A dynamic programming model for operation decision-making in bicycle sharing systems under a sustainable development perspective

Abstract

Maintaining a balance between revenue and expenditure is the key to the sustainable development of a bicycle sharing system (BSS), and is a challenge for almost all systems worldwide. This article proposes a dynamic programming approach to obtain the optimal strategy to maximize the revenue of overall BSS. The Variable Granularity-Depth First Search (VG-DFS) algorithm is designed to speed up the solution. A numerical experiment is presented to verify the applicability of the model through a comparison with real data from the BSS in Hangzhou. Results indicated that the BSS could achieve break-even, or even obtain a substantial income by utilizing our model to make operational decisions, especially when the region it is located in has a relatively high GDP. Moreover, the best investment strategy proved is to involve stations in the initial construction period of the BSS as much as possible.