Parallelization and surrogate modelling
in multi objective optimization

About

We daily need to solve the problems that have several objectives that are typically contradictory. For solving such problems, a multi objective optimization techniques are often being used. One of the main problems of multi objective optimization is time complexity needed by the algorithms, since they require more problem evaluations than single objective optimization approaches. This special session will concentrate on reducing this problem by using two established techniques: parallelization and/or surrogate modeling. While the later allows faster simulation and assessment of the optimization objectives, the former allows faster calculation of objectives. If properly used they can be an efficient tool to speed up the search process and make multi objective optimization more usable in real world scenarios.