A Clustering-Based Approach For Balancing And Scheduling Bicycle-Sharing Systems

This paper addresses an inventory regulation problem in bicycle sharing-systems. The problem is to balance a network consisting of a set of stations by using a single vehicle, with the aim of minimizing the weighted sum of the waiting times during which some stations remain imbalanced. Motivated by the complexity of this problem, we propose a two-stage procedure based on decomposition. First, the network is divided into multiple zones by using two different clustering strategies. Then, the balancing problem is solved in each zone. Finally, the order in which the zones must be visited is defined. To solve these problems, different algorithms based on approximate, greedy and exact methods are developed. The numerical results show the effectiveness of the proposed regulation methodology.