A new branch-and-bound algorithm for the maximum edge-weighted clique problem

San Segundo, Pablo, Coniglio, Stefano, Furini, Fabio and Ljubic, Ivana (2019) A new branch-and-bound algorithm for the maximum edge-weighted clique problem. European Journal of Operational Research, 278 (1), 76-90. (doi:10.1016/j.ejor.2019.03.047).

Record type: Article

Abstract

We study the maximum edge-weighted clique problem, a problem related to the maximum (vertex-weighted) clique problem which asks for finding a complete subgraph (i.e., a clique) of maximum total weight on its edges. The problem appears in a wide range of applications, including bioinformatics, material science, computer vision, robotics, and many more. In this work, we propose a new combinatorial branch-and-bound algorithm for the problem which relies on a novel bounding procedure capable of pruning a very large amount of nodes of the branch-and-bound tree. Extensive computational experiments on random and structured graphs, encompassing standard benchmarks used in the literature as well as recently introduced real-world large-scale graphs, show that our new algorithm outperforms the state-of-the-art by several orders of magnitude on many instances.

Text

mewc_BB_rev - Accepted Manuscript

Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (510kB)