林涛, 冯嘉冀, 赵伊. 基于精英策略与正弦余弦机制的蜻蜓算法研究[J]. 微电子学与计算机, 2020, 37(9): 24-30.
引用本文: 林涛, 冯嘉冀, 赵伊. 基于精英策略与正弦余弦机制的蜻蜓算法研究[J]. 微电子学与计算机, 2020, 37(9): 24-30.
LIN Tao, FENG Jia-ji, ZHAO Yi. Study on dragonfly algorithm based on elite strategy and sines and cosines mechanism[J]. Microelectronics & Computer, 2020, 37(9): 24-30.
Citation: LIN Tao, FENG Jia-ji, ZHAO Yi. Study on dragonfly algorithm based on elite strategy and sines and cosines mechanism[J]. Microelectronics & Computer, 2020, 37(9): 24-30.

基于精英策略与正弦余弦机制的蜻蜓算法研究

Study on dragonfly algorithm based on elite strategy and sines and cosines mechanism

  • 摘要: 蜻蜓算法模拟蜻蜓觅食行为对个体位置进行更新,为提高蜻蜓算法收敛精度、避免陷入局部最优解,现提出基于精英策略与正弦余弦机制的蜻蜓算法(ESCDA).该算法利用精英策略优化种群初始化,通过正弦余弦机制优化位置更新,改善了算法的收敛效果、提高了寻优精度.将算法应用于14个典型复杂函数优化问题,并与经典的粒子群算法,狼群算法和蚁狮算法相比较,实验结果表明该改进算法具有良好的全局收敛性和寻优精度,尤其适合高维,多峰的复杂函数求解.

     

    Abstract: In order to improve the convergence accuracy of dragonfly algorithm and avoid falling into loc-al optimal solution, dragonfly algorithm based on elite strategy and sines and cosines mechanism(ESCDA) is proposed.The algorithm USES elite strategy to optimize population initialization, optimizes position upd-ate through sines and cosines mechanism, and improves the convergence effect of the algorithm and the o-ptimization accuracy.The algorithm is applied to 14 typical complex function optimization problems, and c-ompared with the classical particle swarm optimization algorithm, Wolf pack algorithm and ant-lion algorit-hm, the experimental results show that the improved algorithm has good global convergence and optimizat-ion accuracy, and is especially suitable for solving complex functions with high dimensions and multiple p-eaks.

     

/

返回文章
返回