Parallel multiobjective evolutionary algorithms

Abstract

The use of evolutionary algorithms (EAs) for solving multiobjective optimization problems has been very active in the last few years. The main reasons for this popularity are their ease of use with respect to classical mathematical programming techniques, their scalability, and their suitability for finding trade-off solutions in a single run. However, these algorithms may be computationally expensive because (1) many real-world optimization problems typically involve tasks demanding high computational resources and (2) they are aimed at finding a whole front of optimal solutions instead of searching for a single optimum. Parallelizing EAs emerges as a possible way of reducing the CPU time down to affordable values, but it also allows researchers to use an advanced search engine-the parallel model-that provides the algorithms with an improved population diversity and enable them to cooperate with other (eventually nonevolutionary) techniques. The goal of this chapter is to provide the reader with an upto-date review of the recent literature on parallel EAs for multiobjective optimization.