电子学报 ›› 2019, Vol. 47 ›› Issue (2): 266-273.DOI: 10.3969/j.issn.0372-2112.2019.02.002

• 学术论文 • 上一篇    下一篇

基于多算子协同进化的自适应并行量子遗传算法

曲志坚, 陈宇航, 李盘靖, 刘晓红, 李彩虹   

  1. 山东理工大学计算机科学与技术学院, 山东淄博 255049
  • 收稿日期:2018-04-23 修回日期:2018-07-01 出版日期:2019-02-25
    • 通讯作者:
    • 李彩虹
    • 作者简介:
    • 曲志坚 男,1980年生于山东青岛.现为山东理工大学计算机科学与技术学院副教授、硕士生导师.主要研究方向为进化算法和网络资源优化.E-mail:zhijianqu@sdut.edu.cn;陈宇航 男,1997年生于浙江丽水.现为山东理工大学计算机科学与技术专业本科生.ACM-ICPC亚洲区域赛银奖获得者.主要研究方向为算法分析与设计.E-mail:sdut_snow@163.com
    • 基金资助:
    • 国家自然科学基金 (No.61473179); 山东省自然科学基金 (No.ZR2016FM18); 山东省高等学校科技计划项目 (No.J16LN20)

Cooperative Evolution of Multiple Operators Based Adaptive Parallel Quantum Genetic Algorithm

QU Zhi-jian, CHEN Yu-hang, LI Pan-jing, LIU Xiao-hong, LI Cai-hong   

  1. School of Computer Science and Technology, Shandong University of Technology, Zibo, Shandong 255049, China
  • Received:2018-04-23 Revised:2018-07-01 Online:2019-02-25 Published:2019-02-25

摘要: 量子遗传算法具有种群规模小,全局搜索能力强的特点被广泛应用于各类优化问题的求解.为了进一步提高量子遗传算法的收敛速度和搜索稳定性,克服算法的早熟问题,本文改进了基于自适应机制的量子遗传算法.在自适应量子遗传算法的基础上根据种群的适应度定义了个体相似度评价算子、个体适应度评价算子和种群变异调整算子及相应算子的计算方法,利用多算子协同评价当前种群状态并根据进化代数的变化,自适应的改变个体的变异概率,提高了算法全局寻优能力和收敛速度,降低了算法陷入局部寻优的概率.此外,为了提高算法的时间效率,将算法采用并行多宇宙的方式实现.实验结果表明,本文提出的算法在全局搜索性能、收敛速度和时间效率方面有较好的综合表现.

关键词: 遗传算法, 并行计算, 自适应机制, 量子变异

Abstract: The characteristics of strong global search ability with small population size lead to the quantum genetic algorithm is well popular in solving optimization problems.In order to further improve the convergence speed,search stability and overcome the pre-matureness of the quantum genetic algorithm,an improved adaptive mechanism based quantum genetic algorithm was presented in the paper.For the presented algorithm,the individual similarity evaluation operator,individual fitness evaluation operator and population mutation adjustment operator were defined and added into the self-adaptive based quantum genetic algorithm.The way of calculating the three operators were also proposed.Therefore,the current population state can be evaluated by the operators cooperatively,and the individual's mutation probability can be determined according to the current population state.The proposed algorithm can improve the global optimization ability and convergence speed,and reduces the probability of falling into local optimization.In addition,a parallel multi-universe mechanism is employed to improve the time efficiency of the algorithm.Experimental results show that the proposed algorithm has a good performance in the global search performance and time efficiency.

Key words: genetic algorithm, parallel computing, adaptive mechanism, quantum mutation

中图分类号: