Network Design in Transportation and Logistics: Models and Algorithms

Network design problems in transportation and logistics are both complex and largescale—often containing millions, or even billions, of decision variables and tens of thousands of constraints. Finding optimal or near-optimal solutions to large-scale network design problems is challenging, if not impossible. To address this issue, we extend and bring together in a unified setting two complementary ideas that have emerged repeatedly in the course of our current research and that of others. These are: 1. Solving network design problems by means of strengthened reformulations in extended variable space (we refer to these reformulated variables as composite variables); and 2. Inducing data-dependent decomposition through bounded approximation.