Handling uncertainty in evolutionary multiobjective optimization: SPGA

This paper presents an extension of the previously developed approach to solve multiobjective optimization problems in deterministic environments by incorporating a stochastic Pareto-based solution ranking procedure. The proposed approach, called stochastic Pareto genetic algorithm (SPGA), employs some statistical analysis on the solution dominance in stochastic problem environments to better discriminate among the competing solutions. Preliminary computational results on three published test problems for different levels of noise with SPGA and NSGA-II are discussed.