改进的差分演化算法及其在资金决策中的应用  

The Improved Differential Evolution Algorithm and Its Application in the Decision-making of Investment

作　　者：鄢靖丰[1] 田胜利[1] YAN Jingfeng;TIAN Shengli(College of Information Technology,Xuchang University,Xuchang 461000,China)

机构地区：[1]许昌学院信息工程学院,河南许昌461000

出　　处：《许昌学院学报》2024年第2期104-108,共5页Journal of Xuchang University

基　　金：教育部高教司产学合作协同育人项目(202101358002);河南省终身教育专项课题(豫教[2022]45742);许昌学院优秀青年骨干教师资助计划。

摘　　要：针对传统的基本差分演化算法存在对参数设置太过于敏感,算法容易陷入早熟收敛、稳定性不强等等缺陷,提出了改进措施,采取了对称算法来保持种群多样性,通过变异个体,对称个体,个体三者进行有效竞争,增强算法的全局搜索能力;同时采用了轮盘赌的选择策略和二次变异算子保证种群个体的有效性与多样性,增加算法的全局稳定性.文中的算法应用于求解资金决策分配问题上,通过模拟仿真,该算法在稳定性、收敛速度、求解的精度方面优于其他被比较的算法.In view of the shortcomings of the traditional basic differential evolution algorithm:the algorithm is too sensitive to parameter settings and the algorithm is easy to fall into precocious convergence and weak stability,this paper proposes improvement measures,and adopts a symmetry algorithm to maintain population diversity,and effectively compete with mutated individuals,symmetric individuals,and individuals,so as to enhance the global search ability of the algorithm.At the same time,the selection strategy of roulette and the quadratic mutation operator are used to ensure the effectiveness and diversity of individuals in the population,and to increase the global stability of the algorithm.The algorithm in this paper is applied to solve the problem of capital decision allocation,and through simulation,the algorithm is better than other compared algorithms in terms of stability,convergence speed and solution accuracy.

关 键 词：差分演化算法 资金决策 收敛速度 稳定性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]