Computer Engineering and Applications ›› 2020, Vol. 56 ›› Issue (7): 147-154.DOI: 10.3778/j.issn.1002-8331.1812-0184

Previous Articles     Next Articles

Interactive Learning Cuckoo Search Algorithm

ZHANG Hainan, YOU Xiaoming, LIU Sheng, LIU Zhongqiang   

  1. 1.School of Electronic and Electrical Engineering, Shanghai University of Engineering Science, Shanghai 201620, China
    2.School of Management, Shanghai University of Engineering Science, Shanghai 201620, China
  • Online:2020-04-01 Published:2020-03-28



  1. 1.上海工程技术大学 电子电气学院,上海 201620
    2.上海工程技术大学 管理学院,上海 201620


In order to solve the traveling salesman problem, the cuckoo search algorithm has the problems of lack of initial information and slow convergence. An Interactive Learning Cuckoo Search Algorithm(ILCSA) is proposed. In order to improve the search efficiency of the cuckoo algorithm, a two-layer interactive learning model is constructed by combining the ant colony algorithm. The ant colony is used as the bottom population, and the cuckoo is used as the high-level population. The two populations learn from each other, cooperate to find the best, and improve the search speed. In addition, cuckoo search algorithm introduces a reinforcement learning strategy, adaptively updates the step size, dynamically adjusts the discovery probability, and deeply optimizes the optimal solution to further improve the quality of the solution. Finally, multiple sets of standard TSPLIB examples of different scales are compared with other optimization algorithms. The results show that the ILCSA algorithm performs better in terms of accuracy and stability.

Key words: interactive learning model, reinforcement learning strategy, adaptive step size, dynamic adjustment mechanism, Cuckoo Search Algorithm(CSA), ant colony optimization algorithm


针对布谷鸟搜索算法在求解旅行商问题时,存在初期信息缺乏严重和收敛速度慢等问题,提出一种交互式学习的布谷鸟搜索算法(Interactive Learning Cuckoo Search Algorithm,ILCSA)。为提高布谷鸟搜索算法的搜索效率,结合蚁群优化算法构建双层交互学习模型,将蚁群作为底层种群,布谷鸟作为高层种群,双种群互相学习,合作寻优,提高搜索速度;此外,在布谷鸟搜索算法中引入强化学习策略,自适应更新步长,并对发现概率进行动态调整,深度优化最优解,进一步提高解的质量。最后采用多组不同规模的标准TSPLIB算例与其他优化算法进行对比,结果表明ILCSA算法在求解精度和稳定性方面表现更优。

关键词: 交互学习模型, 强化学习策略, 自适应步长, 动态调整机制, 布谷鸟搜索算法, 蚁群优化算法