基于模拟退火的粒子群算法在函数优化中的应用  

Application of particle swarm optimization algorithm based on simulated annealing in function optimization

作　　者：李淑香 LI Shu-xiang(Department of Mathematics and Physics,Institute of Mobile Communication of Chongqing University of Posts and Telecommunications,Chongqing 401520,China)

机构地区：[1]重庆邮电大学移通学院数理教学部

出　　处：《沈阳工业大学学报》2019年第6期664-668,共5页Journal of Shenyang University of Technology

基　　金：重庆市教委科学技术研究基金资助项目(KJ1501505);重庆市教委教改项目(173157)

摘　　要：为了克服标准粒子群搜索算法在函数优化中出现的迭代速度慢、精度低且易陷入局部最优等缺点,提出了一种基于模拟退火的粒子群优化算法.该混合算法利用模拟退火算法中的概率突变能力,在接受新解时既能接受好解也能以一定的概率接受坏解,能够跳出算法的局部最优解,不仅提高了算法的灵活性与多样性,还能提高粒子的多样性,从而获得了较强的全局与局部优化能力.对5个非线性基准函数进行仿真实验对比后发现,混合算法在非线性复杂函数优化中具有更好的寻优能力,表现出调节精度高,收敛速度快等优点,同时避免了"早熟"现象和陷入局部最优的问题.In order to overcome the disadvantages,e.g.lower iteration speed,poor precision and liability to local optimality,of standard particle swarm optimization(PSO)algorithm in function optimization,a simulated annealing particle swarm optimization(SAPSO)algorithm was proposed.The probability mutation property in the simulated annealing(SA)algorithm was adopted in this hybrid algorithm,so the hybrid algorithm could accept not only good solutions but also bad solutions with a certain probability.The hybrid algorithm could jump out of local optimal solution,enhancing not only the flexibility and diversity of algorithm but also the variety of particles.As a result,the hybrid algorithm attained strong ability for global and local optimization.Through the comparison among simulated tests among five nonlinear benchmark functions,it is discovered that the hybrid algorithm has better optimization ability in nonlinear complex function optimization.In addition,the hybrid algorithm demonstrates the advantages of high adjustment precision and fast convergence speed,avoiding the premature phenomenon and the entrapment of local optimization.

关 键 词：粒子群算法 遗传算法 模拟退火算法 概率突变 多样性 混合算法 基准函数 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]