关键词: 城市交通, 共享单车动态调度, 双层规划模型, 无桩型共享单车, 多主体仿真

Abstract:

The spontaneous imbalance of the bike- sharing system creates an offset between the distribution of bikes and demands, and reduces the service capacity of the system. Bike reposition is needed to achieve rebalancing of the system. Existing dynamic reposition algorithms haven't considered the "re-take" demand during the trip due to insufficient supply at the origin. It is difficult to accurately identify the real distribution of requests and reposition effect was thus reduced. This paper proposed a bi-level programming model with user choice behavior as the lower layer and truck path planning as the upper layer. The heuristic algorithm with a simulator was used to solve the model. The case study was based on the historical data of Hongkou, Yangpu districts of Shanghai, China, and was grid- processed. The results show that the proposed model can effectively identify the shifting demand and improve the demand- supply matching level. The matching ratio has been improved by 18.07% to 19.89% under different reposition resource configurations, and the efficiency of bike- sharing system management has been enhanced to some extent.

Key words: urban traffic, dynamic sharing-bike repositioning problem, bilevel programming model, free-floating bike-sharing system, multi-agent simulation