中文摘要：

传统粒子群优化算法PSO（Particle Swarm Optimization）概念简单,适应性强,但存在早熟等问题.本文提出了新的基于搜索空间划分（Search Space Division）和Sharing函数的智能分布粒子群优化算法（SDSIR-PSO）.创新点包括：（1）保优的重布粒子算法;（2）引入Sharing函数阻止重分布的粒子陷入同一局部最优;（3）划分搜索空间,子空间中寻优,再优中选优,作全局最优.通过对典型测试函数的详细测试验证了新算法的有效性,在相同条件下较传统算法的解精度提高了80%以上,并有效避免了早熟,提高了收敛速度.

英文摘要：

Traditional PSO algorithm may lead to prematurity by local optimum trap. This paper presents a novel, intelligent redistributed PSO based on search space partition and sharing function, named SDSIR-PSO. The main innovation of this paper include, （1） holding the best global solution while redistribute particles. （2） Introduces the Sharing function to avoid plunging same local trap. （3） Divides search space into sub-spaces and selects the best value as global optimum among optima selected from sub-spaces. The effectiveness of the new algorithm is proved by several well-known benchmark functions.