基于自适应子种群和动态反向学习的改进鸡群算法

An Improved Chicken Swarm Optimization Algorithm based on Adaptive Subpopulation and Dynamic Reverse Learning

针对鸡群算法(Chicken swarm optimization,CSO)求解复杂高维问题收敛精度低、容易陷入局部极值等问题,提出了一种基于自适应子种群和动态反向学习的改进鸡群(ICSO)算法.根据鸡群算法迭代进化进程,自适应确定公鸡种群规模大小,并据此将母鸡种群和小鸡分成若干个子种群;设计进化停滞判定机制,并引入动态反向学习因子以改进算法个体更新方式,有效保持鸡群样本多样性和算法全局深度搜索能力.典型测试函数仿真实验结果表明,与SFLA算法、PSO等智能优化算法相比,ICSO算法具有更高的收敛精度和更优的复杂函数优化能力.

An improved chicken swarm optimization algorithm based on adaptive subpopulation and dynamic reverse learning(ICSO) is proposed to solve complex high-dimensional problems,such as low convergence accuracy and easy to fall into local extremum.According to the iterative evolution process of the chicken group algorithm,the size of the cock population is adaptively determined,and the hen population and the chick are divided into several subpopulations.The evolution stagnation judgment mechanism is designed,and the dynamic reverse learning factor is introduced to improve the individual update mode of the algorithm,effectively maintaining the diversity of the chicken group samples and the global depth search ability of the algorithm.The simulation results of typical test functions show that ICSO has higher convergence accuracy and better complex function optimization ability than other intelligent optimization algorithms such as SFLA algorithm and PSO.