A Classification-based Mixture-of-Kriging Assisted Evolutionary Algorithm for Expensive Many-objective Optimization

Abstract

Surrogate-assisted evolutionary algorithms (SAEAs) have been widely used to solve expensive optimization problems (EOPs). However, most studies only focus on solving single or multiobjective EOPs. The study of using SAEAs to solve many-objective EOPs has not received much attention. To fill this research gap, this paper presents a new SAEA by using mixture-of-Kriging as a surrogate to approximate the objective values in many-objecitve EOPs. In this algorithm, a fitness-based classification method is employed for choosing data to train the models. Experimental results demonstrate that the proposed algorithm is very promising in performance comparison with the state-of-the-art SAEAs on a number of benchmark problems.