The capacitated edge activation (CEA) problem consists of activating a minimum cost set of capacitated edges to ensure the routing of some traffic demands. Most of the MIP-based heuristics proposed for network design problems are based on the so-called flow formulation that includes both activation and routing variables. Indeed, there also exists a capacity formulation that includes only activation variables. This formulation is, however, non-compact. Here, we investigate the price to pay to use the non-compact capacity formulation instead of the compact flow formulation in a MIP-based rounding heuristic for the CEA problem. Both splittable and unsplittable flows are considered. The experiments show that, indeed, the capacity formulation requires more time and solves less instances than the flow formulation, due to the time spent in separating feasibility cuts, in particular for unsplittable flows.
Capacitated edge activation problem Capacity formulation Heuristics
This is a preview of subscription content, log in to check access.
The author has been partially supported by Grants MIUR PRIN 2015B5F27W and 20153TXRX9.
Compliance with ethical standards
Conflict of interest
The author has no conflict of interest concerning this study.
Human and animal participants
This article does not contain any studies with human participants or animals performed the author.
Addis B, Capone A, Carello G, Gianoli L, Sansó B (2014) On the energy cost of robustness and resiliency in IP networks. Comput Netw 75(Part A):239–259CrossRefGoogle Scholar
Crainic T, Gendreau M, Hernu G (2004) A slope scaling/lagrangian perturbation heuristic with long-term memory for multicommodity capacitated fixed-charge network design. J Heur 10:525–545CrossRefzbMATHGoogle Scholar
Crainic T, Li Y, Toulouse M (2006) A first multilevel cooperative algorithm for capacitated multicommodity network design. Comput Oper Res 33:2602–22CrossRefzbMATHGoogle Scholar
Mattia S, Poss M (2018) A comparison of different routing schemes for the robust network loading problem: polyhedral results and computation. Comput Optim Appl 69:753–800MathSciNetCrossRefzbMATHGoogle Scholar