Commodity Representations and Cut-Set-Based Inequalities for Multicommodity Capacitated Fixed-Charge Network Design

We improve the mixed-integer programming formulation of the multicommodity capacitated fixed-charge network design problem by incorporating valid inequalities into a cutting-plane algorithm. We use five classes of valid inequalities: the strong, cover, minimum cardinality, flow cover, and flow pack inequalities. The first class is particularly useful when a disaggregated representation of the commodities is chosen, while the last four are expressed in terms of network cutsets. We develop efficient separation and lifting procedures for these classes of inequalities. We present computational results on a large set of instances of various characteristics, allowing us to measure the impact of the different classes of valid inequalities on the quality of the lower bounds, in particular with respect to the representation of the commodities.