Considering user behavior in free-floating bike sharing system design: A data-informed spatial agent-based model

Abstract

Although bike-sharing has been recognized as an active and sustainable transportation mode, the dramatic expansion of free-floating bike sharing (FFBS) services generates problems such as illegal parking and low utilization. An effective FFBS system needs to be highly regulated. This study combines Big Data and spatial agent-based modeling to understand the interactions between stakeholders to assist the bike-sharing system design. The key design decisions considered are the locations and capacities of bicycle parking lots in the system, as well as the connected bike lanes between parking lots. The model has been applied to the case of Hong Kong for demonstration. The results show that the parking lots with higher capacities are mostly close to the metro stations, and the cycleways are disconnected even for those that have high cycling occupancy. The results indicate that for most target people to be willing to change the parking location, the minimum fare discount rate for doing so should be set to 30%. The average trip time can be reduced by 3.8%, and per user cost can be reduced by 2.4% with an expected investment of 0.12 million USD to build new cycle tracks and connect existing cycleways.