Estimation of Optimal Inventory Levels at Stations of a Bicycle Sharing System

Abstract

Bicycle sharing systems are getting increasingly popular around the world as a sustainable and eco- friendly transportation mode which can address many urban transportation issues. However, the highly asymmetric usage of such systems leads to an imbalance in the system. The agencies try to balance the system by repositioning bicycles within the system. A very good tool for calculating the desired inventory levels based on the future usage of the system is essential for this activity and ensuring a minimum desired level-of-service to the users. This paper proposes a model for predicting the optimal inventory levels at the stations to take care of the future usage using the historic usage data of the system. The usage at the stations is modeled as a Markov process and the expected penalty values are calculated for the different starting inventory levels. A Mixed Integer Program (MIP) is developed to ensure that the number of bicycles in the system is conserved. The approach is tested for a bicycle sharing system in Antwerp, Velo Antwerpen. The computational times are found to be extremely reasonable. The results show that the levels-of-service can be improved and considerable savings can be obtained for the systems by reducing the resources used in the repositioning operation.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 8 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.