通信学报 ›› 2017, Vol. 38 ›› Issue (8): 201-212.doi: 10.11959/j.issn.1000-436x.2017051

• 学术通信 • 上一篇    下一篇

基于双变异策略的自适应骨架差分进化算法

刘会宇1,2,韩继红1,袁霖1,于波1,3   

  1. 1 解放军信息工程大学,河南 郑州 450001
    2 中国人民解放军75741部队,广东 广州 510510
    3 中国人民解放军国防信息学院,湖北 武汉 430010
  • 修回日期:2016-11-30 出版日期:2017-08-01 发布日期:2017-09-07
  • 作者简介:刘会宇(1987-),男,湖北咸宁人,解放军信息工程大学硕士生,主要研究方向为人工智能。|韩继红(1966-),女,山西定襄人,博士,解放军信息工程大学教授、博士生导师,主要研究方向为网络与信息安全、安全协议形式化分析与自动化验证。|袁霖(1981-),男,河南商丘人,博士,解放军信息工程大学副教授,主要研究方向为安全协议形式化分析与自动化验证、软件可信性分析。|于波(1991-),男,重庆人,解放军信息工程大学硕士生,主要研究方向为云计算和数据中心。
  • 基金资助:
    国家自然科学基金资助项目(61309018)

Self-adaptive bare-bones differential evolution based on bi-mutation strategy

Hui-yu LIU1,2,Ji-hong HAN1,Lin YUAN1,Bo YU1,3   

  1. 1 PLA Information Engineering University,Zhengzhou 450001,China
    2 PLA Unit 75741,Guangzhou 510510,China
    3 Chinese People's Liberation Army National Defense Information Institute,Wuhan 430010,China
  • Revised:2016-11-30 Online:2017-08-01 Published:2017-09-07
  • Supported by:
    The Nationa1 Natura1 Science Foundation of China(61309018)

摘要:

骨架差分进化算法能够较好规避差分进化算法控制参数和变异策略选择问题。针对基于双变异策略的经典骨架差分算法(MGBDE)没有根据个体进化差异选择适合的变异策略和考虑早熟收敛的问题,提出一种改进算法。该算法引入变异策略选择因子,并借鉴自适应差分进化算法的设计思想,将选择因子随个体共同参与进化,使个体执行当前最为适合的变异策略,克服原始算法进化过程的盲目性,同时选择因子的动态自适应特性保持了骨架算法近似无参数的优点;该算法加入停滞扰动策略,降低陷入局部最优的风险。采用18个标准测试函数进行实验,结果表明,新算法在收敛精度、收敛速度和顽健性上整体优于多种同类骨架算法以及知名的差分进化算法。

关键词: 差分进化, 骨架算法, 双变异策略, 自适应

Abstract:

Bare-bones differentia1 evo1ution(BBDE)can e1egant1y so1ve the se1ection prob1em of contro1 parameters and mutation strategy in differentia1 evo1ution(DE).MGBDE is a c1assica1 BBDE based on bi-mutation strategy.However,it random1y assigns a mutation strategy to each individua1,not considering their differences during evo1ution process,meanwhi1e it may suffer from premature convergence.To overcome these drawbacks,a modified a1gorithm based on MGBDE was proposed.A mutation strategy choice factor that guided the individua1 to choose a preferab1e mutation strategy at each mutation operation was introduced,avoiding the evo1ution b1indness brought by the random se1ection of mutation strategy.To retain the a1most parameter-free characteristic of bare-bones a1gorithm,the tuning of choice factor to be adapted was invo1ved in the individua1 evo1ution,inspired by the concept of se1f-adaptive DE.The a1gorithm a1so inc1uded a we11-designed stagnation perturbation mechanism to reduce the risk of trapping into the 1oca1 optima1.Experimenta1 resu1ts on 18 benchmark functions show that the proposed a1gorithm genera11y achieves better performance than state-of-the-art BBDE variants and severa1 we11-known DE a1gorithms in terms of convergence and robustness.

Key words: differentia1 evo1ution, bare-bones a1gorithm, bi-mutation strategy, se1f-adaptive

中图分类号: 

No Suggested Reading articles found!