Laplacian eigenvalues and the maximum cut problem

Laplacian eigenvalues and the maximum cut problem

0.00 Avg rating0 Votes
Article ID: iaor19951896
Country: Netherlands
Volume: 62
Issue: 3
Start Page Number: 557
End Page Number: 574
Publication Date: Dec 1993
Journal: Mathematical Programming (Series A)
Authors: ,
Abstract:

The authors introduce and study an eigenvalue upper bound ℝrsquo;(G) on the maximum cut mc(G) of a weighted graph. The function ℝrsquo;(G) has several interesting properties that resemble the behaviour of mc(G). The following results are presented. The authors show that ℝrsquo; is subadditive with respect to amalgam, and additive with respect to disjoint sum and 1-sum. They prove that ℝrsquo;(G) is never worse that 1.131 mc(G) for a planar, or more generally, a weakly bipartite graph with nonnegative edge weights. The authors give a dual characterization of ℝrsquo;(G), and show that ℝrsquo;(G) is computable in polynomial time with an arbitrary precision.

Reviews

Required fields are marked *. Your email address will not be published.