Optimising station stocking in bicycle-sharing systems

Abstract

The main aim of this study is to explore the application of Bicycle use models, Machine learning and bicycle-sharing system (BSS) restocking approaches that could increase bicycle availability and utilisation. The main approach is centred around the modelling, analysis, predictive techniques and restocking alternatives based on a dataset with around 8 million records from a full year of bicycle-sharing system operation on 1127 stations for New York City's most popular bike-sharing service, Citi Bike, as a surrogate for system demand. Users in BSS instigate imbalance in the system by creating demand in an asymmetric pattern, which may lead to unavailability of bikes when demand is high and thus potential loss of sales. In some situations, only restocking a BSS (i.e. replenishing bicycle stations with its original inventory) after regular operational hours is not enough to warrant a well-balanced system. This study focuses on the exploration of different means of how such efficiency could be increased, where the main target is the rebalancing of stocking/inventory in bike stations. A precursor to this is the prediction of trip demand at a station level. In this context, demand refers to the number of bicycles taken out of a station at a given time. Knowing the number of future trips that might occur will aid in the planning of when and how to restock during system operation. Hence, a data analytic approach with an element of machine learning was applied to the problem. Data exploration was used to identify patterns and relationships in trip usage, and machine learning techniques were necessary for prediction of hourly demand of trips per station at a particular hour of a day. In addition, this study presents a selection of bicycle restocking strategies in the form of algorithms, as well as an array of predictive models used to predict this hourly demand. The outcomes of the optimised best predictive model, combined with restocking algorithms determine if system efficiency can be improved (in terms of fewer opportunities lost in bike deposit and pickup) which would optimistically result in better availability of bikes in stations. This, in turn, could possibly increase customer satisfaction and which could result in more sales.