Online freight exchange (OFEX) platforms serve the purpose of matching demand and supply for freight in real time. This paper studies a truck routing problem that aims to leverage the power of an OFEX platform. The OFEX routing problem is formulated as a Markov decision problem, which we solve by finding the bidding strategy at each possible location and time along the route that maximizes the expected profit. At the core of the OFEX routing problem is a combined pricing and bidding model that simultaneously (1) considers the probability of winning a load at a given bid price and current market competition, (2) anticipates the future profit corresponding to the current decision, and (3) prioritizes the bidding order among possible load options. Results from numerical experiments constructed using real-world data from a Chinese OFEX platform indicate that the proposed routing model could (1) improve a truck’s expected profit substantially, compared with the benchmark solutions built to represent the state of the practice, and (2) enhance the robustness of the overall profitability against the impact of market competition and spatial variations.
A metaheuristic method for the multireturn-to-depot petrol truck routing problem with time windows
