一种新型的改进果蝇优化算法

A New Improved Fruit Fly Optimization Algorithm

为了克服原始果蝇优化算法在求解多峰值函数和高维度函数时容易早熟、效率低等问题,针对原始果蝇优化算法中存在的弊端,创新性地提出一种基于随机策略及非线性步长递减的果蝇优化算法。随机策略的引入能够增加算法寻优时解的多样性,有效避免算法陷入局部极值。非线性步长递减可以使算法在寻优初期保持较大步长,以提高收敛速度,随着寻优的不断深入,步长值不断减小,以提高算法的寻优精度。将新改进的果蝇算法与原始果蝇算法、改进算法DS-FOA和改进算法LGMS-FOA进行寻优对比,实验结果表明,所提出的算法在寻优精度以及收敛速度上有明显提高。

In order to overcome the original fruit flies optimization algorithm in solving the problem that the multimodal function and high dimension function tends to premature and low efficiency,in view of the deficiencies in the original fruit flies optimization algorithm,innovatively pro poses a fruit flies optimization algorithm based on a random strategy and nonlinear diminishing step size.The introduction of stochastic strategy can increase the diversity of the algorithm's optimal solution and effectively avoid the algorithm falling into the local extremum.Nonlinear diminishing step size can keep relatively large at the beginning of the optimization algorithm.In order to improve the conver gence speed,with the deepening of the optimization,the step size decreases continuously to improve the optimization precision of the algo rithm.Compared with the original fruit flies algorithm,DS-FOA and LGMS-FOA optimization algorithm,the experimental results show that the proposed algorithm is improved obviously on the optimization precision and convergence speed.