 304.

2008



* 545(<-528): IDENTIFICATION OF MATERIAL PARAMETERS BY HYBRID METHOD

Accurate identification of material parameters is an important task to
model the forming processes. In this paper several experimental tests
are used simultaneously to identify an elastoplastic behaviour model
with isotropic hardening. Yet, while this identification of the
material parameters is converted into a multi-objective optimization,
it results into a prohibitive computing time is required. In order to
overcome this issue, in this work a hybrid optimization approach based
both on the finite element and the artificial neural networks
computations is presented. The proposed strategy is used to identify
the Karafillis & Boyce criterion and the Voce parameters model of the
Stainless Steel AISI 304.

2008


* 546(<- 17): Aerothermal shape optimization for a double row of discrete film cooling holes on the suction surface of a turbine vane

A multiple-objective optimization is implemented for a double row of staggered film holes on the suction surface of a turbine vane. The optimization aims to maximize the film cooling performance, which is assessed using the cooling effectiveness, while minimizing the corresponding aerodynamic loss, which is measured with a mass-averaged total pressure coefficient. Three geometric variables defining the hole shape are optimized: the conical expansion angle, compound angle and length to diameter ratio of the non-diffused portion of the hole. The optimization employs a non-dominated sorting genetic algorithm coupled with an artificial neural network to generate the Pareto front. Reynolds-averaged Navier-Stokes simulations are employed to construct the neural network and investigate the aerodynamic and thermal optimum solutions. The optimum designs exhibit improved performance in comparison to the reference design. The optimization methodology allowed investigation into the impact of varying the geometric variables on the cooling effectiveness and the aerodynamic loss.

2015


* 547(<- 60): Aerothermal Optimization and Experimental Verification for Discrete Turbine Airfoil Film Cooling

The optimization aims to maximize the film cooling performance while minimizing the corresponding aerodynamic penalty. The cooling performance is assessed using the adiabatic film cooling effectiveness, while the aerodynamic penalty is measured with a mass-averaged total pressure loss coefficient. Two design variables are selected: the coolant-to-mainstream temperature ratio and the coolant-to-mainstream total pressure ratio. Two staggered rows of discrete cylindrical film cooling holes on the suction surface of a turbine vane are considered. Anondominated sorting genetic algorithm (NSGA-II) is coupled with an artificial neural network (ANN) to perform a multiple-objective optimization of the coolant flow parameters on the vane suction surface. Three-dimensional Reynolds-averaged Navier-Stokes (RANS) simulations are employed to construct the ANN network that produces low-fidelity predictions of the objective functions during the optimization. The effect of varying the coolant flow parameters on the adiabatic film cooling effectiveness and the aerodynamic loss is analyzed using the optimization method and RANS simulations. The computational fluid dynamics predictions of the adiabatic film cooling effectiveness and aerodynamic performance are assessed and validated against corresponding experimental measurements. The optimal solutions are reproduced in the experimental facility and the Pareto front is substantiated with experimental data.

2015


* 548(<-229): CFD modeling and multi-objective optimization of cyclone geometry using desirability function, artificial neural networks and genetic algorithms

The low-mass loading gas cyclone separator has two performance parameters, the pressure drop and the collection efficiency (cut-off diameter). In this paper, a multi-objective optimization study of a gas cyclone separator has been performed using the response surface methodology (RSM) and CFD data. The effects of the inlet height, the inlet width, the vortex finder diameter and the cyclone total height on the cyclone performance have been investigated. The analysis of design of experiment shows a strong interaction between the inlet dimensions and the vortex finder diameter. No interaction between the cyclone height and the other three factors was observed. The desirability function approach has been used for the multi-objective optimization. A new set of geometrical ratios (design) has been obtained to achieve the best performance. A numerical comparison between the new design and the Stairmand design confirms the superior performance of the new design. As an alternative approach for applying RSM as a meta-model, two radial basis function neural networks (RBFNNs) have been used. Furthermore, the genetic algorithms technique has been used instead of the desirability function approach. A multi-objective optimization study using NSGA-II technique has been performed to obtain the Pareto front for the best performance cyclone separator. (C) 2012 Elsevier Inc. All rights reserved.

2013


* 549(<-304): Modeling and Pareto optimization of gas cyclone separator performance using RBF type artificial neural networks and genetic algorithms

Both the pressure drop and the cut-off diameter are important performance parameters in the design of the cyclone separator. In this paper, a multi-objective optimization study of the gas cyclone separator is performed. In order to predict accurately the complex non linear relationships between the performance parameters (pressure drop and cut-off diameter) and the geometrical dimensions, two radial basis function neural networks (RBFNNs) are developed and employed to model the pressure drop and the cut-off diameter for cyclone separators. The artificial neural networks have been trained and tested by the experimental data available in literatures for the pressure drop and the lozia and Leith model for the cut-off diameter. The results demonstrate that artificial neural networks can offer an alternative and powerful approach to model the cyclone performance parameters. The analysis indicates the significant effect of the vortex finder diameter D-x, the vortex finder lengths, the inlet width b and the total height H-t. The response surface methodology has been used to fit a second-order polynomial to the RBFNN. The second-order polynomial has been used to study the interaction between the geometrical parameters. The two trained artificial neural networks have been used as two objective functions to get new optimal ratios for minimum pressure drop and minimum cut-off diameter using the multi-objective genetic algorithm optimization technique. Sometimes, the main concern is minimizing the pressure drop, so a single objective optimization study has been performed to obtain the cyclone geometrical ratio for minimum pressure drop. The comparison of numerical simulation of the new optimal design and the Stairmand design confirms the superior performance of the new design. (C) 2011 Elsevier B.V. All rights reserved.

2012


* 550(<-335): CFD modeling and multi-objective optimization of compact heat exchanger using CAN method

Thermal modeling and optimal design of compact heat exchanger is presented in this paper. Fin pitch, fin height, cold stream flow length, no-flow length and hot stream flow length were considered as five design parameters. A CFD analysis coupled with artificial neural network was used to develop a relation between Colburn factor and Fanning friction factor for the triangle fin geometry with acceptable precision. Then, fast and elitist non-dominated sorting genetic algorithm (NSGA-II) was applied to obtain the maximum effectiveness and the minimum total pressure drop as two objective functions. The results of optimal designs were a set of multiple optimum solutions, called 'Pareto-optimal solutions'. It reveals that any geometrical changes which decrease the pressure drop in the optimum situation, lead to a decrease in the effectiveness and vice versa. Finally sensitivity analysis shows the increases of heat transfer surface area necessarily do not increases the pressure drop and it is case sensitive. (C) 2011 Elsevier Ltd. All rights reserved.

2011


* 551(<-412): Thermal-economic multi-objective optimization of plate fin heat exchanger using genetic algorithm

Thermal modeling and optimal design of compact heat exchangers are presented in this paper. epsilon-NTU method was applied to estimate the heat exchanger pressure drop and effectiveness. Fin pitch, fin height, fin offset length, cold stream flow length, no-flow length and hot stream flow length were considered as six design parameters. Fast and elitist non-dominated sorting genetic-algorithm (NSGA-II) was applied to obtain the maximum effectiveness and the minimum total annual cost (sum of investment and operation costs) as two objective functions. The results of optimal designs were a set of multiple optimum solutions, called 'Pareto optimal solutions'. The sensitivity analysis of change in optimum effectiveness and total annual cost with change in design parameters of the plate fin heat exchanger was also performed and the results are reported. As a short cut for choosing the system optimal design parameters the correlations between two objectives and six decision variables with acceptable precision were presented using artificial neural network analysis. (C) 2009 Elsevier Ltd. All rights reserved.

2010




* 567(<-383): Optimization of a Centrifugal Compressor Impeller II-Artificial Neural Network and Genetic Algorithm

The optimization of a centrifugal compressor was conducted. The ANN
(Artificial Neural Network) was adopted as an optimization algorithm,
and it was learned and trained with the DOE (Design of Experiment). In
the DOE, it was predicted the main effect and the interaction effect
of design variables to the objective function. The ANN was improved in
the optimization process using the GA (Genetic Algorithm). When any
output at each generation was reached a standard level, it was
re-calculated by the CFD (Computational Fluid Dynamics) and it was
applied to develop a new ANN. After 6th generation, the prediction
difference between ANN and CFD was less than 1%. A pareto of the
efficiency versus the pressure ratio was obtained through the 21th
generation. Using this method, the computational time for the
optimization was equivalent to the time consumed by the gradient
method, and the optimized results of multi-objective function were
obtained.

2011




* 568(<- 47): A hybrid evolutionary performance improvement procedure for optimisation of continuous variable discharge concentrators

An iterative hybrid performance improvement approach integrating
artificial neural network modelling and Pareto genetic algorithm
optimisation was developed and tested. The optimisation procedure,
code named NNREGA, was tested for tuning pilot scale Continuous
Variable Discharge Concentrator (CVD) in order to simultaneously
maximise recovery and upgrade ratio of gold bearing sulphides from a
polymetallic massive sulphide ore. For the tests the CVD was
retrofitted during normal operation on the flotation tailings
stream. On the basis of mineralogical data showing strong pyrite-gold
association in the flotation tailings, iron assays were used as an
indicator of CVD performance on recovery of gold bearing
sulphides. Initially, 17 pilot scale statistically designed tests were
conducted to assess metallurgical performance. The Matlab 2010a
software was used to train and simulate back propagation ANNs on
experimental results. Regression models developed from simulation data
were used to formulate the objective functions used to optimise the
CVD using the NSGA-II genetic algorithm. Results show that the NNREGA
procedure provides an efficient way of exploring the design space to
learn the relationship between interacting variables and outputs and
is capable of generating the operating line, which is a non-dominated
recovery/grade line. The paper forms a basis for future work aiming to
model and scale up processing equipment. (C) 2015 Elsevier B.V. All
rights reserved.

2015




* 569(<-258): Multidisciplinary optimization of collapsible cylindrical energy absorbers under axial impact load

In this article, the multi-objective optimization of cylindrical aluminum tubes under axial impact load is presented. The specific absorbed energy and the maximum crushing force are considered as objective functions. The geometric dimensions of tubes including diameter, length and thickness are chosen as design variables. D/t and L/D ratios are constricted in the range of which collapsing of tubes occurs in concertina or diamond mode. The Non-dominated Sorting Genetic Algorithm-II is applied to obtain the Pareto optimal solutions. A back-propagation neural network is constructed as the surrogate model to formulate the mapping between the design variables and the objective functions. The finite element software ABAQUS/Explicit is used to generate the training and test sets for the artificial neural networks. To validate the results of finite element model, several impact tests are carried out using drop hammer testing machine.

2012





* 570(<- 62): Pareto genetic design of group method of data handling type neural network for prediction discharge coefficient in rectangular side orifices

The powerful method of Group Method of Data Handling (GMDH) was used
for estimating the discharge coefficient of a rectangular side
orifice. First, the existing equations for calculating the discharge
coefficient were studied making use of experimental results. On the
first hand, the factors affecting the discharge coefficient were
determined, then five models were constructed in order to analyze the
sensitivity in achieving accuracy by using different parameters. The
results, obtained using statistical indexes (MARE=0.021 and
RMSE=0.017), showed that one model out of the five models, on
estimation using the dimensionless parameters of the ratio of depth of
flow in main channel to width of rectangular orifice (Y-m/L), Froude
number (Fr), the ratio of sill height to width of rectangular orifice
(W/L) and width of main channel to width of rectangular orifice (B/L),
presented the best results. (C) 2014 Elsevier Ltd. All rights
reserved.

2015



* 571(<-159): APPLICATION OF ANN AND GA FOR THE PREDICTION AND OPTIMIZATION OF THERMAL AND FLOW CHARACTERISTICS IN A RECTANGULAR CHANNEL FITTED WITH TWISTED TAPE VORTEX GENERATORS

This study reports an application of the hybrid model, including back
propagation network and genetic algorithm, for predicting the thermal
and flow characteristics in a rectangular channel fitted with multiple
twisted tape vortex generators (MT-VG). Dimensionless geometric
parameters and Reynolds number were considered as network inputs, and
Nusselt number and friction factor were the output variables. The
performance of the developed neural networks was found to be superior
in comparison with the empirical correlations. In addition, the
proposed networks have been considered as two objective functions in
order to obtain optimal operation conditions. Since mentioned
objectives are conflicting, the multi-objective optimization using
genetic algorithm was used for the optimization.

