摘要
针对粒子群优化算法容易陷入局部极值,进化后期收敛速度慢、精度低等缺点,本文将粒子群优化算法与遗传算法相结合,在基本粒子群优化算法中引入了正态变异算子,提出了一种新的混合进化算法,新算法增加了种群的多样性,增强了算法的全局寻优能力,提高了算法的搜索效率。使用新算法对经典函数进行优化测试,结果表明,本算法保持了粒子群优化算法简捷快速、容易实现的特点;同时,正态变异算子的引入提升了算法后期的收敛速度与全局搜索能力。新的算法能够以更小的种群数和进化代数获得较好的优化能力,在克服陷入局部最优和收敛速度方面均优于基本粒子群优化算法、遗传算法以及加入混沌扰动的粒子群优化算法(CPSO)。
The basic Particle Swarm Optimization （bPSO） algorithm suffers from some defects,such as the tendency to converge into a local extremum,the slow convergence rate and the low convergence accuracy in the late stage of evolution.A new algorithm HPSO based on hybrid PSO-GA （Particle Swarm Optimization and Genetic Algorithm） is proposed in this paper.The normal mutation operator is introduced into the basic particle swarm optimization algorithm.By taking advantage of the searching abilities of these two methods,the population diversity is enhanced;the global search ability and search efficiency are improved.The new HPSO is used in several typical function optimizations,and it is shown that the proposed method,while retaining the advantages of bPSO,such as the ease to realize and operate and high speed in calculation,with the introduction of the normal mutation operator,greatly improves the search ability and search efficiency in the late stage of evolution.The new Hybrid algorithm enjoys higher optimization capability with less particles and less generations than bPSO,GA and CPSO.
出处
《科技导报》
CAS
CSCD
北大核心
2010年第22期74-76,共3页
Science & Technology Review
基金
高等学校博士学科点专项科研基金项目(20060532026)
关键词
粒子群优化算法
遗传算法
全局搜索
局部搜索
种群多样性
particle swarm optimization
genetic algorithm
global search
local search
population diversity
