Differential evolution using local search for multi-objective optimization

Abstract

Differential evolution has the characteristics of fast convergence, less parameters, and ease of implementation. This paper proposes an enhanced DE using the local search for multi-objective optimization, which is called DEMOLS. In DEMOLS, two candidate mutation variants are randomly chosen to enhance the search ability by taking their advantages and strengths and two local search mechanisms are designed to improve the ability of local adjustment. Numerical experiments are performed on a set of multi-objective optimization problems, and the experimental results show that DEMOLS has the ability to solve multi-objective optimization problems. © 2012 Springer-Verlag.