2014


* 572(<-380): Multi-objective shape optimization of helico-axial multiphase pump impeller based on NSGA-II and ANN

In order to improve the prototype's performance of the helico-axial multiphase pump, a multi-objective optimal method for the pump impeller was developed by combining the artificial neural network (ANN) with non-dominated sorting genetic algorithm-II (NSGA-II). The main geometric parameters influencing the impeller's performance were chosen as the optimization variables, and the sample spaces were structured according to the orthogonal experimental design method. Then the pressure rise and efficiency in specific working conditions were obtained about all the elements in the sample space by numerical simulation. With the simulated results as the input specimen, a multiphase pump performance prediction model was designed through BP neural network. With the obtained prediction model as the fitness value evaluation method, the pump impeller was optimized using the NSGA-II multi-objective genetic algorithm, which finally offered an improved impeller structure with enhanced pressure rise and efficiency. Furthermore, five stages of optimized compression cells were manufactured and applied in experiment test. The result shows compared to the original design, the pressure rise of the optimized pump has increased by similar to 10% and the efficiency has increased by similar to 3%, which is in keeping with our optimal result and confirms our method is feasible. (C) 2010 Elsevier Ltd. All rights reserved.

2011


* 573(<-401): An optimization strategy for die design in the low-density polyethylene annular extrusion process based on FES/BPNN/NSGA-II

An optimization strategy for die design in the polymer extrusion process is proposed in the study based on the finite element simulation, the back-propagation neural network, and the non-dominated sorting genetic algorithm II (NSGA-II). The three-dimensional simulation of polymer melts flow in the extrusion process is conducted using the penalty finite element method. The model for predicting the flow patterns in the extrusion process is established with the artificial neural network based on the simulated results. The non-dominated sorting genetic algorithm II is performed for the search of globally optimal design variables with its objective functions evaluated by the established neural network model. The proposed optimization strategy is successfully applied to the die design in low-density polyethylene (LDPE) annular extrusion process. A constrained multi-objective optimization model is established according to the characteristics of annular extrusion process. The minimum of velocity relative difference, delta u, and the minimum of swell ratio, S (w), that, respectively, ensure the extrinsic feature, mechanical property, and dimensional precision of the final products are taken as optimization objectives with a constrained condition on the maximum shear stress. Three important die structure parameters, including the die contraction angle alpha, the ratio of parallel length to inner radius L/R (i), and the ratio of outer to inner radius R (o) /R (i), are taken as design variables. The Phan-Thien-Tanner constitutive model is adopted to describe the viscoelastic rheological characteristics of LDPE whose parameters are fitted by the distributions of material functions detected on the strain-controlled rheometer. The penalty finite element model of polymer melts flowing through out of the extrusion die is derived. A decoupled method is employed to solve the viscoelastic flow problem with the discrete elastic-viscous split-stress algorithm. The simulated results are selected and extracted to constitute the learning samples according to the orthogonal experimental design method. The back propagation algorithm is adopted for the training and the establishment of the predicting model for the optimization objective. A Pareto-optimal set for the constrained multi-objective optimization is obtained using the constrained NSGA-II, and the optimal solution is extracted based on the fuzzy set theory. The optimization for die parameters in the annular extrusion process of low-density polyethylene is performed and the optimization objective is successfully achieved.

2010



* 574(<- 66): Application of multi-objective genetic algorithm to optimize energy efficiency and thermal comfort in building design

Several conflicting criteria exist in building design optimization, especially energy consumption and indoor environment thermal performance. This paper presents a novel multi-objective optimization model that can assist designers in green building design. The Pareto solution was used to obtain a set of optimal solutions for building design optimization, and uses an improved multi-objective genetic algorithm (NSGA-II) as a theoretical basis for building design multi-objective optimization model. Based on the simulation data on energy consumption and indoor thermal comfort, the study also used a simulation-based improved back-propagation (BP) network which is optimized by a genetic algorithm (GA) to characterize building behavior, and then establishes a GA-BP network model for rapidly predicting the energy consumption and indoor thermal comfort status of residential buildings; Third, the building design multi-objective optimization model was established by using the GA-BP network as a fitness function of the multi-objective Genetic Algorithm (NSGA-II); Finally, a case study is presented with the aid of the multi-objective approach in which dozens of potential designs are revealed for a typical building design in China, with a wide range of trade-offs between thermal comfort and energy consumption. (C) 2014 Elsevier B.V. All rights reserved.

2015




* 576(<-201): Airflow and temperature distribution optimization in data centers using artificial neural networks

To control energy usage in data center rooms, reduced order models are
important in order to perform real-time assessment of the optimum
operating conditions to reduce energy usage. Here computational fluid
dynamics (CFD) simulation-based Artificial Neural Network (ANN) models
were developed and applied to a basic hot aisle/cold aisle data center
configuration in order to predict thermal operating conditions for a
specified set of control variables. Once trained, the ANN-based model
predictions were shown to agree well with the CFD results for
arbitrary values of the input variables within the specified
limits. In addition, the ANN model was combined with a cost function
based multi-objective Genetic Algorithm (GA), which enabled the
operating conditions to be inversely predicted for specified values of
the output variable (e.g., server rack inlet temperatures). The ANN-GA
optimization approach considerably reduces the total computation time
compared to a fully CFD-based response surface optimization
methodology. Consequently, operating conditions are capable of being
reliably predicted in seconds, even for configurations outside of the
original ANN training set. These results show that an ANN based model
can yield an effective real-time thermal management design tool for
data centers. (c) 2013 Elsevier Ltd. All rights reserved.

2013



* 577(<-326): Multi-Objective Optimization of aluminum hollow tubes for vehicle crash energy absorption using a genetic algorithm and neural networks

A numerical study of the crushing of thin-walled circular aluminum tubes has been carried out to investigate their behaviors under axial impact loading. These kinds of tubes are usually used in automobile and train structures to absorb the impact energy. A Multi-Objective Optimization of circular aluminum tubes undergoing axial compressive loading for vehicle crash energy absorption is performed for five crushing parameters using the weighted summation method. To improve the accuracy of the optimization process, artificial neural networks are used to reproduce the behavior of the crushing parameters in crush dynamics conditions. An explicit finite element method (FEM) is used to model and analyzed the behavior. A series of aluminum cylindrical tubes are simulated under axial impact condition for the experimental validation of the numerical solutions. A finite element code, capable of evaluating parameters crush, is prepared of which the outputs are used for training and testing the developed neural networks. In order to find the optimal solution, a genetic algorithm is implemented. With the purpose of illustrating optimum dimensional ratios, numerical results are presented for thin-walled circular aluminum AA6060-T5 and AA6060-T4 tubes. Multi-Objective Optimization of circular aluminum tubes has been performed in the basis of different priorities to create the ability for designer to select the optimum dimension ratio. Also, crush parameters of two aluminum alloys has been compared. (C) 2011 Elsevier Ltd. All rights reserved.

2011


* 578(<-374): Multi-point and multi-objective optimization design method for industrial axial compressor cascades

Modern aerodynamic optimization design methods for the industrial axial compressor cascade mainly aim at improving both design point and off-design point performance. In this study, a multi-point and multi-objective optimization design method is established for the cascade, particularly aiming at widening the operating range while maintaining good performance at the acceptable expense of computational load. The design objectives are to maximize the static pressure ratio and minimize the total pressure loss coefficient at the design point, and to maximize the operating range for the positive and negative incidences. To alleviate the computational load, a design of experiment (DOE)-based GA-BP-ANN model is constructed to rapidly approximate the cascade aerodynamic performance in the optimization process. The artificial neural network (ANN) is trained by the genetic algorithm (GA) technique and back propagation (BP) algorithm, where the training cascades are sampled by the DOE method and analysed by the computational fluid dynamics method. The multi-objective genetic algorithm is used to search for a series of Pareto-optimum solutions, from which an optimal cascade is found out whose objectives are all better than (ABT) those of the original design. The ABT cascade is characterized by the lower camber and higher turning angle, leading to better aerodynamic performance in a widened operating range. Compared with the original design, the ABT cascade decreases the total pressure loss coefficient by 1.54 per cent, 23.4 per cent, and 7.87 per cent at the incidences of 5 degrees, -9 degrees, and 13 degrees, respectively. The established optimization design method can be extended to the three-dimensional aerodynamic design of axial compressor blade.

2011



* 579(<-106): Multiobjective optimization of composite cylindrical shells for strength and frequency using genetic algorithm and neural networks

In this paper, the optimal fiber orientations relative to the
principle axis of composite cylindrical shell composed of four and six
layers were determined so that the natural frequency and strength of
the shell are optimized. For this purpose, first, the free vibration
analysis of the shell was carried out based on 3D elasticity. Then,
for calculation of the strength objective function, the inverse form
of Tsai-Hill yield criteria was used and the functions of strength and
frequency were developed in terms of fiber orientation. Once the
correctness of the above solutions was ensured, the objective
functions were modeled with artificial neural network (ANN). The model
made was then introduced into genetic algorithm (GA) and the maximum
fitness function and optimal staking sequence of the layers with
respect to the fibers angles were obtained. Optimal solutions obtained
by combination of ANN and GA are compared to the solutions obtained by
analytical solution and GA; eventually, the tables and diagrams are
presented and different fiber orientations as optimization solutions
are presented as the final results of the composite shell analysis.

2014




* 581(<-262): Soft computing based multi-objective optimization of steam cycle power plant using NSGA-II and ANN

In this paper a steam turbine power plant is thermo-economically modeled and optimized. For this purpose, the data for actual running power plant are used for modeling, verifying the results and optimization. Turbine inlet temperature, boiler pressure, turbines extraction pressures, turbines and pumps isentropic efficiency, reheat pressure as well as condenser pressure are selected as fifteen design variables. Then, the fast and elitist Non-dominated Sorting Genetic Algorithm (NSGA-II) is applied to maximize the thermal efficiency and minimize the total cost rate (sum of investment cost, fuel cost, and maintenance cost) simultaneously. The results of the optimal design are a set of multiple optimum solutions, called 'Pareto optimal solutions'. The optimization results in some points show 3.76% increase in efficiency and 3.84% decrease in total cost rate simultaneously, when it compared with the actual data of the running power plant. Finally as a short cut to choose the system optimal design parameters a correlation between two objectives and fifteen decision variables with acceptable precision are presented using Artificial Neural Network (ANN). (c) 2012 Elsevier B.V. All rights reserved.

2012


* 582(<- 43): Experimental and numerical analysis of the optimized finned-tube heat exchanger for OM314 diesel exhaust exergy recovery

In this research, a multi objective optimization based on Artificial Neural Network (ANN) and Genetic Algorithm (GA) are applied on the obtained results from numerical outcomes for a finned-tube heat exchanger (HEX) in diesel exhaust heat recovery. Thirty heat exchangers with different fin length, thickness and fin numbers are modeled and those results in three engine loads are optimized with weight functions for pressure drop, recovered heat and HEX weight. Finally, two cases of HEXs (an optimized and a non-optimized) are produced experimentally and mounted on the exhaust of an OM314 diesel engine to compare their results in heat and exergy recovery. All experiments are done for five engine loads (0%, 20%, 40%, 60% and 80% of full load) and four water mass flow rates (50, 40, 30 and 20 g/s). Results show that maximum exergy recovers occurs in high engine loads and optimized HEX with 10 fins have averagely 8% second law efficiency in exergy recovery. (C) 2015 Elsevier Ltd. All rights reserved.

2015


* 583(<-105): Multi-objective optimization of nanofluid flow in flat tubes using CFD, Artificial Neural Networks and genetic algorithms

In this article, multi-objective optimization of Al2O3-water nanofluid parameters in flat tubes is performed using Computational Fluid Dynamics (CFD) techniques, Artificial Neural Networks (ANN) and Non-dominated Sorting Genetic Algorithms (NSGA II). At first, nanofluid flow is solved numerically in various flat tubes using CFD techniques and heat transfer coefficient ((h) over bar) and pressure drop (Delta P) in tubes are calculated. In this step, two phase mixture model is applied for nanofluid flow analysis and the flow regime is also laminar. In the next step, numerical data of the previous step will be applied for modeling (h) over bar and Delta P using Grouped Method of Data Handling (GMDH) type ANN. Finally, the modeling achieved by GMDH will be used for Pareto based multi-objective optimization of nanofluid parameters in horizontal flat tubes using NSGA II algorithm. It is shown that the achieved Pareto solution includes important design information on nanofluid parameters in flat tubes. (C) 2014 The Society of Powder Technology Japan. Published by Elsevier B. V. and The Society of Powder Technology Japan. All rights reserved.

