系统工程与电子技术 ›› 2020, Vol. 42 ›› Issue (1): 148-156.doi: 10.3969/j.issn.1001-506X.2020.01.20

• 系统工程 • 上一篇    下一篇

混合方法优化的自适应引力搜索算法

娄奥1(), 姚敏立1(), 贾维敏2(), 袁丁1()   

  1. 1. 火箭军工程大学作战保障学院, 陕西 西安 710025
    2. 火箭军工程大学核工程学院, 陕西 西安 710025
  • 收稿日期:2019-05-27 出版日期:2020-01-01 发布日期:2019-12-23
  • 作者简介:娄奥(1995-),男,硕士研究生,主要研究方向为智能算法优化、系统辨识。E-mail:la5310@qq.com|姚敏立(1966-),男,教授,博士研究生导师,博士,主要研究方向为智能算法优化、宽带移动卫星通信。E-mail:yaominli66@163.com|贾维敏(1971-),女,教授,博士研究生导师,博士,主要研究方向为智能算法优化、阵列信号处理。E-mail:jwm602@163.com|袁丁(1983-),男,讲师,博士,主要研究方向为导航制导、宽带移动卫星通信。E-mail:yuanding31@126.com
  • 基金资助:
    国家自然科学基金(61179004);国家自然科学基金(61179005)

Adaptive gravitational search algorithm improved by hybrid methods

Ao LOU1(), Minli YAO1(), Weimin JIA2(), Ding YUAN1()   

  1. 1. School of Military Operational Support, Rocket Force University of Engineering, Xi'an 710025, China
    2. School of Nuclear Engineering, Rocket Force University of Engineering, Xi'an 710025, China
  • Received:2019-05-27 Online:2020-01-01 Published:2019-12-23
  • Supported by:
    国家自然科学基金(61179004);国家自然科学基金(61179005)

摘要:

针对引力搜索算法存在的易早熟收敛、易陷入局部最优、搜索精度有待提高等缺陷,提出一种混合方法优化的自适应引力搜索算法(gravitational search algorithm,GSA)。首先利用Sobol序列初始化种群,增强算法全局搜索能力;其次引入Hamming贴进度计算种群成熟度,判断种群是否早熟;然后引入Logistic混沌对种群作混沌搜索,变异已陷入局部最优的粒子位置;最后基于早熟收敛判断因子改进引力系数,并为粒子位置公式添加收缩因子,促使种群加快脱离局部最优。对9个不同类型的基准测试函数做仿真实验,结果表明新算法能有效改善种群的早熟问题,具备更好的寻优性能。

关键词: 引力搜索算法, 低差异序列, 贴进度, 混沌

Abstract:

In order to overcome the shortcomings of premature convergence, trapping in local optimum easily and lower search accuracy of gravitational search algorithm (GSA), an adaptive GSA improved by hybrid methods is proposed. Firstly, sobol sequence is used to initialize the population and enhance the global search ability. Secondly, hamming nearness degree is introduced to calculate the population maturity and judge whether the population is premature. Thirdly, logistic chaos is introduced to search the population chaotically and update the particle which has fallen into the local optimum. Finally, based on the precocious convergence judgment factor, the gravitational coefficient is improved, and the shrinkage factor is added to the particle position formula to accelerate the population departure from the local optimum. The simulation results of nine different types of benchmark functions show that the new algorithm can effectively improve the premature convergence problem and has better optimization performance.

Key words: gravitational search algorithm (GSA), low-discrepancy sequence, nearness degree, chaos

中图分类号: