Integrated pre-design step methodology based on multi-objective evolutionary optimization

In this paper we propose an integrated pre-design step framework using multiobjective evolutionary optimization and a decision support tool. The tailored genetic algorithm relies on specific fitness function which enables to deal with a high number of objectives. Moreover, surrogate models have been integrated so as to speed up objective functions evaluations which are usually expensive in case of mechanical product pre-design step. An automatic post-treatment of Pareto optimal solutions is proposed in order to synthesize a large multidimensional database into a restricted number of typings. This latter step is of particular importance since it affords designer a powerful decision support tool.