Bike sharing systems: stochastic formulation and mean field analysis

This thesis presents a stochastic approach to analyzing a bike-sharing network, composed of a set of stations and a fleet of bikes that travel between them. Analyzing such networks is important for optimizing user experience and minimizing maintenance costs. However, the system’s complexity makes analysis challenging, particularly from a computational point of view. To address this, we study a stochastic model and establish probabilistic results that help understand the system’s behavior and predict its evolution. The work includes the description of the model, the proof of a mean-field limit theorem in a simplified setting, and the examination of a few specific examples.