A Partial Demand Fulfilling Capacity Constrained Clustering Algorithm to Static Bike Rebalancing Problem

Abstract

Nowadays, bike sharing systems have been widely used in major cities around the world. One of the major challenges of bike sharing systems is to rebalance the number of bikes for each station such that user demands can be satisfied as much as possible. To execute rebalancing operations, operators usually have a fleet of vehicles to be routed through stations. When rebalancing operations are executing at nighttime, user demands usually are small enough to be ignored and this is regarded as the static bike rebalancing problem. In this paper, we propose a Partial Demand Fulfilling Capacity Constrained Clustering (PDF3C) algorithm to reduce the problem scale of the static bike rebalancing problem. The proposed PDF3C algorithm can discover outlier stations and group remaining stations into several clusters where stations having large demands can be included by different clusters. Finally, the clustering result will be applied to multi-vehicle route optimization. Experiment results verified that our PDF3C algorithm outperforms existing methods.