基于正弦惯性权值反向蜉蝣优化算法

Oppositional mayfly optimization algorithm based on sinusoidal inertia weights

针对蜉蝣优化算法(Mayfly Algorithm,MA)早熟问题,提出一种基于正弦惯性权值反向蜉蝣优化算法。该算法采用准反向学习初始化替代随机初始化,在雌雄蜉蝣位置更新过程中引入准反向学习策略,将当前位置与反向位置反向学习,择优选择位置进入下一次迭代。基于正弦函数的周期性,将正弦函数、随机策略与惯性权值融合成正弦惯性权值策略以增强MA的搜索能力。仿真结果表明,所提算法在全局收敛性和收敛速度方面均明显优于其他算法,并在焊接梁设计中验证了该算法的实用性。

For the prematurity problem of the mayfly algorithm(MA),an oppositional mayfly optimization algorithm is proposed based on the sinusoidal inertia weights.The algorithm adopts the quasi-oppositional learning initialization instead of the random initialization,introduces a quasi-oppositional learning strategy in the male and female mayfly position updating process,in which the current and the oppositional positions are learnt oppositely,and are optimized to enter the next iteration.Based on the periodicity of the sinusoidal function,the sinusoidal function,stochastic strategy,and the inertia weights are fused to form a sinusoidal inertia weights strategy,in order to enhance the search capability of MA.Simulation results show that significantly outperforms the comparison algorithms in terms of global convergence and convergence speed.The practicality of the proposed algorithm can be verified by the welded beam design.