Setting inventory levels in bike a sharing network

摘要

Bike sharing systems (BSSs) allow customers to rent bicycles at automatic rental stations for short periods and return them to any station within the city. This is an environment-friendly transport that can be easily integrated into the public transport system to increase its utility. The major problem affecting BSSs is the non-availability of bikes and lockers which involves demand for the bicycles and the demand for the lockers. If there are more bikes then there will be less lockers and vice versa. The capacity of each station has to be taken into account. Due to asymmetric demands, there will be a shortage of bicycles or lockers. Several studies have been conducted to provide solutions to this problem. One method is reserving a parking spot at a desired location while renting, which will reduce the uncertainty in demand. Another approach is to choose two destinations so that the system will direct them to one where there is a vacant locker. The most common method adopted to handle this issue is repositioning bikes among stations using trucks to carry several bikes. This is performed as dynamic repositioning or static repositioning. Dynamic repositioning is done when the system is active to react to the repositioning. Static repositioning assumes that the target inventory levels are known. Nair and Miller-Hooks (Ref. 1) considered a stochastic mixed integer program to minimize the cost of redistribution for a required service level. Most of the available studies have considered only one station independently of others and the interactions among stations are not taken into account. One study has taken the multiple stations into account but ignored the interactions among them. This article considers the interactions among BSSs stations to set target inventory levels.