一种基于欧氏干扰因子的微粒群优化算法的改进

An Improvement of Particle Swarm Optimization Algorithm with Euclidean Distance Based Interference Factor

针对基本微粒群优化(PSO,Particle Swarm Optimization)算法在应用于具有极多局部极值和维数被优化问题时易陷入局部最优和早熟收敛的不足,提出了一种新的改进算法称之为欧氏微粒群算法.此改进算法的主要思想是当算法陷入局部最优时,给微粒一个扰动因子,它的大小会因当前微粒与全局最优微粒的欧式距离的大小而自适应变化,促使微粒跳出局部最优.在实验中选取典型标准函数对算法进行测试,实验结果表明,本文算法优于标准微粒群算法(SPSO)和高斯微粒群算法(GPSO),而且随着问题复杂性的提高其性能优越性越明显.

This paper develops a new improved particle swarm optimization （PSO） algorithm named Euclidean PSO（EPSO） to solve the problems such as the insufficiency of local optima and premature convergence when PSO used in the issues of rich local extremum and dimension optimized.The main improvement of the algorithm is to develop an interference factor for the particles.The value of interference factor will be self-adaptive according to the Euclidean distance between the current particle and the global best particle.And it has confirmed its excellent performance in benchmark functions compared with Standard PSO （SPSO） and Gaussian PSO（GPSO）.