基于杂交策略的自适应灰狼优化算法

Adaptive gray wolf optimization algorithm based on hybridization strategy

针对灰狼优化算法(GWO)存在较为严重的收敛性缺陷问题,提出了一种基于杂交策略的自适应灰狼优化算法(AGWO)。首先引入非线性收敛因子,以平衡算法的全局搜索性和局部开发性;其次引进遗传杂交策略,对灰狼群体以一定概率两两杂交以产生新个体,从而有效增强灰狼群体的多样性;同时为避免算法后期陷入局部最优解,受蝠鲼觅食策略的启发,引入蝠鲼觅食策略并加入了动态自适应调节因子以调节群体的多样性,有效提升算法的收敛精度及全局寻优性能。通过选取CEC2014中11个基准测试函数进行实验,与其他相关算法横纵向对比分析,多方位验证了AGWO算法的综合寻优性能。实验结果表明,在相同参数设置下,AGWO算法的收敛速度及综合寻优性能明显优于其他比较算法。

Aiming at solving the problem of convergence defect in the traditional gray wolf optimization algorithm(GWO),this paper proposed an adaptive gray wolf optimization algorithm(AGWO)with hybridization strategy.The algorithm firstly introduced a non-linear convergence factor to balance the global search and local development of the algorithm.Secondly,it introduced a genetic hybrid strategy,which crossed the gray wolf population in pairs with a certain probability to produce new individuals,thereby effectively enhancing the diversity of the gray wolf population.At the same time,in order to avoid the algorithm falling into the local optimal solution in the later stage,inspired by the manta ray foraging strategy,the manta ray foraging strategy was introduced and a dynamic adaptive adjustment factor was added to adjust the diversity of the population,which effectively improved the convergence accuracy of the algorithm and global optimization performance.Numerical experiments of selecting 11 benchmark test functions in CEC2014 demonstrate that the proposed algorithm achieves better performance of higher convergence speed and optimization.