A Massive-Cargo Transportation Routing Model Based on Reinforcement Learning

Abstract

Considering the size (length, width, and height) and weight of massive cargos and on the basis of satisfying the throughput capacity of the road network, a massive-cargo transportation routing model was proposed to find the shortest transportation route based on the origin and destination. According to the characteristics of the routing problem, a reinforcement learning algorithm was designed to solve the problem and determine the strategy, reward function, value function, and environment model. Additionally, based on a case study, an optimized transportation plan was developed and the model sensitivity was analyzed to verify the validity and correctness of the proposed model and algorithm.