Journées de l'optimisation 2007 Optimizations days

This talk will describe a dynamic reassignment procedure to reduce the time that empty cars spend in rail yards. Computational results based on real-life data gathered from one of the major railroad companies in North America will be presented.

Shipping companies must maintain a sufficient number of empty containers in inland depots in anticipation of future loaded movements. They can dispatch empty containers among depots, build stocks, perform substitutions among container types and rent containers, according to rental contracts clauses. We formulate an optimization model to solve this issue.

We are addressing the service network design problem in freight rail transportation. The problem simultaneously considers the service selection (with scheduling), traffic distribution, blocking policy, and train make-up policy. To explicitly describe the flows of services, blocks, and cars/commodities, a cyclic time-space network structure with three layers is constructed. Based on the network structure, we present a linear mixed-integer formulation. Solution method ideas are also discussed.

Empty containers must be repositioned among ports to correct trade imbalances. Several sources of uncertainty affect this issue, like supplies and demands at ports, storage and transportation capacities. Although low-quality statistics are available, we present and solve an optimization model considering several realizations of uncertain parameters through expert-based scenarios.