A new approach for clustering problem based on binary small world optimization algorithms

This paper presents a new clustering approach based on the binary small world optimization algorithm (BSWOA). Each node is represented as a binary string which is transformed from a decimal string. The i^th element of the decimal string denotes the group number assigned to object i. An integer vector corresponds to a candidate solution for the clustering problem. A swarm of nodes are initiated and fly through the solution space for targeting the optimal solution. To verify the efficiency of the proposed BSWOA algorithm, comparisons with traditional K-means algorithm and the genetic K-means algorithm are performed. Computational results show that the proposed BSWOA algorithm is very competitive and outperforms traditional K-means algorithm and a genetic K-means algorithm.