作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2022, Vol. 48 ›› Issue (7): 73-81. doi: 10.19678/j.issn.1000-3428.0062074

• 人工智能与模式识别 • 上一篇    下一篇

动态调整成长方式的郊狼优化算法及其应用

严逍亚1, 王振雷1, 王昕2   

  1. 1. 华东理工大学 能源化工过程智能制造教育部重点实验室, 上海 200237;
    2. 上海交通大学 电工与电子技术中心, 上海 200240
  • 收稿日期:2021-07-13 修回日期:2021-09-01 出版日期:2022-07-15 发布日期:2021-09-09
  • 作者简介:严逍亚(1997—),女,硕士研究生,主研方向为智能优化算法;王振雷,教授、博士;王昕,副教授、博士。
  • 基金资助:
    国家重点研发计划 (2018YFB1701103);国家自然科学基金重大项目(61890930-3);国家杰出青年科学基金(61725301);中央高校基本科研业务费专项资金(222202117006)。

Coyote Optimization Algorithm with Dynamically Adjusting Growth Mode and Its Application

YAN Xiaoya1, WANG Zhenlei1, WANG Xin2   

  1. 1. Key Laboratory of Smart Manufacturing in Energy Chemical Process, Ministry of Education, East China University of Science and Technology, Shanghai 200237, China;
    2. Center of Electrical & Electronic Technology, Shanghai Jiao Tong University, Shanghai 200240, China
  • Received:2021-07-13 Revised:2021-09-01 Online:2022-07-15 Published:2021-09-09

摘要: 郊狼优化算法在迭代运行时种群多样性降低,收敛速度变慢,易陷入局部最优,并且在求解约束优化问题时难以获得可行解。提出一种动态调整成长方式的郊狼优化算法(DGCOA)。在狼群进化中引入变异交叉策略,增强种群多样性,基于郊狼成长策略加入全局最优个体指导搜索,使得每个子种群中的个体从不同的方向快速逼近最优解位置,并根据种群中个体相似度对郊狼位置更新方式进行调整,平衡算法的全局探索与局部开发能力。在求解约束优化问题时,利用自适应约束处理方法构建新的适应度函数,协调优化目标和约束违反度。基于CEC2006对22个测试函数和3个工程设计问题进行仿真,结果表明,与COA、ICTLBO、ODPSO等算法相比,DGCOA算法具有较高的收敛精度和稳定性,适用于求解复杂优化问题。

关键词: 郊狼优化算法, 变异交叉, 约束处理, 测试函数, 工程优化

Abstract: As the search continues, Coyote Optimization Algorithm(COA) has several disadvantages of the population reduced diversity, slow convergence speed, ease to fall into local optimum and unable to obtain a feasible solution when solving a constrained optimization problem.To overcome these disadvantages, Coyote Optimization Algorithm with Dynamically adjusting Growth mode(DGCOA) is proposed.First, the mutation crossover strategy is introduced to enhance the population diversity.Subsequently, the global optimal individual is added to guide the search so that individuals in each sub-population can quickly approach the optimal solution position from different directions, and based on the similarity degree of individuals in the population, the updating method of coyote growth was adjusted to effectively enhance the ability of balance global exploration and local development of the algorithm.When solving constrained optimization problems, a new fitness function is constructed by using the adaptive constraint processing method to coordinate the optimization objective and constraint violation degree.The simulation results of 22 test functions and 3 engineering design problems on CEC2006 reveal that the DGCOA algorithm has higher convergence accuracy and stability than COA, ICTLBO, and ODPSO algorithms. Consequently, DGCOA can solve complex optimization problems more effectively.

Key words: Coyote Optimization Algorithm(COA), mutation crossover, constraint handling, test function, engineering optimization

中图分类号: