Analyzing bike repositioning strategies based on simulations for public bike sharing systems: Simulating bike repositioning strategies for bike sharing systems

With the contributions on reducing the traffic congestion and air pollution, bike sharing systems become more popular recently in many metropolitan areas worldwide. Without effective bike redistribution strategies, a bike rental station may easily become out or full of bikes, which incurs the customer inconvenience and conflicts its purpose. In order to evaluate the impacts and performance on different bike redistribution strategies, we propose and simulate several bike redistribution strategies with and without different levels of real-time or historical bike rental information. In particular, for a system that conducts no or simple bike repositioning operations, we further consider whether the system is capable of learning preferred bike-return destinations specified by commuters, suggesting bike-return destinations to customers, or exploiting the historical trend of commuter traffics.