In order to solve premature phenomenon and slow convergence problems in particle swarm algorithm, an improved mean particle swarm algorithm is provided. The algorithm apply to the nonlinear weight. At the same time in each iteration step, the history optimal particle and the global optimal population were averaged and multiplied a nonlinear weight so that it improve the global search capacity and convergence speed. Through four standard function test, The results show that the effectiveness of the proposed algorithm.%为了解决粒子群算法存在“早熟”现象和收敛速度慢的问题，本文提出一种改进的均值粒子群算法。该算法采用非线性惯性权重，同时在每个迭代步，将粒子历史最优和种群全局最优取均值再乘以一个非线性权重的方法，以提高算法的全局搜索能力和收敛速度。通过4个标准函数的测试，实验结果表明该算法的有效性。
展开▼
