Capacitated path-aggregation constraint model for arc disruption in networks

Path-aggregation constraint (PAC) models can be used to represent flows across networks with specific origin-destination demand pairs without requiring explicit enumeration of all potential paths, greatly reducing the computational burden needed analyze a network's performance in the event of arc disruptions. This paper presents new PAC models for examining arc disruptions on networks, providing the first known models to extend such PACs to capacitated arcs. Extensive computational testing demonstrates that these novel arc-capacitated models do not significantly increase computational burden beyond that required by arc-uncapacitated models, while avoiding the suboptimal performance achieved when uncapacitated models are applied to arc-capacitated networks.