Computer Engineering and Applications ›› 2021, Vol. 57 ›› Issue (19): 84-96.DOI: 10.3778/j.issn.1002-8331.2009-0199

Previous Articles     Next Articles

Fireworks Algorithm Based on Tournament Elite Learning and Covariance Mutation

WAN Da, LI Jun   

  1. 1.College of Computer Science and Technology, Wuhan University of Science and Technology, Wuhan 430065, China
    2.Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System, Wuhan 430065, China
  • Online:2021-10-01 Published:2021-09-29

锦标赛精英学习与协方差变异的烟花算法

万达,李俊   

  1. 1.武汉科技大学 计算机科学与技术学院,武汉 430065
    2.智能信息处理与实时工业系统湖北省重点实验室,武汉 430065

Abstract:

To overcome the shortcomings of FireWorks Algorithm(FWA) such as low convergence precision, slow convergence speed and easy to fall into local optimum, a fireworks algorithm(GLFWA-CM) based on tournament elite learning and covariance mutation is proposed. Firstly, in the process of the explosion operator, the update information of the core fireworks is used to determine the explosion radius of the core fireworks in each dimension and guide the core fireworks to generate more explosion sparks in the update direction, which improves the search ability of the core fireworks. Secondly, covariance mutation is used in the mutation operator instead the original Gaussian mutation that makes full use of the information of explosion sparks, effectively balancing the local search and global search capabilities of the algorithm. Finally, in the fireworks selection process, an elite learning strategy based on tournament is proposed, which effectively speeds up the convergence speed of the algorithm. The simulation experiments on CEC2015 test function show that the algorithm has better convergence and stability than other fireworks algorithms.

Key words: fireworks algorithm, core fireworks, update information guide, covariance mutation, tournament, elite learning

摘要:

针对传统烟花算法收敛精度低,收敛速度慢,容易陷入局部最优等问题,提出一种基于锦标赛精英学习与协方差变异的烟花算法(GLFWA-CM)。该算法在爆炸算子过程中利用核心烟花更新信息确定核心烟花在每一维上的爆炸半径,并引导核心烟花在更新方向上产生更多的爆炸火花,提高了核心烟花的搜索能力;在变异算子中用协方差变异代替原来的高斯变异,充分利用爆炸火花的信息,有效平衡了算法的局部搜索和全局搜索能力;在烟花选择过程中提出了一种基于锦标赛的精英学习策略,有效加快了算法收敛速度。在CEC2015测试函数上做仿真实验,结果表明,与多种经典烟花算法相比,该算法在收敛性和稳定性上都具有较好表现。

关键词: 烟花算法, 核心烟花, 更新信息引导, 协方差变异, 锦标赛, 精英学习