一种增强型双种群粒子群算法的设计与实现  

Design and implementation of an enhanced dual-population Particle Swarm Optimization algorithm

作　　者：张彦超 王晓丽 苏奎[1] ZHANG Yanchao;WANG Xiaoli;SU Kui(College of Medical Imaging,Mudanjiang Medical University,Mudanjiang 157011,Heilongjiang,China;Modern Educational&Technical Center,Mudanjiang Medical University,Mudanjiang 157011,Heilongjiang,China)

机构地区：[1]牡丹江医学院医学影像学院,黑龙江牡丹江157011 [2]牡丹江医学院现代教育技术中心,黑龙江牡丹江157011

出　　处：《智能计算机与应用》2024年第5期194-198,共5页Intelligent Computer and Applications

基　　金：牡丹江市应用技术研究与开发计划项目(HT2022JG130);牡丹江医学院教育教学改革项目(MYPY20210009)。

摘　　要：粒子群算法作为一种具有收敛速度快、易于实现和参数调节少等优点的群体智能算法,被广泛应用于函数与组合优化、机器学习等众多领域。在粒子群及其众多修改算法中惯性权重选择对算法性能起到较大的作用。为此,在经典粒子群算法上针对惯性权重提出了改进策略,设计了一种增强型双种群粒子群算法(EDUPSO)。当粒子在进化中将较小的惯性因子赋予到此次进化到最优位置的种群,将较大的惯性因子赋予到此次没有进化到当前最优位的种群。通过此思路提出了算法的设计和实现方式,并通过多个不同的测试函数分析了该算法与其他经典改进算法在性能上的差异,通过测试的结果可以看出相对于经典的粒子群算法及经典改进算法,此算法在搜索的速度、稳定性及精度上都有明显的优势。Particle Swarm Optimization is widely used in many fields such as function and combinatorial optimization,machine learning,etc.,due to its advantages of fast convergence,easy implementation and less parameter adjustment.For PSO and its many modified algorithms,the selection of inertia weight has a significant impact on the algorithm's performance.Therefore,an improved strategy is proposed for the inertia weight in the classical PSO algorithm,which led to the development of Enhanced Dual-Particle Swarm Optimization(EDUPSO).In EDUPSO,smaller inertia factors are assigned to the population that evolves to the best position during evolution,while larger inertia factors are assigned to the population that did not reach the current optimal position.The algorithm's design and implementation approach are developed using this idea,and the algorithm's performance is analyzed in comparison to other classical improved algorithms using various test functions.The results of the tests show that,relative to the classical PSO algorithm and other classical improved algorithms,this algorithm has significant advantages in terms of search speed,stability and accuracy.

关 键 词：粒子群算法改进 神经网络优化 无线定位优化 人工智能与数据挖掘 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] TP18[自动化与计算机技术—计算机科学与技术]