Affiliation:
1. Department of Mathematics, Simon Fraser University, Surrey, BC V5A 1S6, Canada
Abstract
In this paper, we study the knapsack problem with conflict pair constraints. After a thorough literature survey on the topic, our study focuses on the special case of bipartite conflict graphs. For complete bipartite (multipartite) conflict graphs, the problem is shown to be NP-hard but solvable in pseudo-polynomial time, and it admits an FPTAS. Extensions of these results to more general classes of graphs are also presented. Further, a class of integer programming models for the general knapsack problem with conflict pair constraints is presented, which generalizes and unifies the existing formulations. The strength of the LP relaxations of these formulations is analyzed, and we discuss different ways to tighten them. Experimental comparisons of these models are also presented to assess their relative strengths. This analysis disclosed various strong and weak points of different formulations of the problem and their relationships to different types of problem data. This information can be used in designing special purpose algorithms for KPCC involving a learning component.
Funder
Natural Sciences and Engineering Research Council
Reference49 articles.
1. Bandyapadhyay, S. (2014). A variant of the maximum weight independent set problem. arXiv.
2. Maximum weighted independent with a budget;Gaur;Proceedings of the Algorithms and Discrete Applied Mathematics: Third International Conference, CALDAM 2017,2017
3. The exact weighted independent set problem in perfect graphs and related classes;Milanic;Electron. Notes Discret. Math.,2009
4. Local branching-based algorithms for the disjunctively constrained knapsack problem;Akeb;Comput. Ind. Eng.,2011
5. A reactive local search algorithm for the disjunctively constrained knapsack problem;Hifi;J. Oper. Res. Soc.,2006