基于灰狼算法的改进研究

Study on improvement of gray wolf algorithm

针对灰狼算法具有易陷于局部最优并且收敛速度不理想的缺点,提出基于改进收敛因子策略和引入动态权重策略以及两种策略混合改进的灰狼优化算法,并且用于求解函数优化问题。提出一种非线性收敛因子公式,能够动态地调整算法的全局搜索能力,引入的动态权重使算法在收敛过程中能够加快算法的收敛速度。通过15个基准测试函数验证改进后算法的全局搜索能力、局部搜索能力与收敛速度,实验结果表明,改进后的算法无论在搜索能力还是收敛速度上都强于标准灰狼算法。

According to gray wolf algorithm is easily trapped in local optimum and the convergence rate is not ideal,based on the improved convergence factor strategy and dynamic weighting strategy and two mixed strategies,this paper improved the wolf optimization algorithm and used to solve the function optimization problem. This paper proposed a nonlinear convergent factor formula,which could dynamically adjust the global searching ability of the algorithm,and introduced the dynamic weight to accelerate the convergence rate of the algorithm. 15 benchmark test functions verified the global search ability and local search ability and convergence speed of the improved algorithm. The experimental results show that the improved algorithm is bettter than the standard wolf algorithm in terms of search ability and convergence rate.