Ranking-Based Elitist Differential Evolution for Many-Objective Optimization
Jing Xiao, Kejun Wang · 2013
In this paper, a novel evolutionary algorithm for many-objective optimization is proposed. The algorithm adopts a new global ranking method to favor convergence and an improved crowding distance to maintain diversity, new elitist selection strategy Based on fitness evaluation is also designed to guide the search towards a representative approximation of the Pareto-optimal front. In order to validate the proposed algorithm, we perform a comparative study where three state-of-the-art representative approaches are considered. In such a study, a well-known scalable test problem is adopted as well as six different problem sizes, ranging from 3 to 8 objectives. Experimental results prove that our proposed algorithm is highly effective for many-objective problems in comparison to existing multi-objective evolutionary algorithms.