计算机应用

• 人工智能与仿真 •    下一篇

基于动态D向分割和混沌扰动的阴阳对算法

李大海1,刘庆腾1,艾志刚2,王振东3   

  1. 1. 江西理工大学信息工程学院
    2. 江西理工大学
    3. 江西理工大学信
  • 收稿日期:2021-07-27 修回日期:2021-09-17 发布日期:2021-09-27 出版日期:2021-09-27
  • 通讯作者: 刘庆腾

Yin-Yang-pair algorithm based on dynamic D-way splitting strategy and chaos perturbation

  • Received:2021-07-27 Revised:2021-09-17 Online:2021-09-27 Published:2021-09-27

摘要: 摘 要: 为增强YYPO-SA1的性能,提出了一种动态D向分割和混沌扰动的阴阳对算法N-YYPO,其首先基于牛顿衰减机制动态调整YYPO-SA1中的D向分割概率,然后在分割阶段加入混沌扰动策略。N-YYPO利用动态调整机制在搜索前期使用较大的D向分割概率,在搜索后期则使用较小的D向分割概率,增强了算法的全局搜索能力,同时混沌扰动策略丰富了解的多样性,增强了算法跳出局部最优的能力。最后,将N-YYPO应用于风力发电机的参数优化设计问题。文章选用了15个单峰、多峰和组合测试函数进行性能评估,将N-YYPO YYPO-SA1,以及6个代表性的单目标优化算法:粒子群优化算法,乌鸦优化算法,灰狼算法,鲸鱼算法,花授粉算法,麻雀算法进行性能评测比较,相较于YYPO-SA1算法在f1函数上有着12个数量级的提升。在Friedman检验中NYYPO的排名分别为:2.87,2.0,1.93,均排名第一。实验结果表明NYYPO在统计学意义上具有显著的性能优势。在风力发电机参数优化设计问题中N-YYPO也取得了更好的优化结果。

Abstract: Abstract: To improve the performance of Yin-Yang-pair Optimization- Simulated annealing1 ( YYPO-SA1 ), an enhanced algorithm Newton- Yin-Yang-pair Optimization ( N-YYPO ) was proposed. In order to dynamically tune the probability of D-way splitting ,Newton decay mechanism was adopted and chaos perturbation mechanism was applied in splitting stage. The dynamic adjustment mechanism was applied to enable N-YYPO to use a large D-direction segmentation probability in the early stage of search, and use a smaller D-direction segmentation probability in the late stage of search, which enhanced the global search ability of the algorithm. Meanwhile, the diversity of solution was enriched ,and the ability of the algorithm to jump out of local optimal was enhanced by chaotic perturbation strategy. At last, N-YYPO was applied to the parametric optimization design problem of wind-driven generator.15 functions was selected, including unimodal, multimodal, and composited functions, as benchmark to evaluate performance of N-YYPO ,with YYPO-SA1, and 6 state-of the art single-objective: Particle Swarm Optimization ( PSO ), Flower Pollination Algorithm ( FPA ), Gray Wolf Optimizer ( GWO ), Whale Optimization Algorithm ( WOA ), Crow Search Algorithm ( CSA ), and Sparrow Search Algorithm ( SPA ), compared with YYPO-SA1 algorithm, N-YYPO obtains 12 orders of magnitude improvement in f1 function. In Friedman test, NYYPO ranks 2.87, 2.0 and 1.93 respectively, ranking first in all of them. Experimental result show that N-YYPO can achieve significant performance advantage in statistical sense. N-YYPO also achieves better optimization results in the parameter optimization problem of wind-driven generator.

中图分类号: