Location Selection of Shared Bicycle Distribution Points Based on User Demand

Abstract

Shared bicycle systems must select the best delivery point locations to meet user demand. Various types of users have different needs, and here we used the RFM (Recency, Frequency and Monetary) model to classify user types and then used the analytic hierarchy process to determine the weights of user types. We selected the shortest weighted total walking distance of users as the optimization objective, constructed a siting model of shared bicycle delivery points based on user demand, and solved this problem using genetic algorithm. Finally, we validated the model using the bike-sharing data of a region in Shanghai. The results showed that for 600 delivery points, the average walking distance of each user was 206.98 m. Thus, this model can effectively optimize the layout of delivery points of a bike-sharing system.