Abstract: in order to solve the problem of clustering center improper selection in the traditional K-means algorithm which leads to the clustering result into local optimum, the initial clustering center of K-means algorithm is searched by the bat algorithm. The simulated annealing and the niche technology based on crowding out is added into the bat algorithm, in order to overcome some problems such as slow-speed convergence in later and weak search capability, its validity is verified by test functions. Finally the initial clustering center of K-means algorithm is optimized by the improved bat algorithm. The improved algorithm is compared to the traditional K-means algorithm, and the experimental results show that the improved algorithm of clustering performance has improved greater than the traditional K-means algorithm.

Key words: bat algorithm, initial clustering center, K-means clustering