A branch-and-cut algorithm based on semidefinite programming for the minimum k-partition problem

Ghaddar B, Anjos M, Liers F (2011)

Publication Status: Published

Publication Type: Journal article, Original article

Publication year: 2011

Journal

Annals of Operations Research Springer Verlag (Germany)

Publisher: Springer Verlag (Germany)

Book Volume: 188

Pages Range: 155-174

Journal Issue: 1

DOI: 10.1007/s10479-008-0481-4

Abstract

The minimum k-partition (MkP) problem is the problem of partitioning the set of vertices of a graph into k disjoint subsets so as to minimize the total weight of the edges joining vertices in the same partition. The main contribution of this paper is the design and implementation of a branch-and-cut algorithm based on semidefinite programming (SBC) for the MkP problem. The two key ingredients for this algorithm are: the combination of semidefinite programming with polyhedral results; and a novel iterative clustering heuristic (ICH) that finds feasible solutions for the MkP problem. We compare ICH to the hyperplane rounding techniques of Goemans and Williamson and of Frieze and Jerrum, and the computational results support the conclusion that ICH consistently provides better feasible solutions for the MkP problem. ICH is used in our SBC algorithm to provide feasible solutions at each node of the branch-and-bound tree. The SBC algorithm computes globally optimal solutions for dense graphs with up to 60 vertices, for grid graphs with up to 100 vertices, and for different values of k, providing a fast exact approach for k≥3. © 2008 Springer Science+Business Media, LLC.

Authors with CRIS profile

Frauke Liers-Bergmann Professur für Optimization under Uncertainty & Data Analysis

Involved external institutions

University of Waterloo (UW) CA Canada (CA)

How to cite

APA:

Ghaddar, B., Anjos, M., & Liers, F. (2011). A branch-and-cut algorithm based on semidefinite programming for the minimum k-partition problem. Annals of Operations Research, 188(1), 155-174. https://dx.doi.org/10.1007/s10479-008-0481-4

MLA:

Ghaddar, Bissan, Miguel Anjos, and Frauke Liers. "A branch-and-cut algorithm based on semidefinite programming for the minimum k-partition problem." Annals of Operations Research 188.1 (2011): 155-174.

BibTeX: Download