自适应协同进化蝙蝠算法

蝙蝠算法作为一种新型元启发式进化算法,不可避免在进化过程中存在陷入局部极值的危险.为了有效提高蝙蝠算法的进化性能,提出一种自适应协同进化的蝙蝠算法(ACEBA).为保证算法具有良好的进化结构,提出采用自适应进化种群结构,使得种群结构能够依据种群多样性在集中式结构与分布式结构之间进行切换.为协调实现主种群的勘探和子种群的开采,引入优良个体解对速度和位置进行更新,并在主种群和子种群内采用相适应的更新方式,同时将原有固定参数推广到自适应变化,并对蝙蝠行为的多普勒效应进行补偿.最后对所提出的算法进行收敛性分析和仿真验证,并与相关算法进行对比分析,充分验证了算法的正确性和有效性.

Adaptive cooperation evolutionary bat algorithm

The bat algorithm is a novel meta-heuristic nature-inspired algorithm, and also easy to trap into local optimum inevitably, therefore, the paper proposes an adaptive cooperation evolutionary bat algorithm (ACEBA). In order to ensure the proper framework for the algorithm, the evolutionary framework can be switched between the centralized and distributed framework according to the diversity judgment criteria in order to ensure the favorable evolutionary framework for the algorithm. In order to ensure the exploration ability of the main population and the exploitation ability of the sub-population, the position and velocity for the bat are updated, and the update way in main population is different from the sub-population. The compensation for Doppler effect in echoes is considered and the former fixed constant can change adaptively. Finally, the convergence of the algorithm is also deduced and verified by simulation results show the effectiveness and correctness of the proposed algorithm.