SMARTBIKE: Policy making and decision support for bike share systems

Abstract

Bike share systems have been implemented in many cities to provide an alternative low-cost environmentally friendly and healthy public transportation. Citizens can pick up and drop off bikes among stations around the city. The system may be used for both work-related and leisure purposes including supporting last-mile transportation and tourism. In order to sustain and attract more users into the bike system, understanding demand and trip patterns is a key objective. Further, due to the asymmetric and dynamic nature of user demands, some stations tend to run out of bikes whereas other stations tend to be full (impeding returns). Rebalancing stations to provide high quality services to fit pick up and drop off demand is a major challenge in bike share systems. In this paper, we present a three-layer SMARTBIKE system to assist city administrators in Fortaleza, Brazil with bike share system decision making. This is an international cooperative effort between academia and government to aid in building a Smart City. The first layer of the proposed tool includes ?-means clustering technology to understand the true demand of the city bike share system. We designed a novel station network analysis in the second layer to provide insights on system usage patterns and to help with the rebalancing strategy. The third layer provides rebalancing support at the facility and operation level. Finally, we discuss a dynamic visualization tool to support decision making.