2014


* 584(<-175): Thermal modeling of gas engine driven air to water heat pump systems in heating mode using genetic algorithm and Artificial Neural Network methods

The gas-engine driven air-to-water heat pump, type air conditioning system, is composed of two major thermodynamic cycles (including the vapor compression refrigeration cycle and the internal combustion gas engine cycle) as well as a refrigerant-water plate heat exchanger. The thermal modeling of gas engine driven air-to-water heat pump system with engine heat recovery heat exchangers was performed here for the heating mode of operation (in which it was required to model engine heat recovery heat exchanger). The modeling was performed using typical thermodynamic characteristics of system components, Artificial Neural Network and the multi-objective genetic algorithm optimization method. The comparison of modeling results with experimental ones showed average differences of 5.08%, 5.93%, 5.21%, 2.88% and 6.2% which shows acceptable agreement for operating pressure, gas engine fuel consumption, outlet water temperature, engine rotational speed, and system primary energy ratio. (C) 2013 Elsevier Ltd and HR. All rights reserved.

2013



* 585(<-446): Optimization of the core configuration design using a hybrid artificial intelligence algorithm for research reactors

To successfully carry out material irradiation experiments and
radioisotope productions, a high thermal neutron flux at irradiation
box over a desired life time of a core configuration is needed. On the
other hand, reactor safety and operational constraints must be
preserved during core configuration selection. Two main objectives and
two safety and operational constraints are suggested to optimize
reactor core configuration design. Suggested parameters and conditions
are considered as two separate fitness functions composed of two main
objectives and two penalty functions. This is a constrained and
combinatorial type of a multi-objective optimization problem. in this
paper, a fast and effective hybrid artificial intelligence algorithm
is introduced and developed to reach a Pareto optimal set. The hybrid
algorithm is composed of a fast and elitist multi-objective genetic
algorithm (GA) and a fast fitness function evaluating system based on
the cascade feed forward artificial neural networks (ANNs). A specific
GA representation of core configuration and also special GA operators
are introduced and used to overcome the combinatorial constraints of
this optimization problem. A software package (Core Pattern
Calculator 1) is developed to prepare and reform required data for
ANNs training and also to revise the optimization results. Some
practical test parameters and conditions are suggested to adjust main
parameters of the hybrid algorithm. Results show that introduced ANNs
can be trained and estimate selected core parameters of a research
reactor very quickly. It improves effectively optimization
process. Final optimization results show that a uniform and dense
diversity of Pareto fronts are gained over a wide range of fitness
function values. To take a more careful selection of Pareto optimal
solutions, a revision system is introduced and used. The revision of
gained Pareto optimal set is performed by using developed software
package. Also some secondary operational and safety terms are
suggested to help for final trade-off. Results show that the selected
benchmark case study is dominated by gained Pareto fronts according to
the main objectives while safety and operational constraints are
preserved. (C) 2009 Elsevier B.V. All rights reserved.

2009


* 586(<-460): Estimation of research reactor core parameters using cascade feed forward artificial neural networks

The pattern of the core reload program is very important for an optimize use of research reactors. Reactor safety issues and economic efficiency should be considered during pattern studies. In order to find the best core pattern for a research reactor, its reloading program should be solved as a multi-objective and constrained optimization problem. If considered objective functions of the optimization problem can be estimated in very short time, the optimal fuel reloading pattern can be used effectively. In this research a very fast estimation system for suggested core parameters has been developed using cascade feed-forward type of artificial neural networks (ANNs). Four main core parameters are suggested to optimize reactor core adequately. And also to get larger thermal fluxes in the desired irradiation box, a new flexible method was selected. A Software package has been developed to prepare and reform required data for ANNs training. The gradient descent method with momentum weight/bias learning rule has been used to train ANNs. To get the best conditions for considered ANNs training a vast study has been performed. It includes the effects of variation of hidden neurons, hidden layers, activation functions, learning and momentum coefficients, and also the number of training data sets on the training and simulation results. Some experimental convergence criteria are used to study them. A comparison selection rule has been used to adjust desirable conditions. Final training and simulation results show that developed ANNs can be trained and estimate suggested core parameters of research reactors very quickly. It improves effectively pattern optimization process of core reload program. (C) 2009 Elsevier Ltd. All rights reserved.

2009





* 598(<- 33): An optimal design of wind turbine and ship structure based on neuro-response surface method

The geometry of engineering systems affects their performances. For this reason, the shape of engineering systems needs to be optimized in the initial design stage. However, engineering system design problems consist of multi-objective optimization and the performance analysis using commercial code or numerical analysis is generally time-consuming. To solve these problems, many engineers perform the optimization using the approximation model (response surface). The Response Surface Method (RSM) is generally used to predict the system performance in engineering research field, but RSM presents some prediction errors for highly nonlinear systems. The major objective of this research is to establish an optimal design method for multi-objective problems and confirm its applicability. The proposed process is composed of three parts: definition of geometry, generation of response surface, and optimization process. To reduce the time for performance analysis and minimize the prediction errors, the approximation model is generated using the Backpropagation Artificial Neural Network (BPANN) which is considered as Neuro-Response Surface Method (NRSM). The optimization is done for the generated response surface by non-dominated sorting genetic algorithm-II (NSGA-II). Through case studies of marine system and ship structure (substructure of floating offshore wind turbine considering hydrodynamics performances and bulk carrier bottom stiffened panels considering structure performance), we have confirmed the applicability of the proposed method for multi-objective side constraint optimization problems.

2015




* 600(<-198): Multi-Objective Prediction Model and Parameter Optimization Model for the Sputtering of Aluminum Zinc Oxide Semiconducting Transparent Thin Films

The industry demand for both high conductivity and transmittance in
thin films has made it essential to develop a multi-objective
prediction model for resistivity and transmittance. This study
combined Taguchi methods and artificial neural networks (ANN) to
construct a multi-objective prediction model for the sputtering of AZO
(ZnO:Al = 97:3 wt%) to produce semiconducting transparent thin
films. The Levenberg-Marquardt method was incorporated into the
multi-objective prediction model to construct a multi-objective
parameter optimization model for AZO semiconducting transparent thin
films. The squared difference of the objective values and the
predicted values of each objective served as the error function, which
was then multiplied by the individual weight values and summed to
derive the objective function of the system. In conjunction with the
Levenberg-Marquardt method and reasonable convergence criteria, the
optimal combination of parameters for the sputtering objectives was
obtained. These parameters included radio frequency power
(R. F. power) power of 120 W, process pressure of 15 mTorr, film
thickness of 300 nm, and substrate temperature of 74 degrees C. The
objective resistivity was 11.4 x 10(-3) Omega . cm, and the objective
transmittance was 88.9%. In this experiment, resistivity resulted in
10.6 x 10(-3) Omega . cm, with an error of 7.5% between the predicted
value and the experiment results. Transmittance reached 89.1% in the
experiment, accounting for an error of -0.2%.

2013





* 602(<-495): Optimization of composition and technology for phosphate graphite mold

In present work, the characteristics of three methods such as the
orthogonal design, Fuzzy optimum method and artificial neural network
modeling technique were made on the basis of the optimization and
evaluation of the performance of the phosphate graphite mold. The
variance analysis indicates that the phosphoric acid has greatest
influence on both compression strength and tension strength of
phosphate graphite mold, both drying temperature and drying time
greater, and Al2O3 minor, respectively. The Fuzzy multi-objective
comprehensive evaluation shows that the optimum technology for
phosphate graphite mold designed by us is phosphoric acid 30%, Al2O3
30%, drying temperature 400 degrees C and drying time 60 min. In
addition, the ANN can be used to establish mono- and multi-objective
models for the prediction of other tests outside orthogonal test with
rather high accuracy. However, the predicted results are worse for the
linear regressive equations by the orthogonal analysis. (C) 2008
Elsevier Ltd. All rights reserved.

2008




* 604(<-346): Application of artificial bee colony-based neural network in bottom hole pressure prediction in underbalanced drilling

Two phase flow through annulus is a complex area of study in
evaluating the bottom hole circulating pressure (BHCP). Based on the
over-prediction of empirical correlations and the erroneous assumption
of hydraulic diameter concept, both methods suffer from a great deal
of error. As a result, it is investigated in this work how artificial
neural network (ANN) evolution with artificial bee colony (ABC)
improves the efficiency and prediction capability of artificial neural
network. The proposed methodology adopts a hybrid ABC-back propagation
(BP) strategy (ABC-BP). The proposed algorithm combines the local
searching ability of the gradient-based back-propagation (BP) strategy
with the global searching ability of artificial bee colony. For an
evaluation purpose, the performance and generalization capabilities of
ABC-BP are compared with those of models developed with the common
technique of BP. The results demonstrate that carefully designed
hybrid artificial bee colony-back propagation neural network
outperforms the gradient descent-based neural network. (C) 2011
Elsevier B.V. All rights reserved.

2011




* 607(<- 92): Multi Objective Optimization of Friction Stir Welding Parameters Using FEM and Neural Network

In this study the inuence of rotational and traverse speed on the friction stir welding of AA5083 aluminum alloy has been investigated For this purpose a thermo-mechanically coupled, 3D FEM analysis was used to study the effect of rotational and traverse speed on welding force, peak temperature and HAZ width. Then, an Articial Neural Network (ANN) model was employed to understand the correlation between the welding parameters (rotational and traverse speed) and peak temperature, HAZ width and welding force values in the weld zone. Performance of the ANN model was found excellent and the model can be used to predict peak temperature, HAZ width and welding force. Furthermore, in order to find optimum values of traverse and rotational speed, the multi-objective optimization was used to obtain the Pareto front. Finally, the Technique for Order Preference by Similarity to the Ideal Solution (TOPSIS) was employed to obtain the best compromised solution.

2014




* 614(<-  4): MODELING ENGINE FUEL CONSUMPTION AND NOx WITH RBF NEURAL NETWORK AND MOPSO ALGORITHM

In this study, artificial neural network (ANN) modeling is used to
predict the fuel consumption and NOx emission of a four stroke spark
ignition (SI) engine. Calibration engineers frequently want to know
the responses of an engine for the entire range of operating
conditions in order to change engine control parameters in the
electronic control unit (ECU), to improve performance and reduce
emissions. However, testing the engine for the complete range of
operating conditions is a very time and labor consuming task. As
alternative, ANN is used in order to predict fuel consumption and NOx
emission. In the proposed approach, the multi-objective particle swarm
optimization (MOPSO) is used to determine weights of radial basis
function (RBF) neural networks. The goal is to minimize performance
criteria as root mean square error (RMSE) and model complexity. A
sensitivity analysis is performed on MOPSO parameters in order to
provide better solutions along the optimal Pareto front. In order to
select a compromised solution among the obtained Pareto solutions, a
fuzzy decision maker is employed. The correlation coefficient R-2 is
used to compare the engine responses with the obtained by the proposed
approach.

2015


* 615(<- 12): Combining artificial neural network and multi-objective optimization to reduce a heavy-duty diesel engine emissions and fuel consumption

Nondominated sorting genetic algorithm II (NSGA-II) is well known for
engine optimization problem. Artificial neural networks (ANNs)
followed by multi-objective optimization including a NSGA-II and
strength pareto evolutionary algorithm (SPEA2) were used to optimize
the operating parameters of a compression ignition (CI) heavy-duty
diesel engine. First, a multi-layer perception (MLP) network was used
for the ANN modeling and the back propagation algorithm was utilized
as training algorithm. Then, two different multi-objective
evolutionary algorithms were implemented to determine the optimal
engine parameters. The objective of the present study is to decide
which algorithm is preferable in terms of performance in engine
emission and fuel consumption optimization problem.

2015


* 616(<-121): Artificial neural network based on genetic algorithm for emissions prediction of a SI gasoline engine

This paper proposes a hybrid learning of artificial neural network (ANN) with the nondominated sorting genetic algorithm-II (NSGAII) to improve accuracy in order to predict the exhaust emissions of a four stroke spark ignition (SI) engine. In the proposed approach, the genetic algorithm (GA) determines initial weights of local linear model tree (LOLIMOT) neural networks. A multi-objective optimization problem is determined. A sensitivity analysis is performed on NSGA-II parameters in order to provide better solutions along the optimal Pareto front. Then, a fuzzy decision maker and the technique for order preference by similarity to ideal solution (TOPSIS) are employed to select compromised solutions among the obtained Pareto solutions. The LOLIMOT-GA responses are compared with the provided by radial basis function (RBF) and multilayer perceptron (MLP) neural networks in terms of correlation coefficient R (2).

2014


* 617(<-192): A hybrid method of modified NSGA-II and TOPSIS to optimize performance and emissions of a diesel engine using biodiesel

This paper addresses artificial neural network (ANN) modeling followed by multi-objective optimization process to determine optimum biodiesel blends and speed ranges of a diesel engine fueled with castor oil biodiesel (COB) blends. First, an ANN model was developed based on standard back-propagation algorithm to model and predict brake power, brake specific fuel consumption (BSFC) and the emissions of engine. In this way, multi-layer perception (MLP) network was used for non-linear mapping between the input and output parameters. Second, modified NSGA-II by incorporating diversity preserving mechanism called the e-elimination algorithm was used for multi-objective optimization process. Six objectives, maximization of brake power and minimization of BSFC, PM, NOx, CO and CO2 were simultaneously considered in this step. Optimization procedure resulted in creating of non-dominated optimal points which gave an insight on the best operating conditions of the engine. Third, an approach based on TOPSIS method was used for finding the best compromise solution from the obtained set of Pareto solutions. (C) 2013 Elsevier Ltd. All rights reserved.

2013


* 618(<-197): Modeling and multi-objective optimization of a gasoline engine using neural networks and evolutionary algorithms

In this paper, a multi-objective particle swarm optimization (MOPSO) algorithm and a nondominated sorting genetic algorithm II (NSGA-II) are used to optimize the operating parameters of a 1.6 L, spark ignition (SI) gasoline engine. The aim of this optimization is to reduce engine emissions in terms of carbon monoxide (CO), hydrocarbons (HC), and nitrogen oxides (NO (x) ), which are the causes of diverse environmental problems such as air pollution and global warming. Stationary engine tests were performed for data generation, covering 60 operating conditions. Artificial neural networks (ANNs) were used to predict exhaust emissions, whose inputs were from six engine operating parameters, and the outputs were three resulting exhaust emissions. The outputs of ANNs were used to evaluate objective functions within the optimization algorithms: NSGA-II and MOPSO. Then a decision-making process was conducted, using a fuzzy method to select a Pareto solution with which the best emission reductions can be achieved. The NSGA-II algorithm achieved reductions of at least 9.84%, 82.44%, and 13.78% for CO, HC, and NO (x) , respectively. With a MOPSO algorithm the reached reductions were at least 13.68%, 83.80%, and 7.67% for CO, HC, and NO (x) , respectively.

2013


* 619(<-613): Multi-objective optimization of heavy-duty diesel engines under stationary conditions

New technological developments are helping to control contaminants in diesel engines but, as new degrees of freedom become available, the assessment of optimal values that combine to reduce different emissions has become a difficult task. This paper studies the feasibility of using artificial neural networks (ANNs) as models to be integrated in the optimization of diesel engine settings, with the objective of complying with the increasingly stringent emission regulations while also keeping, or even reducing, the fuel consumption. A large database of stationary engine tests covering a wide range of experimental conditions was used for the development of the ANN models. The optimization was developed within the frame of the European legislation for heavy duty diesel engines. Experimental validation of the optimized results was carried out and compared with the ANN predictions, showing a high level of accuracy, especially for fuel consumption and nitrogen oxides (NOx).

2005




* 625(<- 37): Computational intelligence based designing of microalloyed pipeline steel

Computational intelligence based modeling and optimization techniques
are employed primarily to investigate the role of the composition and
processing parameters on the mechanical properties of API grade
microalloyed pipeline steel and then to design steel having improved
performance in respect to its strength, impact toughness and
ductility. Artificial Neural Network (ANN) models, capable of
prediction and diagnosis in non-linear and complex systems, are used
to obtain the relationship of composition and processing parameters
with said mechanical properties. Then the models are used as objective
functions for the multi-objective genetic algorithms for evolving the
tradeoffs between the conflicting objectives of achieving improved
strength, ductility and impact toughness. The Pareto optimal solutions
are analyzed successfully to study the role of various parameters for
designing pipeline steel with such improved performance. (C) 2015
Elsevier B.V. All rights reserved.

2015




* 627(<-244): Modelling and Pareto optimization of mechanical properties of friction stir welded AA7075/AA5083 butt joints using neural network and particle swarm algorithm

Friction Stir Welding (FSW) has been successfully used to weld similar and dissimilar cast and wrought aluminium alloys, especially for aircraft aluminium alloys, that generally present with low weldability by the traditional fusion welding process. This paper focuses on the microstructural and mechanical properties of the Friction Stir Welding (FSW) of AA7075-O to AA5083-O aluminium alloys. Weld microstructures, hardness and tensile properties were evaluated in as-welded condition. Tensile tests indicated that mechanical properties of the joint were better than in the base metals. An Artificial Neural Network (ANN) model was developed to simulate the correlation between the Friction Stir Welding parameters and mechanical properties. Performance of the ANN model was excellent and the model was employed to predict the ultimate tensile strength and hardness of butt joint of AA7075-AA5083 as functions of weld and rotational speeds. The multi-objective particle swarm optimization was used to obtain the Pareto-optimal set. Finally, the Technique for Order Preference by Similarity to the Ideal Solution (TOPSIS) was applied to determine the best compromised solution. (C) 2012 Elsevier Ltd. All rights reserved.

2013


* 628(<-352): Designing cold rolled IF steel sheets with optimized tensile properties using ANN and GA

Artificial neural network (ANN) models, correlating the mechanical properties (yield strength, tensile strength, %elongation and f) of the cold rolled interstitial free (IF) steel sheets with compositional and processing parameters, are used to find the importance of different variables. Further the above models are used as objective functions for the evolutionary multi-objective optimization algorithms by evolving tradeoffs, which gives a range of combination of strength and ductility. The optimal solutions are also analysed to extract further knowledge on the role of various parameters, which could be used for designing the chemistry of the alloy as well as process parameters for producing cold rolled strips of IF steel with tailor made property. (C) 2011 Elsevier B.V. All rights reserved.

2011


* 629(<-458): An approach for the aging process optimization of Al-Zn-Mg-Cu series alloys

A new model based on least square support vector machines (LSSVM) and capable of forecasting mechanical and electrical properties of Al-Zn-Mg-Cu series alloys has been proposed for the first time. Data mining and artificial intelligence techniques of aluminum alloys are used to examine the forecasting capability of the model. In order to improve predictive accuracy and generalization ability of LSSVM model, a grid algorithm and cross-validation technique has been adopted to determine the optimal hyper-parameters of LSSVM automatically. The forecasting performance of the LSSVM model and the artificial neural network (ANN) has been compared with the experimental values. The result shows that the LSSVM model provides slightly better capability of generalized prediction compared to back propagation network (BPN) in combination with the gradient descent training algorithm. Considering its advantages of the computation speed, unique optimal solution, and generalization performance, the LSSVM model is therefore considered to be used as an alternative powerful modeling tool for the aging process optimization of aluminum alloys. Furthermore. a novel methodology hybridizing nondominated sorting-based multi-objective genetic algorithm (MOGA) and LSSVM has been proposed to make tradeoffs between the mechanical and electrical properties. A desirable nondominated solution set has been obtained and reported. (c) 2008 Elsevier Ltd. All rights reserved.

2009


* 630(<-549): Aging process optimization for a copper alloy considering hardness and electrical conductivity

A multi-objective optimization methodology for the aging process parameters is proposed which simultaneously considers the mechanical performance and the electrical conductivity. An optimal model of the aging processes for Cu-Cr-Zr-Mg is constructed using artificial neural networks and genetic algorithms. A supervised artificial neural network (ANN) to model the non-linear relationship between parameters of aging treatment and hardness and conductivity properties is considered for a Cu-Cr-Zr-Mg lead frame alloy. Based on the successfully trained ANN model, a genetic algorithm is adopted as the optimization scheme to optimize the input parameters. The result indicates that an artificial neural network combined with a genetic algorithm is effective for the multi-objective optimization of the aging process parameters. (c) 2006 Elsevier B.V. All rights reserved.

2007



* 631(<-168): DEVELOPMENT OF A CLOSED-LOOP DIAGNOSIS SYSTEM FOR REFLOW SOLDERING USING NEURAL NETWORKS AND SUPPORT VECTOR REGRESSION

This study presents an industrial application of artificial neural
network (ANN) and support vector regression (SVR) to diagnose control
reflow soldering process in a closed-loop framework. Reflow soldering
is the principal process for the fabrication of a variety of modern
computer, communication, and consumer (3C) electronics products. It is
important to achieve robust electrical connections without changing
the mechanical and electronic characteristics of the components during
reflow soldering process. In this study, a 3(8-4) experimental design
was conducted to collect the structured process information. The
experimental data was then used for data-training via the ANN and SVR
techniques to investigate both the forward and backward relationships
between the heating factors and the resultant reflow thermal profile
(RTP) and so as to develop a closed-loop reflow soldering diagnosis
system. The proposed system includes two modules: (1) a forward-flow
module used to predict the output elements of the RTP and evaluate its
performance based on ANN and a multi-criteria decision-making (MCDM)
criterion; (2) a backward-flow module employed to ascertain the set of
heating parameter combinations which best fulfill the production
requirements of the expected throughput rate, product configuration,
and the desired solderability. The efficiency and cost-effectiveness
of this methodology were empirically evaluated and the results show
the promising to improve soldering quality and
productivity. Significance: The proposed closed-loop reflow soldering
process diagnosis system can predict the output elements of a reflow
temperature profile according to process inputs. This system is also
able to ascertain the set of heating parameter combinations which best
fulfill the production requirements and the desired solderability. The
empirical evaluation demonstrates the efficiency and
cost-effectiveness for the improvements of soldering quality and
productivity.

2014




* 633(<-203): Multi-objective prediction model for the establishment of sputtered GZO semiconducting transparent thin films

In recent years, semiconducting transparent thin films have undergone rapid development. Today, excellent conductivity and transmittance are the qualities sought in the manufacturing of these films. Most manufacturers have the objective of enhancing both conductivity and transmittance, developing a multi-objective model for the prediction of resistivity and transmittance in semiconducting transparent thin films is essential. Taguchi analysis results indicate that among the factors influencing resistivity, radio frequency power (R. F. power) is the most significant, followed by process pressure. Among the factors influencing transmittance, target-to-substrate distance is the most significant, followed by R. F. power. This study proposed a progressive Taguchi-neural network model, combining Taguchi method with an artificial neural network for the development of a multi-objective prediction model for use with sputtered gallium zinc oxide (GZO, ZnO:Ga=97:3 wt%) semiconducting transparent thin films. Analysis results have shown that in the Stage-1 of the initial network, prediction results were ineffective due to insufficient network training examples. The refined network in the Stage-3 however, provided improved global prediction results.

2013





* 634(<-239): Integration of finite element simulation and intelligent methods for evaluation of thermo-mechanical loads during hard turning process

The machined surfaces are mainly affected by thermo-mechanical loads
during machining processes. In this regard, thermal loads increase
tensile residual stress and heat-affected zone; however, mechanical
loads increase fatigue strength and compressive residual stress on the
machined workpiece during the process. Since experimental
investigation is difficult, the problem becomes more difficult if the
aim is minimizing thermal loads, while maximizing mechanical loads
during the hard turning process. This article presents a hybrid method
based on the artificial neural networks, multiobjective optimization,
and finite element analysis for evaluation of thermo-mechanical loads
during the orthogonal turning of AISI H13-hardened die steel
(52HRC). First, using an iterative procedure, controllable parameters
of simulation (including contact conditions and flow stress) are
determined by comparison between finite element and experimental
results from the literature. Then, the results of finite element
simulation at the different cutting conditions and tool geometries
were employed for training neural networks by genetic
algorithm. Finally, the functions implemented by neural networks were
considered as objective functions of nondominated genetic algorithm
and optimal nondominated solution set were determined at the different
states of thermal loads (workpiece temperature) and mechanical loads
(workpiece effective strain). Comparison between the obtained results
of nondominated genetic algorithm and predicted results of finite
element simulation showed that the hybrid technique of finite element
method-artificial neural networks-multiobjective optimization provides
a robust framework for machining simulation of AISI H13.

