Multi-objective evolutionary algorithms based on decomposition (MOEA/D) has demonstrated excellent performance in dealing with multi-objective optimization problems. As two essential issues in MOEA/D, mating neighborhood sizes and reproduction operators determine the exploitation and exploration abilities of the algorithm. This paper proposes a new decomposition-based multi-objective evolutionary algorithm with mating neighborhood sizes and reproduction operators adaptation (MOEA/D-ATO), which adaptively assigns the suitable combination of mating neighborhood size and reproduction operator to each subproblem at different searching stages. Numerical results indicate that the proposed adaptation is effective. Moreover, comparison with other adaptive steady-state MOEA/D variants and state-of-art generational MOEA/D variants shows that the proposed MOEA/D-ATO algorithm performs significantly better in terms of solution quality and CPU time.

EB00000002751K

Available