A Bi-level Approach for Optimizing Rental Station Locations in Public Bicycle System

This paper discusses the problem of optimizing the location of rental station in the public bicycle system. First, the effect of the new-added rental station on the public bicycle system is studied, and a time series model is used to forecast the amount of users (rentals). Second, a bi-level model is established to comprehensively consider the concerns of decision makers and travelers. The upper-level model minimizes the construction cost, and considers the effect of new station on the adjacent stations. The lower-level model minimize the travel time and maximize the degree of satisfaction of bicycle users. Finally, a heuristic algorithm is designed to solve the model. The proposed method was implemented in a real-world public bicycle system in Nanjing, China. Results proved the feasibility of the proposed method, and show that the optimal location of rental station can improve the performance of the public bicycle system.