A Multi-Objective Evolutionary Algorithm based on complete-linkage clustering to enhance the solution space diversity
Kamyab Tahernezhad, Kimia Bazargan Lari, Ali Hamzeh, Sattar Hashemi · 2012
Multi-Objective Evolutionary Algorithm (MOEA) is a leader framework to solve multi-objective optimization problems due to its capability of obtaining a set of compromise solutions in a single run. Most of MOEAs try to converge to the Pareto optimal front in purpose of maintaining the population diversity in the objective space. Here, we are going to present a novel MOEA for enhancing the population diversity of non-dominated vectors in the solution space. In this paper, a novel approach, which is inspired from geometrical information of candidate solutions, is proposed to adopt the innovative clustering-based scheme during the optimization cycle. This approach intends to obtain more diverse and well-distributed non-dominated vectors (i.e. Pareto-set) in the solution space. The present work is applied to a wide range of well established test problems. The obtained results validate the motivation on the basis of diversity and performance measures in comparison to state of the art algorithms.