基于圆形搜索机制的多反向复合鲸鱼优化算法

Whale algorithm based on circle search and multiple opposition-based learning

为了改进鲸鱼优化算法存在的种群多样性和勘探开采能力不足等问题,提出了基于圆形搜索机制的多反向复合鲸鱼优化算法(CSOWOA).首先,针对种群多样性进行了改进.通过折射反向学习初始化种群,以便于搜索到更为隐蔽的空间,加强初始种群的多样性;在算法寻优过程中,通过适应度值大小来划分优势种群和劣势种群,分别对其进行折射反向学习和随机反向学习的多反向复合方式,确保算法寻优过程中种群分布的多样性,便于算法寻优。其次,针对算法勘探开采能力进行了改进.采用结合种群成功率的自适应权重来加强鲸鱼的包围搜索能力,同时在包围搜索过程中通过两种圆形搜索机制加强算法的勘探和开采能力,提升算法的收敛速度和寻优精度.最后,加入正态变异来扰动精英个体的位置,带动可能陷入停滞的鲸鱼种群,避免算法陷入局部最优.仿真实验在13个基准测试函数中与几个知名改进鲸鱼算法和经典智能优化算法进行比较,比较结果显示CSOWOA有明显的提升效果.

In order to improve the population diversity and insufficient exploration and exploitation capacity of whale optimization algorithm,a multi-inverse composite whale optimization algorithm(CSOWOA)based on circular search mechanism is proposed.First,improvements were made for population diversity.The initial population is initialized by refractive backward learning through refraction,so as to find more hidden space and strengthen the diversity of the initial population.In the optimization process of the algorithm,the dominant population and the inferior population are divided by the size of the fitness value,and the multi-inverse composite method of refractive backward learning and random backward learning is applied to them respectively to ensure the diversity of the population in the process of algorithm seeking.Secondly,to improve the exploration and exploitation ability of the algorithm,adaptive weights combined with the population success rate are used to strengthen the bracketing search ability of algorithm,while the exploration and exploitation ability of the algorithm is strengthened by two circular search mechanisms in the bracketing search process to improve the convergence speed and the accuracy of the algorithm.Finally,normal variation is added to perturb the position of elite individuals to drive the whale populations that may fall into stagnation to avoid the algorithm from falling into local optimum.The simulation experiments are compared with several well-known improved whale algorithms and classical intelligent optimization algorithms in 13 benchmark test functions,and the comparison results show that CSOWOA has a significant enhancement effect.t.