Probabilistic Petri Nets Model for Assessing Temporal Variations of the Number of Bikes in Bike-Sharing Stations

Bike sharing systems belong to a type of mobility solutions that are increasingly of relevance when facing the issues of crowded traffic, pollution and sedentary lifestyle. Demand prediction is one of the main directions for research done on bike sharing systems aiming to address the existing problems regarding bike station overfilling and emptying. This paper proposes a probabilistic model for the evolution of available resources at station level. The method presented is tested on a compact set of time intervals. Both the tests and the validations are performed on real world data obtained from Citi Bike New York bike-sharing system.