A New Hyper-Heuristic Multi-Objective Optimisation Approach Based on MOEA/D Framework

Abstract

A multi-objective evolutionary algorithm based on decomposition (MOEA/D) serves as a robust framework for addressing multi-objective optimization problems (MOPs). However, it is widely recognized that the applicability of a fixed offspring-generating strategy in MOEA/D can be limited, despite its foundation in the MOEA/D methodology. Consequently, hybrid algorithms have gained popularity in recent years. This study proposes a novel hyper-heuristic approach that integrates the estimation of distribution (ED) and crossover (CX) strategies into the MOEA/D framework based on the view of successful replacement rate (SSR) and attempts to explain the potential reasons for the advantages of hybrid algorithms. The proposed approach dynamically switches from the differential evolution (DE) operator to the covariance matrix adaptation evolution strategy (CMA-ES) operator. Simultaneously, certain subproblems in the neighbourhood denoted as B(i) employ the Improved Differential Evolution (IDE) operator to generate new individuals for balancing the high evaluation costs associated with CMA-ES. Numerical experiments unequivocally demonstrate that the suggested approach offers distinct advantages when applied to a three-objective test suite. These experiments also validate a significant enhancement in the efficiency (SRR) of the DE operator within this context. The perspectives and experimental findings, with a focus on the Success Rate Ratio (SRR), have the potential to provide valuable insights and inspire further research in related domains.

Optimizing the Three-Dimensional Multi-Objective of Feeder Bus Routes Considering the Timetable

“…NSGA-II [50], MOPSO [51], SPEA2 [52], MOEA/D [53], NSGA-III [54], and INSGA-II, respectively, are used in the MATLAB 2022a environment to solve the Pareto front solutions of 6 biased constrained test functions, including ZDT1, ZDT2, ZDT3 and DTLZ1, DTLZ2, DTLZ3 [55], etc., to verify the feasibility of the algorithms. Each algorithm has been configured with a population size of 100 and a maximum number of 500 iterations.…”

“…In this section, NSGA-II [50], MOPSO [51], SPEA2 [52], MOEA/D [53], NSGA-III [54], and INSGA-II are used to solve the test functions with full constraints on the independent variables, respectively, the test functions are shown in Table 3, and all objective functions are to be minimized. The Pareto fronts of INSGA-II are shown in Figures 11 and 12.…”