2013



* 635(<-301): Particle swarm optimization in multi-stage operations for operation sequence and DT allocation

Improved operation sequence and economic tolerance allocation directly
influence product quality and manufacturing costs. The purpose of this
study is to generate the optimal operation sequence and allocate
economic tolerances to cutting surfaces to achieve the specified
quality and minimize the manufacturing costs. Because this type of
problem is a multi-objective optimization problem subject to various
constraints, it is defined as an NP-hard problem. A three-step
procedure is used to solve the problem. First, a mathematical model is
developed to define the relationships between manufacturing costs and
tolerances. Second, an artificial neural network (ANN) is applied to
obtain the best fitting cost-tolerance function. Finally, the
formulated mathematical models are solved by using particle swarm
optimization (PSO) in order to determine the optimal operation
sequence. In addition, both the effectiveness and efficiency of the
proposed methodologies are tested and verified for a given workpiece
that needs multi-stage operations. The key contributions of this study
are the generation of the optimal operation sequence and the effective
allocation of the optimal dimensional tolerance (DT) using an advanced
computational intelligence algorithm with consideration for
multi-stage operations. (C) 2011 Elsevier Ltd. All rights reserved.

2012


* 636(<-319): Constrained optimum surface roughness prediction in turning of X20Cr13 by coupling novel modified harmony search-based neural network and modified harmony search algorithm

Nowadays, manufacturers rely on trustworthy methods to predict the optimal cutting conditions which result in the best surface roughness with respect to the fact that some constraining functions should not exceed their critical values because of current restrictions considering competition found among them in delivering economical and high-quality products to the stringent customers in the shortest time. The present research deals with a modified optimization algorithm of harmony search (MHS) coupled with modified harmony search-based neural networks (MHSNN) to predict the cutting condition in longitudinal turning of X20Cr13 leading to optimum surface roughness. To this end, several experiments were carried out on X20Cr13 stainless steel to attain the required data for training of MHSNN. Feed-forward artificial neural network was utilized to create predictive models of surface roughness and cutting forces exploiting experimental data, and the MHS algorithm was used to find the constrained optimum of surface roughness. Furthermore, simple HS algorithm was used to solve the same optimization problem to illustrate the capabilities of the MHS algorithm. The obtained results demonstrate that the MHS algorithm is more effective and authoritative in approaching the global solution compared with the HS algorithm.

2012


* 637(<-375): Modelling Power Consumption in Ball-End Milling Operations

Power consumption is a factor of increasing interest in manufacturing due to its obvious impact on production costs and the environment. The aim of this work is to analyze the influence of process parameters on power consumption in high-speed ball-end milling operations carried out on AISI H13 steel. A total of 300 experiments were carried out in a 3-axis vertical milling center, the Deckel-Maho 105V linear. The power consumed by the spindle and by the X, Y, and Z machine tool axes was measured using four ammeters located in the respective power cables. The data collected was used to develop an artificial neural network (ANN) which was used to predict power consumption during operations. The results obtained from the ANN are very accurate. Power consumption predictions can help operators to determine the most effective cutting parameters for saving energy and money while bringing the milling process closer to the goal of environmentally sensitive manufacturing which has become a topic of general importance.

2011


* 638(<- 29): MULTI-OBJECTIVE OPTIMIZATION OF CUT QUALITY CHARACTERISTICS IN CO2 LASER CUTTING OF STAINLESS STEEL

In this paper, multi-objective optimization of the cut quality characteristics in CO2 laser cutting of AISI 304 stainless steel was discussed. Three mathematical models for the prediction of cut quality characteristics such as surface roughness, kerf width and heat affected zone were developed using the artificial neural networks (ANNs). The laser cutting experiment was planned and conducted according to the Taguchi's L-27 orthogonal array and the experimental data were used to train single hidden layer ANNs using the Levenberg-Marquardt algorithm. The ANN mathematical models were developed considering laser power, cutting speed, assist gas pressure, and focus position as the input parameters. Multi-objective optimization problem was formulated using the weighting sum method in which the weighting factors that are used to combine cut quality characteristics into the single objective function were determined using the analytic hierarchy process method.

2015


* 639(<-255): MULTI-OBJECTIVE OPTIMIZATION OF SURFACE ROUGHNESS AND MATERIAL REMOVAL RATE IN CO2 LASER CUTTING USING ANN AND NSGA-II

The prediction of optimal laser cutting conditions for satisfying different requirements is of great importance in process planning. hi this paper, multi-objective optimization of CO2 laser cutting AISI 304 stainless steel using the non-dominated sorting genetic algorithm (NSGA-II), with surface roughness and material removal rate (MRR) as the objective functions was presented. The laser cutting experiments were conducted based on Taguchi's experimental design using L-27 orthogonal array by varying the laser power, cutting speed, assist gas pressure and focus position at three levels. Using these experimental data, the mathematical models of surface roughness and kerf width were developed using artificial neural network (ANN). The later ANN model was then used for calculating the MRR considering that the MRR is the function of cutting speed, workpiece thickness and kerf width. On the basis of a computer code written for ANN function models, the optimization problem was formulated and solved using NSGA-II. The obtained optimal solution set was plotted as Pareto optimal front. It was observed that the functional dependence between the surface roughness and material removal rate is nonlinear and can be expressed with a second degree polynomial.

2013


* 640(<-165): Artificial intelligence based modeling and optimization of heat affected zone in Nd:YAG laser cutting of duralumin sheet

Duralumin is an alloy of aluminium which has some unique properties such as high strength to weight ratio, high resistance to corrosion, light in weight, and more demanding alloy in various sectors such as space craft, marine, chemical industries, construction and automobile. These applications require very precise and complex shapes which may not be obtained with conventional machining. Pulsed Nd:YAG laser cutting may be used to fulfill these objectives by using optimum setting of process parameters. The present research paper has experimentally investigated the modeling and optimization of heat affected zone in the pulsed Nd:YAG laser cutting of Duralumin sheet with the aim to minimize heat affected zone. The quality is improved by the proper control of different process parameters such as gas pressure, pulse width, pulse frequency and scanning speed. Artificial intelligence (AI) algorithms have been used to solve the many engineering problems successfully through development of Genetic Algorithm (GA), Fuzzy Logic (FL) and Artificial Neural Network (ANN) systems. The optimization of heat affected zone has been carried out by using Hybrid Approach of Multiple Regression Analysis (MRA) and GA. In this methodology, the second order regression model has been developed by using MRA with the help of experimental data obtained by L-27 orthogonal array (OA). Further this equation has been used as objective function in GA based optimization. The significant factors have been found with further discussion of their effect on the heat affected zone.

2014


* 641(<-456): Modeling and optimization on Nd:YAG laser turned micro-grooving of cylindrical ceramic material

Nd:YAG laser turning is a new technique for manufacturing micro-grooves on cylindrical surface of ceramic materials needed for the present day precision industries. The importance of laser turning has directed the researchers to search how accurately micro-grooves can be obtained in cylindrical parts. In this paper, laser turning process parameters have been determined for producing square micro-grooves on cylindrical surface. The experiments have been performed based on the statistical five level central composite design techniques. The effects of laser turning process parameters i.e. lamp current, pulse frequency, pulse width, cutting speed (revolution per minute, rpm) and assist gas pressure on the quality of the laser turned micro-grooves have been studied. A predictive model for laser turning process parameters is created using a feed-forward artificial neural network (ANN) technique utilized the experimental observation data based on response surface methodology (RSM). The optimization problem has been constructed based on RSM and solved using multi-objective genetic algorithm (GA). The neural network coupled with genetic algorithm can be effectively utilized to find the optimum parameter value for a specific laser micro-turning condition in ceramic materials. The optimal process parameter settings are found as lamp current of 19 A, pulse frequency of 3.2 kHz, pulse width of 6% duty cycle, cutting speed as 22 rpm and assist air pressure of 0.13 N/mm(2) for achieving the predicted minimum deviation of upper width of -0.0101 mm, lower width 0.0098 mm and depth -0.0069 mm of laser turned micro-grooves. (C) 2009 Elsevier Ltd. All rights reserved.

2009


* 642(<-488): Neural Network Modeling and Particle Swarm Optimization (PSO) of Process Parameters in Pulsed Laser Micromachining of Hardened AISI H13 Steel

This article focuses on modeling and optimizing process parameters in pulsed laser micromachining. Use of continuous wave or pulsed lasers to perform micromachining of 3-D geometrical features on difficult-to-cut metals is a feasible option due the advantages offered such as tool-free and high precision material removal over conventional machining processes. Despite these advantages, pulsed laser micromachining is complex, highly dependent upon material absorption reflectivity, and ablation characteristics. Selection of process operational parameters is highly critical for successful laser micromachining. A set of designed experiments is carried out in a pulsed Nd:YAG laser system using AISI H13 hardened tool steel as work material. Several T-shaped deep features with straight and tapered walls have been machining as representative mold cavities on the hardened tool steel. The relation between process parameters and quality characteristics has been modeled with artificial neural networks (ANN). Predictions with ANNs have been compared with experimental work. Multiobjective particle swarm optimization (PSO) of process parameters for minimum surface roughness and minimum volume error is carried out. This result shows that proposed models and swarm optimization approach are suitable to identify optimum process settings.

2009


* 643(<-588): Multi criteria optimization of laser percussion drilling process using artificial neural network model combined with genetic algorithm

This is a study of laser percussion drilling optimization by combining the neural network method with the genetic algorithm. First, optimum input parameters of the process were obtained in order to optimize every single output parameter (response) of the zprocess regardless of their effect on each other (single criterion optimization). Then, optimum input parameters were obtained in order to optimize the effect of all output parameters in a multicriteria manner. Artificial neural network (ANN) method was employed to develop an experimental model of the process according to the experimental results. Then optimum input parameters (peak power, pulse width, pulse frequency, number of pulses, assist gas pressure, and focal plane position) were specified by using the genetic algorithm (GA). The output parameters include the hole entrance diameter, circularity of hole entrance and hole exit, and hole taper. The tests were carried out on mild steel EN3 sheets, with 2.5 mm thickness. The sheets were drilled by a 400 w pulsed Nd:YAG laser emitting at 1.06 mu m wave length. Oxygen was employed as the assist gas. Considering the accuracy of the optimum numerical results and the high capability of the neural network in modeling, this method is reliable and precise and confirms the qualitative results in the previous studies. As a result, one can use this method to optimally adjust input parameters of the process in multicriteria optimization mode, which indicates substitute application of the method for optimizing the laser percussion drilling process.

2006


* 644(<-436): Determination of Optimal Pulse Metal Inert Gas Welding Parameters with a Neuro-GA Technique

Optimization of a manufacturing process is a rigorous task because it has to take into account all the factors that influence the product quality and productivity. Welding is a multi-variable process, which is influenced by a lot of process uncertainties. Therefore, the optimization of welding process parameters is considerably complex. Advancement in computational methods, evolutionary algorithms, and multiobjective optimization methods create ever-more effective solutions to this problem. This work concerns the selection of optimal parameters setting of pulsed metal inert gas welding (PMIGW) process for any desired output parameters setting. Six process parameters, namely pulse voltage, background voltage, pulse frequency, pulse duty factor, wire feed rate and table feed rate were used as input variables, and the strength of the welded plate, weld bead geometry, transverse shrinkage, angular distortion and deposition efficiency were considered as the output variables. Artificial neural network (ANN) models were used for mapping input and output parameters. Neuro genetic algorithm (Neuro-GA) technique was used to determine the optimal PMIGW process parameters. Experimental result shows that the designed parameter setting of PMIGW process, which was obtained from Neuro-GA optimization, indeed produced the desired weld-quality.

2010



* 646(<-540): Optimization of the characteristic parameters in milling using the PSO evolution technique

The selection of machining parameters is an important step in process
planning; therefore, a new evolutionary computation technique has been
developed to optimize the machining process. In this paper Particle
Swarm Optimization (PSO) is used to efficiently optimize the machining
parameters simultaneously in milling processes where multiple
conflicting objectives are present. First, an artificial neural
network (AAW) predictive model is used to predict the cutting forces
during machining and then the PSO algorithm is used to obtain the
optimum cutting speeds and feed rates. The goal of the optimization is
to determine the objective function maximum (the predicted
cutting-force surface) by considering the cutting constraints. During
optimization the particles fly intelligently in the solution space and
search for optimal cutting conditions according to the strategies of
the PSO algorithm. The results showed that an integrated system of
neural networks and swarm intelligence is an effective method for
solving multi-objective optimization problems. The high accuracy of
the results within a wide range of machining parameters indicates that
the system can be practically applied in industry. The simulation
results show that compared with genetic algorithms (GAs) and simulated
annealing (SA) the proposed algorithm can improve the quality of the
solution while speeding up the convergence process. The new
computational technique has several advantages and benefits and is
suitable for use when combined with ANN-based models where no explicit
relation between the inputs and the outputs is available. This
research opens the door for a new class of optimization techniques
that are based on evolution computation in the area of machining. (C)
2007 Journal of Mechanical Engineering. All rights reserved.

