Comparing loading strategies for auto-trains: balancing efficiency with information requirements

Auto-trains serve as the favored transportation mode for distributing passenger vehicles, offering cost-effective and environmental benefits over transportation by trucks. This paper addresses the auto-train loading problem (ATLP) with the objective to improve the capacity utilization of auto-trains. We propose five different problem cases of the ATLP with varying levels of available information on the availability and accessibility of the vehicles and develop suitable strategies to solve them. This paper proposes the first contribution to the ATLP in the domain of finished vehicle logistics. The ATLP differs from similar problems, such as the motorail transportation problem, as it involves various precedence constraints. We consider the method and order in which the vehicles are parked as well as the order in which they are loaded on the auto-train. For the first two cases, we approach the problem as a multiple knapsack problem and solve it heuristically using a best-fit algorithm. For the other three cases, we formulate the ATLP as a generalized assignment problem and solve it with a commercial solver. We also develop a rolling-horizon heuristic to address the problem’s size for two of these cases. This research contributes by formally defining the ATLP, offering a spectrum of strategies tailored to varying levels of information, and conducting simulations based on real-world data to quantify potential improvements. The findings indicate that, on average, the objective value of strategies for cases with less available information aligns with historical data, while our strategies for cases with higher information could increase capacity utilization by up to 9.73% compared to historical data.