Demand-to-Train Allocation in a Hub-and-Spoke Network

everal planning and scheduling problems are involved within rail freight transportation planning. One of these problems is the demand-to-train allocation problem, which is the assignment of demands to trains. In this thesis a method is presented to solve this demand-to-train allocation problem for a hub-and-spoke network. The problem is divided into two subproblems: the blocking problem and the block-to-train assignment problem. Hereby, demands are first combined into blocks, after which these blocks are assigned to trains. A heuristic is presented to solve the blocking problem, while we also investigate the case where we do not create blocks and solve the demand-to-train allocation problem directly. The block-to-train assignment problem is formulated as an IP model. For this problem a Lagrangian heuristic and a greedy heuristic are created to solve the problem. The IP model and heuristics were tested on several problem instances with different problem sizes, in case of blocking and in the case where the demand-to-train allocation problem is solved directly. For all problem instances it was better, in terms of both solution value and computation time, to first create blocks instead of solving the problem at once.