2007




* 647(<- 55): Applying the ensemble artificial neural network-based hybrid data-driven model to daily total load forecasting

Accurate electricity forecasting has become a very important research
field for high-efficiency electricity production. But the hybrid
data-driven models for load forecasting are rarely studied. This paper
presents a novel hybrid data-driven "PEK" model for predicting the
daily total load. The proposed hybrid model is mainly constructed by
various function approximators, which containing the partial mutual
information (PMI)-based input variable selection (IVS), ensemble
artificial neural network-based output estimation and K-nearest
neighbor regression-based output error estimation. The PMI-based IVS
algorithm is used to select the input variables, resulting in a good
compromise between the parsimony and adequacy of the input
information. After that, the topology and parameter calibration of the
PEK model are implemented by the NSGA-II multi-objective optimization
algorithm. The electricity load demands from years 2010 to 2012 of the
Shuyang hydrothermal station are chosen as a case study to verify the
performance of the PEK model. Simulation results show that this model
obtains significantly better accuracy in the prediction of daily total
load.

2015




* 649(<-163): WELDING PROCESS OPTIMIZATION WITH ARTIFICIAL NEURAL NETWORK APPLICATIONS

Correct detection of input and output parameters of a welding process
is significant for successful development of an automated welding
operation. In welding process literature, we observe that output
parameters are predicted according to given input parameters. As a new
approach to previous efforts, this paper presents a new modeling
approach on prediction and classification of welding parameters. 3
different models are developed on a critical welding process based on
Artificial Neural Networks (ANNs) which are (0 Output parameter
prediction, (ii) Input parameter prediction (reverse application of
output prediction model) and (iii) Classification of products. In this
study, firstly we use Pareto Analysis for determining uncontrollable
input parameters of the welding process based on expert views. With
the help of these analysis, 9 uncontrollable parameters are determined
among 22 potential parameters. Then, the welding process of ammunition
is modeled as a multi-input multi-output process with 9 input and 3
output parameters. 1st model predicts the values of output parameters
according to given input values. 2nd model predicts the values of
correct input parameter combination for a defect-free weld operation
and 3rd model is used to classify the products whether defected or
defect-free. 3rd model is also used for validation of results obtained
by 1st and 2nd models. A high level of performance is attained by all
the methods tested in this study. In addition, the product is a
strategic ammunition in the armed forces inventory which is
manufactured in a limited number of countries in the world. Before
application of this study, the welding process of the product could
not be carried out in a systematic way. The process was conducted by
trial-and-error approach by changing input parameter values at each
operation. This caused a lot of costs. With the help of this study,
best parameter combination is found, tested, validated with ANNs and
operation costs are minimized by 30%.

2014


* 650(<-615): A hybrid analytical-neural network approach to the determination of optimal cutting conditions

In the contribution, a new hybrid optimization technique for complex optimization of cutting parameters is proposed. The developed approach is based on the maximum production rate criterion and incorporates 10 technological constraints. It describes the multi-objective technique of optimization of cutting conditions by means of the artificial neural network (ANN) and OPTIS routine by taking into consideration the technological, economic and organizational limitations. The analytical module OPTIS selects the optimum cutting conditions from commercial databases with respect to minimum machining costs. By selection of optimum cutting conditions, it is possible to reach a favourable ratio between the low machining costs and high productivity taking into account the given limitation of the cutting process. To reach higher precision of the predicted results, a hybrid optimization algorithm is developed and presented to ensure simple, fast and efficient optimization of all important turning parameters. Experimental results show that the proposed optimization algorithm for solving the nonlinear-constrained programming problems (NCP) is both effective and efficient, and can be integrated into an intelligent manufacturing system for solving complex machining optimization problems. To demonstrate the procedure and performance of the proposed approach, an illustrative example is discussed in detail. (C) 2004 Elsevier B.V. All rights reserved.

2004


* 651(<- 74): ARTIFICIAL NEURAL NETWORK MODELING AND OPTIMIZATION OF HALL-HEROULT PROCESS FOR ALUMINUM PRODUCTION

Experience in applying a hybrid artificial neural network (ANN)-genetic algorithm for modeling and optimizing the Hall-Heroult process for aluminum extraction is described in this study. During the stage of modeling, the most important and effective process variables including temperature and cell voltage, metal and bath heights, purity of CaF2 and Al2O3, and bath ratio are chosen as input variables whilst outputs of the model are product purity, ampere efficiency, and product rate. During three years of operation, 19 points were selected for building and training, 7 points for testing, and 7 data points for validating the model. Results show that a feed-forward Artificial Neural Network (ANN) model with 3 neurons in the hidden layer can acceptably simulate the mentioned output variables with the Mean Squared Error (MSE) of 0.002%, 0.108% and 0.407%, respectively. Utilizing the validated model and multi-objective genetic algorithms, aluminum purity and the rate of production are maximized by manipulating decision variables. Results show that setting these decision variables at the optimal values can increase approximately the metal purity, ampere efficiency, and product rate by 0.007%, 0.185%, and 20kg/h, respectively.

2015




* 654(<- 38): Multiobjective optimization of friction welding of UNS S32205 duplex stainless steel

The present study is to optimize the process parameters for friction
welding of duplex stainless steel (DSS UNS S32205). Experiments were
conducted according to central composite design. Process variables, as
inputs of the neural network, included friction pressure, upsetting
pressure, speed and burn-off length. Tensile strength and
microhardness were selected as the outputs of the neural networks. The
weld metals had higher hardness and tensile strength than the base
material due to grain refinement which caused failures away from the
joint interface during tensile testing. Due to shorter heating time,
no secondary phase intermetallic precipitation was observed in the
weld joint. A multi-layer perceptron neural network was established
for modeling purpose. Five various training algorithms, belonging to
three classes, namely gradient descent, genetic algorithm and
Levenberg-Marquardt, were used to train artificial neural network. The
optimization was carried out by using particle swarm optimization
method. Confirmation test was carried out by setting the optimized
parameters. In conformation test, maximum tensile strength and maximum
hardness obtained are 822 MPa and 322 Hv, respectively. The
metallurgical investigations revealed that base metal, partially
deformed zone and weld zone maintain austenite/ferrite proportion of
50:50. Copyright (C) 2015, China Ordnance Society. Production and
hosting by Elsevier B.V. All rights reserved.

2015


* 655(<-234): Influence of Ultrasonic and Microwave Irradiation on Cation Exchange Properties of Clay Material

This study deals with optimization of the clay activation process using artificial neural network models and multi-objective optimization function. Different artificial neural network models were used for description of the relation between clay sorption capacity and the activation treatment process (power and time of clay exposure to ultrasonic and/or microwave irradiation). Two methodologies (feed-forward and cascade-forward) in combination with five different training algorithms (random order incremental training with learning functions, resilient backpropagation, one-step secant backpropagation, Levenberg-Marquardt backpropagation, Bayesian regularization backpropagation) were applied in order to obtain an optimal artificial neural network model. The optimal artificial neural network model showed good predictive ability (relative error 6.02 % based on external validation data set). In-house developed multi-objective criteria function was used in combination with the developed artificial neural network model and calculated optimal activation was determined (5 minutes of ultrasonic 120 W and microwave 60 W treatment) increasing the sorption capacity by 15 %.

2013





* 656(<-236): Artificial Neural Network Modeling of ECAP Process

Equal channel angular pressing (ECAP) is a type of severe plastic
deformation procedure for achieving ultra-fine grain structures. This
article investigates artificial neural network (ANN) modeling of ECAP
process based on experimental and three-dimensional (3D) finite
element methods (FEM).In order to do so, an ECAP die was designed and
manufactured with the channel angle of 90 degrees and the outer corner
angle of 15 degrees. Commercial pure aluminum was ECAPed and the
obtained data was used for validating the FEM model. After confirming
the validity of the model with experimental data, a number of
parameters are considered. These include the die channel angles (angle
between the channels phi and the outer corner angle ) and the number
of passes which were subsequently used for training the ANN. Finally,
experimental and numerical data was used to train neural networks. As
a result, it is shown that a feed forward back propagation ANN can be
used for efficient die design and process determination in the
ECAP. There is satisfactory agreement between results according to
comparisons.

2013



* 657(<-286): Bayesian regularization-based Levenberg-Marquardt neural model combined with BFOA for improving surface finish of FDM processed part

Fused deposition modeling has a complex part building mechanism making
it difficult to obtain reasonably good functional relationship between
responses and process parameters. To solve this problem, present study
proposes use of artificial neural network (ANN) model to determine the
relationship between five input parameters such as layer thickness,
orientation, raster angle, raster width, and air gap with three output
responses viz., roughness in top, bottom, and side surface of the
built part. Bayesian regularization is adopted for selection of
optimum network architecture because of its ability to fix number of
network parameters irrespective of network size. ANN model is trained
using Levenberg-Marquardt algorithm, and the resulting network has
good generalization capability that eliminates the chance of over
fitting. Finally, bacterial foraging optimization algorithm which
attempts to model the individual and group behavior of Escherichia
coli bacteria as a distributed optimization process is used to suggest
theoretical combination of parameter settings to improve overall
roughness of part. This paper also investigates use of chaotic time
series sequence known as logistic function and demonstrates its
superiority in terms of convergence and solution quality.

2012


* 658(<-317): Artificial Neural Network-Based Multiobjective Optimization of Mechanical Alloying Process for Synthesizing of Metal Matrix Nanocomposite Powder

The aim of this article was to optimize the mechanical alloying process for synthesizing of Al-8vol%SiC nanocomposite powders through an artificial neural network based on multiobjective optimization procedure. First, a suitable trained multi-layer perceptron (MLP) neural network was established for modeling purpose. Process variables as inputs of the network included milling time, milling speed, and balls to powders weight ratio. Parameters of the nanocomposite as outputs of the network were the crystallite size and the lattice strain of the aluminum matrix. The optimization was carried out by using two methods: gradient descent and pattern search. The aim of the optimization was to determine the minimum crystallite size and the maximum lattice strain of the aluminum matrix that could be obtained by regulating the mechanical alloying process variables. The response surfaces and the contour plots showed that the combination of the artificial neural network (ANN) and the optimization procedure were able to optimize the mechanical alloying process to synthesize Al-8vol%SiC nanocomposite.

2012





* 659(<-385): Artificial Neural Network Modeling of Forming Limit Diagram

Forming limit diagram (FLD) provides the limiting strains a sheet
metal can sustain whilst being formed. In this article, the
formability of Ti6Al4V titanium alloy and Al6061-T6 aluminum alloy
sheets is investigated experimentally using hydroforming deep
drawing. Hecker's simplified technique [1] was used to obtain
experimental FLDs for these sheet materials. Artificial neural network
(ANN) modeling of the process based on experimental results is
introduced to predict FLDs. It is shown that a feed forward back
propagation (BP) ANN can predict the FLDs, therefore, indicating the
possibility of ANN as a strong tool in simulating the
process. According to comparisons there is a good agreement between
experimental and neural network results.

2011


* 660(<-438): FEM and ANN Analysis in Fine-Blanking Process

Fine-blanking (FB) is an effective and economical shearing process
that offers a precise and clean cutting-edge finish, eliminates
unnecessary secondary operations, and increases quality. In the
traditional blanking product development paradigm, the design of the
formed product and tooling is usually based on know-how and
experience, which are generally obtained through long years of
apprenticeship and skilled craftsmanship. In this study, the
possibility of using finite element method (FEM) together with
artificial neural networks (ANN) was investigated to analysis the
fine-blanking process. Finite element analysis was used to simulate
the process with an isotropic elastic-plastic material model. The
results compare well with experimental results available in the
literature; after confirming the validity of the model with
experimental data, a number of parameters such as V-ring height
effect, punch and holder force on die-roll, hydrostatic pressure
status as an important factor in increasing burnish zone, and accuracy
of part and radial stress status as a factor in increasing die
erosion, which were also used for training the ANN, were
considered. Finally, numerical data were used to train neural
networks. The Levenberg-Marquardt (LM) algorithm with three neurons in
the hidden layer (LM-3) appeared to be the most optimal topology and
gives the best results. It was found that the coefficient of multiple
determinations (R2 value) between the FEM and ANN predicted data is
equal to about 0.999 for the size of die-roll, therefore indicating
the possibility of FEM and ANN as a powerful design tool for the
fine-blanking process.

