A Stochastic, Dynamic Network Model for Railroad Car Distribution

Abstract

Freight cars represent a substantial portion of total railroad investment, and the low utilization of this major capital resource is certainly a contributing factor in the poor financial health of many railroads. A key element in improving utilization is the process of empty car redistribution. This paper describes a dynamic network optimization model for distribution of empty freight cars. It incorporates information on potential revenues from loads available at various points in the network for each period over a specified planning horizon, the costs of not being able to satisfy these demands, the costs of holding empty cars in anticipation of demands, and the costs of moving empty cars over the system. The information on supplies of, and demands for, cars in future periods, as well as their travel times over the network, is represented stochastically to reflect the uncertainty inherent in forecasts of future conditions.