Location of a Conservative Hyperplane for Disjoint Bilinear Programming Based on Distances

The generation of a conservative cut is an effective approach for degeneracy removal in cutting plane methods for disjoint bilinear programming (DBLP), but the location of such a hyperplane seems time-consuming. This paper presents several revised distance-following algorithms within one framework in order to identify a conservative hyperplane. We demonstrate their computational efficiency from both the degree of degeneracy and the dimensionality. Numerical results are reported.