Moment-based rental prediction for bicycle-sharing transportation systems using a hybrid genetic algorithm and machine learning

Bicycle-sharing systems are a new type of transportation service that provides bicycles for shared use; they allow users to rent a bicycle at one station, ride it, and return it to another station in the same city. A predictive model is needed to forecast the rental demand to improve user satisfaction and increase profits. To effectively predict the rental demand in such bicycle-sharing systems, we propose a moment-based model and a new hybrid approach that combines a fuzzy C-means (FCM)-based genetic algorithm (GA) with a back propagation network (BPN). This FCM-based GA is a new unsupervised classification method that is used to pre-classify historical rental records into groups. The classification results are then fed into a BPN predictor, which is trained using these categorized records. After training, the BPN predictor can predict the demand at future moments. Finally, we present a case study based on real-life data to demonstrate the effectiveness and efficiency of the proposed approach.