A Study of Fitness Inheritance and Approximation Techniques for Multi-Objective Particle Swarm Optimization

Abstract

In this paper, we study the use of fitness inheritance and approximation techniques to reduce the number of fitness evaluations into a PSO-based multi-objective algorithm previously proposed by the authors. Fifteen fitness inheritance techniques and four approximation techniques are applied to a set of four well-known test functions taken from the multi-objective optimization literature. A comparison of the best techniques found against other PSO-based multi-objective approaches is carried out using other test functions. The obtained results show a good performance of the enhancement techniques proposed.