A Behavioral Probabilistic Model of Carrier Spatial Repositioning Decision-Making

This paper studies the truckload market with carriers providing transport services between two locations. It aims to provide a modeling methodology to represent the spatial behavior of a carrier dealing with the issue of repositioning. Indeed, due to the imbalance of trade, carriers face the difficulty of finding freight for their return trips. When they operate over long distance shipments, repositioning their empty vehicles from the low - demand zone is necessary to sustain their business. Yet the mechanisms at stake by carriers to understand their repositioning decision - making process are mostly unknown and unobservable. This lack of data on carrier repositioning zone choice issues has major consequences for shipper and forwarder resource planning system. Indeed, repositioning behavior induces hidden costs which makes it difficult to design for example a cost - based pricing strategy. To address this problem we develop a mathematical model to study the spatial repositioning behavior of carriers. We propose a probabilistic approach based on aggregated transport data that consists in a two-steps decision making process. The first one is the probabilistic selection of a set of repositioning candidates based on the microeconomic theory of the consumer. The second step is the choice of a region within this set through the estimation of the spatial distribution of reloading. It makes use of the graph structure of the transport data and combines a spatial interaction model and a random walk model on a graph. Using simulations, we illustrate how our methodology can be used for operational purposes to provide more transparency on carrier behavior. In conclusion, research perspectives are suggested for tackling the problem of freight demand estimation as well as rationalizing the impact of the trade imbalance on the price of a transport. Software development perspectives will also be addressed