A Multi-objective Evolutionary Algorithm Based on Second-Order Differential Operator.

Differential evolution (DE) is a swarm intelligence algorithm based on population, which has been used to solve multi-objective optimization problems (MOP). The distribution and convergence of the non-dominated solutions set are often the key indicators to evaluate the merits of MOP algorithms. In this paper, we propose a decomposition multi-objective evolutionary algorithm based on second-order differential operator (MOEA/D-SODE). By selecting the commonly used ZDT, DTLZ and IMOP benchmark functions, comparing with the existing differential evolution MOEA/D-DE, the solutions set obtained by MOEA/D-SODE algorithm has better convergence and distribution. The experimental results verify the effectiveness of MOEA/D-SODE algorithm, which provides a new and effective method for MOP.