Dynamic programming and lower-bound approaches to the minimum binding problem

Abstract

The Minimum Binding Problem is that given a 0-1 matrix, a sequence of columns is found such that in each row 1 's appear as consecutively as possible. A dynamic programming algorithm succeeds in solving the problem for limited instances. For larger instances, we consider obtaining lower-bounds. A strategy to improve the lower-bounds is also discussed. Computational experiments show the effectiveness of our lower-bounds.