2010




* 663(<-196): Crashworthiness design of multi-component tailor-welded blank (TWB) structures

Crashworthiness of tailor-welded blank (TWB) structures signifies an
increasing concern in lightweight design of vehicle. Although
multiobjective optimization (MOO) has to a considerable extent been
successfully applied to enhance crashworthiness of vehicular
structures, majority of existing designs were restricted to single or
uniform thin-walled components. Limited attention has been paid to
such non-uniform components as TWB structures. In this paper, MOO of a
multi-component TWB structure that involves both the B-pillar and
inner door system subjected to a side impact, is proposed by
considering the structural weight, intrusive displacements and
velocity of the B-pillar component as objectives, and the thickness in
different positions and the height of welding line of B-pillar as the
design variables. The MOO problem is formulated by using a range of
different metamodeling techniques, including response surface
methodology (RSM), artificial neural network (ANN), radial basis
functions (RBF), and Kriging (KRG), to approximate the sophisticated
nonlinear responses. By comparison, it is found that the constructed
metamodels based upon the radial basis function (RBF, especially
multi-quadric model, namely RBF-MQ) fit to the design of experiment
(DoE) checking points well and are employed to carry out the design
optimization. The performance of the TWB B-pillar and indoor panel
system can be improved by optimizing the thickness of the different
parts and height of the welding line. This study demonstrated that the
multi-component TWB structure can be optimized to further enhance the
crashworthiness and reduce the weight, offering a new class of
structural/material configuration for lightweight design.

2013





* 664(<-266): Holistic Approach to Decision Making in the Formulation and Selection of Anti-Icing Products

To effectively fight snow storms in the challenging funding
environment, many maintenance agencies in North America have started
to produce their own anti-icing liquids, instead of procuring
commercial anti-icers. This work demonstrates a systematic approach to
collaborative, data-driven, and multicriteria decision making by
conducting a set of laboratory tests to assess twenty blended
chloride-based anti-icing formulations. The laboratory data were then
used to establish predictive models correlating the multiple design
parameters with the anti-icer performance and effects or with an
anti-icer composite index. The authors used artificial neural networks
for modeling and examined anti-icer performance (characteristic
temperature and ice-melting capacity at 30 and 15 degrees F (-1.1 and
-9.4 degrees C), respectively) and effects (splitting tensile strength
of concrete after ten freeze-thaw cycles and corrosivity to mild
steel) as a function of the formulation design. The anti-icer
composite index was calculated for four different user priority
scenarios (cost-first, performance-first, impacts-first, or a balanced
approach), each of which placed a different set of decision weights on
various target attributes. Three-dimensional response surfaces were
then constructed to illustrate such predicted correlations and to
guide the direction for formulation improvements. DOI:
10.1061/(ASCE)CR.1943-5495.0000039. (C) 2012 American Society of Civil
Engineers.

2012




* 665(<-296): Prediction of surface roughness and delamination in end milling of GFRP using mathematical model and ANN

Glass fiber reinforced plastics (GFRP) composite is considered to be
an alternative to heavy exortic materials. Accordingly, the need for
accurate machining of composites has increased enormously. During
machining, the reduction of delamination and obtaining good surface
roughness is an important aspect. The present investigation deals with
the study and development of a surface roughness and delamination
prediction model for the machining of GFRP plate using mathematical
model and artificial neural network (ANN) multi objective
technique. The mathematical model is developed using RSM in order to
study main and interaction effects of machining parameters. The
competence of the developed model is verified by using coefficient of
determination and residual analysis. ANN models have been developed to
predict the surface roughness and delamination on machining GFRP
components within the range of variables studied. Predicted values of
surface roughness and delamination by both models are compared with
the experimental values. The results of the prediction models are
quite close with experiment values. The influences of different
parameters in machining GFRP composite have been analyzed.

2012




* 668(<-219): Modeling and optimization of laser beam percussion drilling of thin aluminum sheet

Modeling and optimization of machining processes using coupled methodology has been an area of interest for manufacturing engineers in recent times. The present paper deals with the development of a prediction model for Laser Beam Percussion Drilling (LBPD) using the coupled methodology of Finite Element Method (FEM) and Artificial Neural Network (ANN). First, 2D axisymmetric FEM based thermal models for LBPD have been developed, incorporating the temperature-dependent thermal properties, optical properties, and phase change phenomena of aluminum. The model is validated after comparing the results obtained using the FEM model with self-conducted experimental results in terms of hole taper. Secondly, sufficient input and output data generated using the FEM model is used for the training and testing of the ANN model. Further, Grey Relational Analysis (GRA) coupled with Principal Component Analysis (PCA) has been effectively used for the multi-objective optimization of the LBPD process using data predicted by the trained ANN model. The developed ANN model predicts that hole taper and material removal rates are highly affected by pulse width, whereas the pulse frequency plays the most significant role in determining the extent of HAZ. The optimal process parameter setting shows a reduction of hole taper by 67.5%, increase of material removal rate by 605%, and reduction of extent of HAZ by 3.24%. (C) 2012 Elsevier Ltd. All rights reserved.

2013


* 669(<-220): Modeling and optimization of laser beam percussion drilling of nickel-based superalloy sheet using Nd: YAG laser

The creation of small diameter holes in thin sheets (< 3 mm) of superalloys using a laser beam is a challenging task. Knowledge of the effect of laser related process variables on hole related responses with respect to variation of sheet thickness is essential to obtain a hole of requisite quality. Therefore, in this paper a coupled methodology comprising of Finite Element Method (FEM) and Artificial Neural Network (ANN) has been used to develop a prediction model for the Laser Beam Percussion Drilling (LBPD) process. First, a 2D axisymmetric FEM-based thermal model for LBPD has been developed incorporating temperature-dependent thermal properties, optical properties and phase change phenomena of the sheet material. The developed FEM-based thermal model is validated with self-conducted experimental results in terms of hole taper which is further used to generate adequate input and output data for training and testing of the ANN model. Gray Relational Analysis (GRA) coupled with Principal Component Analysis (PCA) has been effectively used for the multi-objective optimization of the LBPD process utilizing the data predicted by the trained ANN model. The developed ANN model has been used to predict the performance characteristics of the LBPD process. The results predicted by the ANN model show that with the increase in pulse width and peak power the hole taper, material removal rate (MRR) and heat-affected zone (HAZ) increases. The acquired combination of optimal process variables produce a hole with good integral quality, i.e., a reduction of hole taper by 32.1%, increase of material removal rate by 28.9% and reduction of extent of HAZ by 4.5%. (C) 2013 Elsevier Ltd. All rights reserved.

2013


* 670(<-282): Selection of wire electrical discharge machining process parameters using non-traditional optimization algorithms

Selection of the optimal values of different process parameters, such as pulse duration, pulse frequency, duty factor, peak current, dielectric flow rate, wire speed, wire tension, effective wire offset of wire electrical discharge machining (WEDM) process is of utmost importance for enhanced process performance. The major performance measures of WEDM process generally include material removal rate, cutting width (kerf), surface roughness and dimensional shift. Although different mathematical techniques, like artificial neural network, gray relational analysis, simulated annealing, desirability function, Pareto optimality approach, etc. have already been applied for searching out the optimal parametric combinations of WEDM processes, but in most of the cases, sub-optimal or near-optimal solutions have been arrived at. In this paper, an attempt is made to apply six most popular population-based non-traditional optimization algorithms, i.e. genetic algorithm, particle swarm optimization, sheep flock algorithm, ant colony optimization, artificial bee colony and biogeography-based optimization for single and multi-objective optimization of two WEDM processes. The performance of these algorithms is also compared and it is observed that biogeography-based optimization algorithm outperforms the others. (C) 2012 Elsevier B.V. All rights reserved.

2012


* 671(<-387): Optimization of WEDM Process Parameters of gamma-TiAl Alloy Using SVM Method

Wire electrical discharge machining (Wire-EDM) process is a widely used method to produce precision tools. There are many parameters that have influence on the process such as pulse on time, pulse off time, voltage, wire tension, wire diameter, material etc. In certain cases values of the parameters (e.g. cutting speed and surface roughness) conflict with each other. Usually surface quality decreases while cutting speed increases, and vice versa. It is difficult to improve both properties simultaneously. Due to the complexity of process, it is convenient to use some stochastic methods to find optimal process parameters. In this study, Wire-EDM process of gamma titanium aluminide alloy was optimized by support vector machines (SVM) method. To achieve this goal, Wire-EDM experimental results of the alloy were used as training set, and then predictions were made using this set. Obtained results were submitted as graphs and Pareto optimal points were determined among predicted points. Lastly, an optimum point was selected according to desired surface roughness value using multi-objective optimization methodology. Results showed that using SVM is effective as much as traditional prediction methods like Artificial Neural Networks (ANN).

2011


* 672(<-437): An Integrated Approach to Optimization of WEDM Combining Single-Pass and Multipass Cutting Operation

This research article presents an integrated approach to optimization of wire electrical discharge machining (WEDM) of gamma titanium aluminide (-TiAl) with the assistance of artificial neural network (ANN) modeling. Four process parameters, pulse on time, peak current, dielectric flow rate, and effective wire offset, were investigated to study their influence on the process outputs; that is, cutting speed, surface roughness, and dimensional shift in the multipass cutting operation. Two ANN models, based on Bayesian (automated) regularization and early stopping method, have been developed and compared. The model based on Bayesian regularization method was selected because the prediction accuracy was superior compared to the early stopping method. The Pareto optimization was applied to determine the maximum cutting speed corresponding to the required surface roughness for the trim cutting process. Finally, by combining the results of the single- and multipass cutting and introducing the new concept of effective cutting speed, a machining strategy based on the novel concept of critical surface roughness has been developed for selecting the machining process, either single cutting or multipass cutting, so that the maximum productivity can be ensured according to the surface finish requirements.

2010


* 673(<-261): Modeling of wire electrical discharge machining of alloy steel (HCHCr)

This study provides predictive models for the functional relationship between input and output variables of wire cut electrical discharge machine (WEDM) environment using alloy steel (HCHCr). Multi-objective optimization of the process parametric combinations is attempted by modeling WEDM process by use of artificial neural networks (ANN). This work provide an optimized input data set to WEDM system and the results show improvement with better productivity, reduced cutting time and product cost at the cutting speed and surface finish. At experimental result, the surface quality decreases as cutting speed increases and 1.371 mm/ min becomes the maximum cutting speed obtained with good surface finish of 0.387 micron. The results show the potential to improve production efficiency and part quality.

2012





* 674(<-667): An interactive multi-objective artificial neural network approach for machine setup optimization

In this paper, we develop an artificial neural network method for
machine setup problems. We show that our new approach solves a very
challenging problem in the area of machining i.e. machine setup. A
review of machine setup concepts and methods, along with feedforward
artificial neural network is presented. We define the problem of
machine setup to assessing the values of machine speed, feed and depth
of cut (process inputs) for a particular objective such as minimize
cost, maximize productivity or maximize surface finish. We use cutting
temperature, cutting force, tool life, and surface roughness (process
outputs) rather than objective functions to communicate with the
decision maker. We show the relationship between process inputs to
process outputs. This relationship is used in determining machine
setup parameters (speed, feed, and depth of cut). Back propagation
neural network is used as a decision support tool. The network maps,
the forward relationship, and backward relationship between process
inputs and process outputs. This mapping facilitates an interactive
session with the decision maker. The process input is appropriately
selected. Our method has the advantage of forecasting machine setup
parameters with very little resource requirement in terms of time,
machine tool, and people. Forecast time is almost
instantaneous. Accuracy of the forecast depends on training and a well
determined training sample provides very high accuracy. Trained
network replaces the knowledge of an experienced worker, hence labor
cost can be potentially reduced.

2000



* 675(<-190): Parametric study along with selection of optimal solutions in dry wire cut machining of cemented tungsten carbide (WC-Co)

This work deals with parametric study of dry wire EDM (WEDM) process of cemented tungsten carbide. Experiments have been conducted using air as dielectric medium to investigate effects of pulse on time, pulse off time, gap set voltage, discharge current and wire tension on cutting velocity (CV) surface roughness (SR) and oversize (OS). Firstly, a series of exploratory experiments were carried out to identify appropriate gas and its pressure. Afterward, preliminary experiments were conducted to investigate effects of process parameters on dry WEDM characteristics and find appropriate ranges for each factor. Then a central composite rotatable method was employed to design experiments based on response surface methodology (RSM). Empirical models were developed to create relationships between process factors and responses by considering to analysis of variances (ANOVA). To increase the predictability of the process, intelligent models have been developed based on back-propagation neural network (BPNN) and accuracy of these models was compared with mathematical models based on root mean square error (RMSE) and prediction error percent (PEP). In order to select optimal solutions in the cases of single-objective and multi-objectives optimization problems, optimization includes two main approaches. First approach was based on mathematical model and desirability function. Also second approach was designed based on neural network and particle swarm optimization. These approaches were applied in both cases of single-objective and multi-objectives optimization problems and their results were compared with together. Results indicated that selection of air at inlet pressure of 1.5 bar is really appropriate for conducting experiments of next stages. Also, the BPNN creates more accurate prediction rather than mathematical model. Moreover, the BPNN-PSO approach was more efficient in optimization of process rather than mathematical model-desirability function in respect with validation tests. (C) 2013 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.

