Data-driven modular vehicle scheduling in scenic areas

Abstract

As tourism demand continues to grow and fluctuate, the problems of increasing empty capacity and high operating costs for tourist shuttle buses have become more acute. Modular vehicles, an emerging transport technology, offer flexible length adjustments and provide innovative solutions to address these challenges. This paper develops a data-driven method to address the problem of scheduling modular vehicles in scenic areas with dynamic passenger demand. The aim is to minimize operating costs and maximize vehicle utilization by exploiting the adjustable capacity of modular vehicles. This approach is applied to tourist shuttle scenarios, and a sensitivity analysis is conducted by varying parameters such as individual vehicle capacity and waiting penalties. Then, we investigate the optimization performance gap between the proposed model and the theoretical global optimum model. The results show that increasing vehicle capacity and varying penalties improve the performance of the data-driven model, and the optimization rate of this model can reach 70.2% of the theoretical optimum, quantifying the effectiveness of the model. The method proposed in this study can effectively reduce the operating cost of shuttle vehicles for scenic areas and meet the challenge of unpredictable passenger demand, which serves as a good reference for fleet management in scenic areas.