Improvement of Differential Evolution Multiobjective Optimization Algorithm Based on Decomposition

Multi-objective optimization problem (MOP) is a challenging field of scientific research in real-life. The effective way to solve multi-objective optimization problems is Multi-objective Evolutionary Algorithm (MOEA). In this paper, enhancements to a Multi-objective Evolutionary algorithm MOEA/D-DE are proposed. The proposed improvement points help to improve both population distribution and algorithmic local search capabilities. In an existing study, in order to better distribute the population, the Monte Carlo method was used for population initialization. Adaptive differential evolution operators are used to improve the local search ability of the algorithm. The algorithm was tested on widely used ZDT and DTLZ family test problems. The experimental results show that the proposed algorithm is better than MOEA/D-DE and has better performance than other excellent multi-objective algorithms.