Solving Computationally Complex Multiobjective Optimization Problems:
Experiences on Industrial Optimization

Vesa Ojalehto

In the modern competition based society, different industries are
increasingly depending on optimization to find new advantages for
their commercial ventures. Due to the complexity of real world
applications, optimization problems solved in different industries
are typically very complex and should be considered in their
entireness, instead of a part at a time. Therefore, industry
problems typically have  several, conflicting objectives and be
computationally very demanding. However, current optimization
methods are usually suitable for only single objective
optimization, and/or they can be applied only to a narrow set of
demanding problems.

In this research, previous experiences obtained from solving
several optimisation problems in collaboration with different
industrial partners are used to determine the main obstacles of
solving complex optimization problems. In order to overcoming
these obstacles, we propose a multiobjective optimization
platform, implementing several different multiobjective methods.
With this platform, an industry engineer can solve complex
multiobjective optimization problems without deep knowledge of
optimisation topics.