Abstract

It has been already documented the fact that estimation of distribution algorithms suffer from loss of population diversity and improper treatment of isolated solutions. This situation is particularly severe in the case of multi-objective optimization, as the loss of solution diversity limits the capacity of an algorithm to explore the Pareto-optimal front at full extent. A set of approaches has been proposed to deal with this problem but —to the best of our knowledge— there has not been a comprehensive comparative study on the outcome of those solutions and at what degree they actually solve the issue. This paper puts forward such study by comparing how current approaches handle diversity loss when confronted to different multi-objective problems.