A stochastic Petri net model for simulation-based performance analysis of public bicycle sharing systems

Abstract

The essence of a public bicycle sharing system (PBSS) is very simple: to allow someone to pick up a bicycle at some point of a region and return it to another place. A redistribution vehicle is used to move bicycles from stations that are almost full to those almost empty; this is usually referred to as balancing operation. This poses an important research problems, i.e., the modeling of PBSS suitable for simulating balancing strategies prior to its application to the real systems. In this paper, we address the problem of modeling PBSS using a generalized stochastic Petri net model. The proposed model has the following advantages: (i) PBSSs with arbitrary number of stations can be modeled; and (ii) it takes into account the stochastic characteristics of PBSSs. Based on this model, an algorithm for carrying out simulations of PBSSs to validate designed balancing strategies has been developed. The model and the simulation algorithm have been applied in the PBSS to be established at the Federal University of Rio de Janeiro.