群体组织启示下的粒子群算法改进

从社会学的视角分析粒子群的组织结构，试图通过增强群体组织管理来防止算法的早熟收敛．借鉴社会管理学中发挥个体能动性和规范成员行为并重的管理理念，构造基于团队式管理的粒子群算法．使用标准测试函数对该算法进行仿真实验，并用正交试验法进行了参数优化．实验结果表明，相对现有的一些改进方案，文中提出的改进算法简单易实现，在一些问题求解上表现出较好的寻优和收敛性能．

Modified particle swarm optimization inspired by group organization

The organizational structure of PSO was analyzed from the view of sociology, and for preventing from premature convergence, a modified approach to enhance organizational management in-group was proposed. Concerning with the sociological management philosophy in which motivating individual initiative must be combined with regulating the behavior of members in society, a modified PSO algorithm based on team-management theory was proposed. Simulation experiments were performed with several benchmark functions, and the orthogonal experiment method was used for parameter selection. Compared PSO with some other improvement strategies by experiment, the team-management based modified PSO algorithm is very simple and easy to operate and shows better performances on optimization and convergence in solving problems.