三态协调搜索多目标粒子群优化算法

提出一种三态协调搜索多目标粒子群优化算法. 该算法提出的三态指导粒子选择策略可以很好地协调算法的局部和全局搜索能力, 且算法改进了传统的外部档案保存机制, 同时引入3 种突变因子, 使获得的非劣解具有更好的分散性. 通过对标准测试函数的求解, 并与其他经典多目标优化算法比较, 表明了新算法在收敛性和多样性方面均有较大的优越性. 最后分析了区域划分系数对所提出算法性能的影响.

Multi-objective particle swarm optimization algorithm based on three status coordinating searching

A multi-objective particle swarm optimization algorithm based on three status coordinating searching(TC-MOPSO) is presented. The three status strategy proposed for choosing guides is useful to coordinate local and global search capability. The traditional external archives update mechanism is improved and three kinds of mutation factors are introduced, which contribute to making the Pareto solutions have a better distribution. By solving several standard test functions and comparing with three classical multi-objective optimization algorithms, it is proved that the new algorithm has high competition in terms of convergence and diversity metrics. Finally, the influence of the regionalism coefficient on the performance of the proposed algorithm is analyzed.