一种基于改进BA算法的K-means聚类

摘要：针对传统BA（蝙蝠）算法易被局部极值吸引、发生过早收敛等问题，将莱维飞行搜索策略引入传统BA算法对蝙蝠的位置和速度更新方式进行改进，从而提高算法的全局搜索能力；通过引入非线性惯性权重平衡算法的全局和局部搜索能力并提高算法搜索精度；结合Limit阈值的思想避免算法过快陷入局部极值。通过对6个标准测试函数的实验表明，改进后的BA算法不仅在全局搜索能力上有所提高，而且具有较好的搜索精度。最后将改进后的BA算法同K-means聚类算法进行结合，提出了一种基于改进BA算法的K-means聚类算法，实验结果表明，改进的算法提高了聚类准确率及算法鲁棒性。

A K-means clustering algorithm based on improved BA algorithm

Abstract: The traditional bat algorithm is easy to fall into local optimum and premature convergence solution, in order to improve the global search capability of the bat algorithm, this paper consider to take Lévy flight search strategy into the bat algorithm to update bat''s position and velocity, at the same time by introducing the nonlinear inertia weight to balance the global and local search capability to improve the search precision of the algorithm . Then combining the Limit threshold theory to avoid getting trapped into local optima. The results on six standard test functions show that the improved bat algorithm not only improves the global search ability and has better accuracy. Finally, combining the improve bat algorithm with K-means clustering algorithm, this paper proposed a K-means clustering algorithm based on the improved bat algorithm. The experimental results show that the algorithm improves the clustering quality and the robustness of the algorithm