Abstract: The standard particle swarm optimization algorithm（PSO） shows a bad performance when optimizing the multimodal and higher dimensional functions.A new formal particle swarm optimization（MPSO） is advanced，which replaces the global best place （p_g） by the center of all individual best places（p_mean）.So，the colonial diversity is increased，and the pre-mature convergence is avoided to some degree.At the same time，the concise iterative formulation is kept.The simulations of 3 testing functions show that the MPSO has better ability to find the global optimum solution than the standard particle swarm optimization algorithm.

Key words: particle swarm optimization, pre-mature convergence, function optimization

摘要： 标准微粒群算法在优化多峰、多维的复杂函数时，其效果并不理想，容易早熟收敛。为了改进微粒群算法处理此类问题的性能，提出了一种新的微粒群算法。该算法将标准微粒群算法迭代公式中的群体最优位置用个体最优位置的中心代替，有利于增强群体的多样性，避免早熟收敛，同时保持了迭代公式的简洁形式。3个常用测试函数的数值模拟表明，新的微粒群算法较标准微粒群算法在寻优能力上有明显的提高。

关键词: 微粒群算法, 早熟收敛, 函数优化