A constraint multi-objective artificial physics optimization algorithm

Abstract

The use of evolutionary algorithms to solve unconstraint multi-objective problems (MOPs) has attracted much attention recently. However, research on constraint multi-objective algorithms is relatively less. The authors introduce a novel evolutionary paradigm of artificial physics optimization (APO) into constraint multi-objective optimization domain and modify the original mass function and virtual force rules in order to fit constraint multi-objective optimization problems. Moreover the authors present a method of virtual force decreasing to improve the efficiency. Finally, simulation tests show that the algorithm is effective.