飞蛾扑火优化算法的研究及改进

Research and Improvement of Moth-Flame Optimization Algorithm

由于标准的飞蛾扑火优化算法(MFO)存在着收敛速度慢、收敛精度低且易于陷入局部最优的缺点,提出了自适应飞蛾扑火优化算法(AMFO).该算法首先在飞蛾与火焰的距离中加入动态自适应步长因子,提高算法的全局寻优的能力;其次,在火焰位置加入动态自适应权重因子,更新寻优方式,从而可以达到全局寻优与局部寻优相平衡,解决易于陷入局部最优的缺陷,使得飞蛾的更新方式更加具有灵活性,促使算法沿着正确的方向进行搜索,因此可以有效地提高算法的精度和收敛速度.采用了8个测试函数对AMFO算法进行仿真实验,结果表明,AMFO算法在收敛速度和收敛精度上都有显著性的变化.

Since the standard Moth-Flame Optimization Algorithm(MFO)has the disadvantages of slow convergence speed,low convergence accuracy and is easy to fall into local optimum,this paper proposes an Adaptive Moth-Flame Optimization Algorithm(AMFO).Firstly,the dynamic adaptive step size factor is added to the distance between moth and flame to improve the global optimization ability of the algorithm.Secondly,the dynamic adaptive weighting factor is added and optimization method is updated in the flame position,so that the global optimization and the local optimization can be balanced,and the defect that is of easy to fall into local optimization can be solved,making the update method of the moth more flexible and prompting search algorithm in the right direction,which can effectively improve the precision and convergence speed of the algorithm.In this paper,eight test functions are used to simulate the AMFO algorithm,and the results show that the AMFO algorithm has significant changes in convergence speed and convergence accuracy.