基于动态和谐搜索的混合粒子群优化算法  

Hybrid particle swarm optimization algorithm based on dynamic harmony search

作　　者：李炜[1] 刘全银[1] 王凯东 

机构地区：[1]兰州理工大学电气工程与信息工程学院,甘肃兰州730050 [2]赛鼎工程有限公司,山西太原030006

出　　处：《兰州理工大学学报》2009年第4期74-77,共4页Journal of Lanzhou University of Technology

基　　金：甘肃省自然科学基金(3ZS051-A25-032);甘肃省高等学校研究生导师基金(050301)

摘　　要：针对粒子群优化算法随维数增大群体多样性相对减小而早熟收敛的问题,在对和谐搜索算法进行适应性改进的基础上,将其引入粒子群算法中,提出一种动态和谐搜索混合粒子群优化算法(DHSPSO).该方法使得粒子在搜索初期更具遍历性,降低算法对初始值的敏感性,并通过和谐搜索算法搜索的随机性和优胜劣汰机制改善粒子群的多样性,使得算法具有更快的收敛速度与更好的全局搜索能力.以多个标准测试函数优化进行仿真测试,结果表明,DHSPSO算法在进行高维优化问题时,在寻优速度、精度和成功率等方面均显示出良好的优化效果.Aimed at the premature convergence problem with the dimension increasing and diversity of swarm relative decreasing in PSO,an improved harmony search algorithm was introduced to particle swarm optimization（PSO）.A hybrid particle swarm optimization algorithm（DHSPSO） based on the improved harmony search algorithm was proposed.At the initial searching stage,the DHSPSO exhibited better ergodicity and reduced the sensitivity to initial value.The diversity of PSO was improved through the random searching and mechanism for the survival of the fittest of HS, making the algorithm of greater converging speed and strong global optimization ability. The DHSPSO was applied to the high-dimensional optimization problems and simulation results showed that DHSPSO manifested better optimization results such as greater optimization speed, accuracy, and successfulness ratio of optimization.

关 键 词：粒子群优化 动态和谐搜索 高维函数优化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]