A branch and bound clustering algorithm

We discuss the clustering problem in a 0-1 matrix in this paper. Although clustering algorithms are available in the literature, many of them cannot produce a solution matrix in a desirable structure. Therefore, additional computation or user intervention is required to obtain submatrices (i.e., clusters) from a solution matrix. To solve the clustering problem effectively, we use a branch and bound approach. The proposed algorithm uses optimal and heuristic branching rules, and therefore, generates optimal and heuristic solutions, respectively. The comparative study shows that our algorithm not only is more efficient but also produces more reliable solutions than many existing algorithms. >