基于饥饿变异检索与局部小生境的麻雀算法

Sparrow Algorithm Based on Hunger Mutation Retrieval and Local Niche

针对麻雀搜索算法在迭代过程中在原点与最优解徘徊,造成迭代代价高,在迭代后期,直接跳跃到极值点附近,易于陷入局部最优,且后期种群多样性减小,全局搜索能力差的问题,提出一种基于饥饿变异检索与局部小生境的麻雀算法。首先改进种群中发现者位置迭代方法,提出饥饿短距变异机制,添加变异因子减少在原点与已有最优解的跃迁,提高全局搜索能力;其次改进跟随者位置迭代过程,引入局部小生境概念,将最优解发现者周围划定一个局部小生境,在小生境范围中的麻雀集中搜索最优解附近,小生境外麻雀逃逸到其它位置进行搜索,增强算法最优解附近的搜索能力与增强算法跃出局部空间的能力。对16个基准函数进行仿真对比实验,结果表明提出的基于饥饿变异检索与局部小生境的麻雀算法能够提高算法的搜索精度、收敛速度和稳定性。

The sparrow search algorithm hovers between the origin and the optimal solution during the iteration process,resulting in high iteration costs.In the later stage of iteration,it directly jumps to the vicinity of the extreme point,making it easy to fall into local optima.In addition,the population diversity decreases in the later stage,and the global search ability is poor.A sparrow algorithm based on hunger mutation retrieval and local niche is proposed to address the above issues.First,improve the position iteration method in the population,put forward the hunger shortrange variation mechanism,and add variation factors to reduce the hopping of the origin and the probable solution,im⁃prove the global search ability;Secondly,improve the follower position iterative process,introduce local small habitat concept to delineate a local small habitat around the best solution,search for the best solution in the sparrow concen⁃tration in the small habitat range,and the small abutment of the sparrow escape to other locations to search,improve the search ability near the optimal solution of the algorithm and enhance the algorithm's ability to jump out of local space.This paper simulated the comparative experiment of 16 reference functions,which indicates that the sparrow al⁃gorithm based on hunger variation and local small habitats proposed herein can improve the search accuracy,conver⁃gence speed and stability of the algorithm.