Artificial Intelligence-driven Optimization Methods for Vehicle Scheduling and Dispatching: An Overview and Further Directions

Vehicle scheduling and dispatching are core optimization problems in large-scale vehicle fleet systems, directly influencing service efficiency, operating cost, and resource utilization. These problems require coordinating multiple vehicles under time-varying demand, limited capacity, and strict real-time constraints, which place stringent requirements on computational efficiency and solution robustness. Classical optimization methods from operations research often struggle to meet these requirements at scale, especially when rapid reoptimization is needed in response to disturbances and uncertainty. Recent studies indicate that incorporating artificial intelligence (AI) techniques into optimization frameworks can improve scalability and adaptability while reducing computational burden. This article reviews AI-driven optimization methods for vehicle fleet scheduling and dispatching and organizes the literature into two technical directions: AI-assisted mathematical programming, where learning supports preprocessing, initialization, and solver-level guidance, and AI-enhanced heuristic and metaheuristic optimizers, where data-driven strategies guide operator selection, parameter adaptation, and population management. The review summarizes how AI could support scheduling and dispatching decisions at different stages of the optimization process and outlines emerging research directions that integrate generative models with optimization in large-scale fleet systems.