带有非线性惯性权重和柯西变异的粒子群优化算法  

Improved Particle Swarm Optimization Algorithm with Nonlinear Inertia Weight and Cauchy Mutation

作　　者：李会荣 彭娇 LI Huirong;PENG Jiao(Department of Mathematics and Computer Application,Shangluo University,Shangluo 726000)

机构地区：[1]商洛学院数学与计算机应用学院,商洛726000

出　　处：《计算机与数字工程》2021年第7期1325-1329,1362,共6页Computer & Digital Engineering

基　　金：陕西省自然科学基金(编号:2020JM-630);陕西省教育厅科研计划项目(编号:17JK024);商洛市科研基金项目(编号:2020-Z-043);商洛学院科研基金项目(编号:18SKY009);商洛学院应用数学科研创新团队(编号:19SCX02);商洛学院重点学科建设项目资助。

摘　　要：粒子群优化算法是一种新型启发式智能优化算法,它运行速度快,收敛性强,但是容易陷入局部极值。为了克服粒子群算法的早熟收敛现象,提出了一种新的带有非线性惯性权重和柯西变异的粒子群优化算法。首先,对算法中的惯性权值进行改进,增强粒子局部收敛能力;然后,利用柯西变异算子,增加种群多样性。数值实验表明,提出的改进粒子群优化算法具有较快收敛速度,寻优能力强,能有效克服早熟收敛现象。Particle swarm optimization is a new heuristic intelligent optimization algorithm,which has strong speed and fast convergence.But it is easy to fall into local optimum.In order to overcome the premature convergence of the algorithm,a new parti⁃cle swarm optimization with nonlinear inertia weight and cauchy mutation is proposed.Firstly,the inertia weight function of the algo⁃rithm is improved to enhance the local convergence ability of particles.Then,cauchy mutation operator is proposed to increase popu⁃lation diversity.The numerical experiments show that the proposed improved particle swarm optimization algorithm has faster conver⁃gence speed and better searching ability.It also can effectively conquer the premature convergence phenomenon.

关 键 词：粒子群优化 惯性权重 早熟收敛 柯西变异 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]