A static bike repositioning model in a hub-and-spoke network framework. (September 2020)

• Record Type: Journal Article
• Authors: Huang, Di; Chen, Xinyuan; Liu, Zhiyuan; Lyu, Cheng; Wang, Shuaian; Chen, Xuewu
• Is Part Of: Transportation research. Volume 141(2020)
• Publication Date: 2020-09
• Subjects: Bike repositioning -- Demand forecasting -- Random forests -- Hub-and-spoke network framework -- Hub-first-route-second; Logistics -- Periodicals; Transportation -- Periodicals; 388.011
• DOI: 10.1016/j.tre.2020.102031
• Languages: English
• ISSNs: 1366-5545

Abstract

Highlights: Apply the Random Forest model to predict the customer demand in bike-sharing system. Propose a hub-first-route-second bike repositioning strategy based on the predicted demand. Develop a solution method incorporating the heuristic algorithm (ABC) and the exact method. Abstract: This paper addresses a static bike repositioning problem by embedding a short-term demand forecasting process, the Random Forest (RF) model, to account for the demand dynamics in the daytime. To tackle the heterogeneous repositioning fleets, a novel repositioning operation strategy constructed on the hub-and-spoke network framework is proposed. The repositioning optimization model is formulated using mixed-integer programming. An artificial bee colony algorithm, integrated with a commercial solver, is applied to address computational complexity. Experimental results show that the RF can achieve a high forecasting accuracy, and the proposed repositioning strategy can efficiently decrease the users' dissatisfaction.