I. Introduction
Multi-objective optimization problems (i.e., a class of optimization problems with two or more conflicting objectives) are ubiquitous in many real-world applications [1]-[3]. Evolutionary multi-objective optimization (EMO) algorithms have been well recognized as one of the effective approaches to solve multi-objective optimization problems [4]. The population-based search nature of EMO algorithms offers the advantage of finding a set of Pareto optimal solutions in a single run. Since a single run of EMO algorithms can find multiple optimal solutions, the main goal of EMO algorithms is to obtain a set of Pareto optimal solutions that are well-distributed over the true Pareto front.