Abstract: A new Multi-Objective Particle Swarm Optimization（MOPSO） algorithm is proposed based on multistrategy to avoid the problems about convergence rate of optimization algorithm and the non-inferior solution diversity cannot be satisfied simultaneously. It divides the external species and evolution population again by the non-dominated sorting and crowded distance ordering strategy. It chooses the best particle as a leader in the external populations to lead the evolution of the evolution of particles in the population by the niche selection strategy. In the evolutionary population, it balances the global searching and local accurate searching algorithm through the multi-scale Gaussian mutation strategy. It adopts the updating strategy of field cognitive individual extremum to update the extreme value of the individual constantly. The algorithm is applied to a typical multi-objective test function and compared to other multi-objective optimization algorithms. It is shown from the results that the four strategies of the algorithm are effective and complementary to each other and shown that the algorithm not only has good convergence and speed, but also has good uniformity and diversity at the same time.

Key words: multi-objective optimization, particle swarm optimization algorithm, multi-strategy, field of cognitive

摘要： 针对目前多目标粒子群优化算法的收敛性能和非劣解的多样性不能同时得到满足等缺陷，提出一种基于多策略的多目标粒子群优化算法（Multi-Objective Particle Swarm Optimization algorithm for Multi-Strategy，MS-MOPSO）。采用非支配排序和拥挤距离排序相结合策略，重新划分外部种群和进化种群；采用小生境选择策略，在外部种群中选择最佳粒子作为领导粒子，用于领导进化种群中粒子的进化；在进化种群中利用多尺度高斯变异策略，平衡算法的全局搜索和局部精确搜索；采用邻域认知个体极值更新策略，不断更新个体极值。将该算法应用到典型的多目标测试函数，并与其他多目标优化算法进行对比分析，测试结果表明该算法中四个策略的有效性和互补性，同时验证了该算法不但具有较好的收敛性和收敛速度，而且该算法最优解的分布具有良好的均匀性和多样性。

关键词: 多目标优化, 粒子群算法, 多策略, 邻域认知