一种具有速度扰动的高斯学习粒子群优化算法  

Velocity Disturbance Based Particle Swarm Optimization with Gaussian Learning

作　　者：王芸[1] 孙辉[1] 

机构地区：[1]南昌工程学院信息工程学院,南昌330099

出　　处：《小型微型计算机系统》2015年第7期1521-1525,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61261039;61305150)资助;教育部人文社科青年基金交叉项目(13YJCZH174)资助;江西省教育厅落地计划项目(KJLD13096)资助;江西省科技厅自然科学基金项目(20122BAB201043)资助;南昌工程学院青年基金项目(2010KJ019)资助

摘　　要：针对标准粒子群(PSO)算法在复杂问题上收敛速度慢和早熟收敛的缺点,提出一种基于速度扰动的高斯学习粒子群优化算法(PGPSO).新算法中,首先在速度更新公式中添加速度扰动项,使得每次迭代进化时粒子速度增量比标准PSO更大,一方面加快了算法的收敛速度,另一方面又减缓了粒子速度快速降低的趋势,有效地维持了种群的多样性;同时引入高斯学习的概念,当算法陷入局部最优时,对全局最优粒子在搜索空间进行高斯学习,以增强算法逃离局部最优的能力.基准测试函数的实验结果表明,相较一些国际上知名的粒子群算法,新算法不仅能提高收敛速度、增强全局搜索能力,而且能有效提高解的精度和稳定性.The standard PSO suffers from the premature convergence problem and the slowconvergence speed problem when solving complex optimal problems,so the article proposes PSO with velocity disturbance and Gaussian learning（ PGPSO）. Firstly,a velocity disturbance item will be introduced to the formula of velocity updating such that it obtains bigger velocity increment than that of the standard PSO in every round of iterative evolution,which can accelerate the algorithm ＇s convergence speed and slowdown the trend of the rapidly decreasing particle＇ s speed so as to help maintain the population＇ s diversity effectively. M oreover,if the algorithm is trapped into local optima,gaussian learning strategy will be conducted to help the global best particle jump out of local optimal regions. Compared with some famous improved PSO algorithms,the experimental results on benchmark test functions showthat PGPSO can enhance the performance of the PSO paradigm,in terms of convergence speed,global optimality,solution accuracy and algorithm reliability.

关 键 词：粒子群优化算法 速度扰动 高斯学习 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]