2013



* 676(<-473): An adaptive neuro-fuzzy inference system (ANFIS) model for wire-EDM

A wire electrical discharge machined (WEDM) surface is characterized
by its roughness and metallographic properties. Surface roughness and
white layer thickness (WLT) are the main indicators of quality of a
component for WEDM. In this paper an adaptive neuro-fuzzy inference
system (ANFIS) model has been developed for the prediction of the
white layer thickness (WLT) and the average surface roughness achieved
as a function of the process parameters. Pulse duration, open circuit
voltage, dielectric flushing pressure and wire feed rate were taken as
model's input features. The model combined modeling function of fuzzy
inference with the learning ability or artificial neural network; and
a set of rules has been generated directly from the experimental
data. The model's predictions were compared with experimental results
for verifying the approach. (C) 2008 Elsevier Ltd. All rights
reserved.

2009


* 677(<-202): Multi-objective optimization of material removal rate and surface roughness in wire electrical discharge turning

Wire electrical discharge turning (WEDT) is an emerging area, and it can be used to generate cylindrical forms on difficult to machine materials by adding a rotary axes to WEDM. The selection of optimum cutting parameters in WEDT is an important step to achieve high productivity while making sure that there is no wire breakage. In the present work, the WEDT process is modelled using an artificial neural network with feed-forward back-propagation algorithm and using adaptive neuro-fuzzy inference system. The experiments were designed based on Taguchi design of experiments to train the neural network and to test its performance. The process is optimized considering the two output process parameters, material removal rate, and surface roughness, which are important for increasing the productivity and quality of the products. Since the output parameters are conflicting in nature, a multi-objective optimization method based on non-dominated sorting genetic algorithm-II is used to optimize the process. A pareto-optimal front leading to the set of optimal solutions for material removal rate and surface roughness is obtained using the proposed algorithms. The results are verified with experiments, and it is found to improve the performance of WEDT process. Using this set of solutions, required input parameters can be selected to achieve higher material removal rate and good surface finish.

2013



* 678(<-305): Intelligent optimization and selection of machining parameters in finish turning and facing of Inconel 718

The heat-resistant super alloy material like Inconel 718 machining is
an inevitable and challenging task even in modern manufacturing
processes. This paper describes the genetic algorithm coupled with
artificial neural network (ANN) as an intelligent optimization
technique for machining parameters optimization of Inconel 718. The
machining experiments were conducted based on the design of
experiments full-factorial type by varying the cutting speed, feed,
and depth of cut as machining parameters against the responses of
flank wear and surface roughness. The combined effects of cutting
speed, feed, and depth of cut on the performance measures of surface
roughness and flank wear were investigated by the analysis of
variance. Using these experimental data, the mathematical model and
ANN model were developed for constraints and fitness function
evaluation in the intelligent optimization process. The optimization
results were plotted as Pareto optimal front. Optimal machining
parameters were obtained from the Pareto front graph. The confirmation
experiments were conducted for the optimal machining parameters, and
the betterment has been proved.

2012


* 679(<-494): Development of multi-objective optimization models for electrochemical machining process

Owing to the complexity of electrochemical machining (ECM), it is very difficult to determine optimal cutting parameters for improving cutting performance. Hence, optimization of operating parameters is an important step in machining, particularly for unconventional machining procedures like ECM. A suitable selection of machining parameters for the ECM process relies heavily on the operator's technologies and experience because of their numerous and diverse range. Machining parameters provided by the machine tool builder cannot meet the operator's requirements. Since for an arbitrary desired machining time for a particular job, they do not provide the optimal conditions. To solve this task, multiple regression model and ANN model are developed as efficient approaches to determine the optimal machining parameters in ECM. In this paper, current, voltage, flow rate and gap are considered as machining parameters and metal removal rate and surface roughness are the objectives. Then by applying grey relational analysis, we calculate the grey grade for representing multi-objective model. Multiple regression model and ANN model have been developed to map the relationship between process parameters and objectives in terms of grade. The experimental data are divided into training and testing data. The predicted grade is found and then the percentage deviation between the experimental grade and predicted grade is calculated for each model. The average percentage deviations for the training data of the linear regression model, logarithmic transformation model, excluding interaction terms and ANN model, are 12.7, 25.6 and 3.03, respectively. The average percentage deviations for the testing data of the three models are 9.83, 26.8 and 2.67. While examining the average percentage deviations of three models, ANN is having less percentage deviation. So ANN is considered as the best prediction model. Based on the testing results of the artificial neural network, the operating parameters are optimized. Finally, ANOVA is used to identify the significance of multiple regression model and ANN model.

2008


* 680(<-285): Multi-objective optimization of electric-discharge machining process using controlled elitist NSGA-II

Parametric optimization of electric discharge machining (EDM) process is a multi-objective optimization task. In general, no single combination of input parameters can provide the best cutting speed and the best surface finish simultaneously. Genetic algorithm has been proven as one of the most popular multi-objective optimization techniques for the parametric optimization of EDM process. In this work, controlled elitist non-dominated sorting genetic algorithm has been used to optimize the process. Experiments have been carried out on die-sinking EDM by taking Inconel 718 as work piece and copper as tool electrode. Artificial neural network (ANN) with back propagation algorithm has been used to model EDM process. ANN has been trained with the experimental data set. Controlled elitist non-dominated sorting genetic algorithm has been employed in the trained network and a set of pareto-optimal solutions is obtained.

2012


* 681(<-542): Modeling of electrical discharge machining process using back propagation neural network and multi-objective optimization using non-dominating sorting genetic algorithm-II

Present study attempts to model and optimize the complex electrical discharge machining (EDM) process using soft computing techniques. Artificial neural network (ANN) with back propagation algorithm is used to model the process. As the output parameters are conflicting in nature so there is no single combination of cutting parameters, which provides the best machining performance. A multi-objective optimization method, non-dominating sorting genetic algorithm-II is used to optimize the process. Experiments have been carried out over a wide range of machining conditions for training and verification of the model. Testing results demonstrate that the model is suitable for predicting the response parameters. A pareto-optimal set has been predicted in this work. (c) 2006 Elsevier B.V. All rights reserved.

2007


* 682(<-316): Intelligent Modeling and Multiobjective Optimization of Die Sinking Electrochemical Spark Machining Process

Die sinking-electrochemical spark machining (DS-ECSM) is one of the hybrid machining processes, combining the features of electrochemical machining (ECM) and electro-discharge machining (EDM), used for machining of nonconducting materials. This article reports an intelligent approach for the modelling of DS-ECSM process using finite element method (FEM) and artificial neural network (ANN) in integrated manner. It primarily comprises development of two models. The first one is the development of a thermal finite element model to estimate the temperature distribution within the heat-affected zone (HAZ) of single spark on the workpiece during DS-ECSM. The estimated temperature field is further post-processed for determination of material removal rate (MRR) and average surface roughness (ASR). The second one is a back propagation neural network (BPNN)-based process model used in a simulation study to find optimal machining parameters. The BPNN model has been trained and tested using the data generated from the FEM simulations. The trained neural network system has been used in predicting MRR and ASR for different input conditions. The ANN model is found to accurately predict DS-ECSM process responses for chosen process conditions. This article also presents an effective approach for multiobjective optimization of DS-ECSM process using grey relational analysis.

2012


* 683(<-435): Modelling and Optimization of Multiple Process Attributes of Electrodischarge Machining Process by Using a New Hybrid Approach of Neuro-Grey Modeling

In the present article, a new hybrid approach of neuro-grey modeling (NGM) technique has been proposed for modeling and optimization of multiple process attributes of the electro discharge machining (EDM) process. It is proposed to simulate through an artificial neural network (ANN) for characterization of multiple process attributes followed by multiple process attributes optimization by using grey relational analysis (GRA) technique. A multineuron ANN of logistic sigmoid activation function has been designed. Levenberg-Marquardt algorithm involving second order error optimization has been chosen for training of the ANN because of its inherent merits. Then, using grey relational analysis (GRA) technique, a grey relational grade has been determined, which effectively represents the aggregate of different process attributes. As a result, a multi-attribute optimization can be converted into optimization of a single grey relational grade. The ANN is simulated first to characterize surface roughness (Ra), depth of heat-affected zone, microhardness value of machined surface, and material removal rate (MRR) with respect to current and pulse duration. Then, optimal values of current and pulse duration have been obtained. The NGM technique is found to be better and easy to implement.

2010


* 684(<-188): Multiobjective optimization of slotted electrical discharge abrasive grinding of metal matrix composite using artificial neural network and nondominated sorting genetic algorithm

The alternative use of electrical discharge grinding and abrasive grinding, which is applied with the application of slotted wheel named as slotted electrodischarge abrasive grinding, is much suitable for machining of metal matrix composites. But the selection of process parameters is a difficult task due to the complexity of the process. The aim of this study is to optimize the process parameters of slotted electrodischarge abrasive grinding process using a combined approach of artificial neural network and nondominated sorting genetic algorithm II. The artificial neural network architecture has been trained and tested with experimental data, and then the developed model is coupled with nondominated sorting genetic algorithm II to develop a hybrid approach of artificial neural network-nondominated sorting genetic algorithm II, which is used for optimization of process parameters. During experimentation, the effect of current, pulse on-time, pulse off-time, wheel speed and grit number has been studied on material removal rate and average surface roughness (Ra). The results have shown that prediction capability of artificial neural network model is within the range of acceptable limits. The developed hybrid approach of artificial neural network-nondominated sorting genetic algorithm II gives optimal solution with correlation coefficient of material removal rate and Ra as 0.9979 and 0.9982, respectively.

2013


* 685(<-195): Intelligent Modeling and Multiobjective Optimization of Electric Discharge Diamond Grinding

The grinding of metal matrix composites (MMCs) is very difficult by conventional techniques due to its improved mechanical properties. It often results in poor surface quality (surface damage) in the form of surface cracks/residual stresses and requires frequent truing and dressing due to clogging of the grinding wheel. The electric discharge diamond grinding (EDDG), a hybrid process of electric discharge machining and grinding may overcome these problems up to some extent. But low material removal rate (MRR) and high wheel wear rate (WWR) are the main problems in EDDG to achieve economic performance. The present paper investigates the EDDG process performance during grinding of copper-iron-graphite composite by modeling and simultaneous optimization of two important performance characteristics such as MRR and WWR. A hybrid approach of artificial neural network, genetic algorithm, and grey relational analysis has been proposed for multi-objective optimization. The verification results show considerable improvement in the performance of both quality characteristics.

2013


* 686(<-526): Parameter optimization model in electrical discharge machining process

Electrical discharge machining (EDM) process, at present is still an experience process, wherein selected parameters are often far from the optimum, and at the same time selecting optimization parameters is costly and time consuming. In this paper, artificial neural network (ANN) and genetic algorithm (GA) are used together to establish the parameter optimization model. An ANN model which adapts Levenberg-Marquardt algorithm has been set up to represent the relationship between material removal rate (MRR) and input parameters, and GA is used to optimize parameters, so that optimization results are obtained. The model is shown to be effective, and MRR is improved using optimized machining parameters.

2008


* 687(<-477): Genetic algorithms based multi-objective optimization of an iron making rotary kiln

Industrial rotary kilns used in iron making are complex reactors having several functions. Raw materials, like iron ore and non-coking coal, are continuously fed whilst product sponge iron is continuously discharged from the downstream end, while the waste gases in counter current flow, exit through the uphill end. The outputs exhibit conflicting trends at the production level - an increase in daily production results in a decrease in the product's metallic iron content and vice versa. The optimization of the operation is thus a typical case of multi-objective optimization within constraints. The relationship between the various inputs and the above outputs, being very complex, is established by Artificial Neural Networks (ANN). As the search spaces for the inputs are not very well defined for the acceptable ranges of each of the outputs, the optimization task was carried out using multi-objective genetic algorithms and the resulting Pareto fronts are further analyzed. The results conform to the existing trends and also suggest some possible improvements. (C) 2008 Elsevier B.V. All rights reserved.

2009