Repositioning activities for bike-sharing systems

Abstract

Bike-sharing systems have become increasingly popular to cities all over the

world, which allow people to rent a bike from any self-served station for short trips

around the city and then return it to any other station. Obviously, satisfying the

fluctuating and asymmetric demand is a crucial problem for a bike-sharing system.

It could be handled through repositioning activities by a fleet of trucks. As moving

docking stations is a development trend, in this thesis the repositioning activities are

defined as transferring both bikes and lockers from some stations to others to

balance the system.

In this thesis, two models are formulated: the "Single Truck Single Period"

model and the "Single Truck Multiple Periods" model. The illustrations of

numerical experiments for the two models are presented, respectively. Through the

sensitivity analysis of the "Single Truck Single Period" model, the test results

indicate how some important parameters affect the performances of the model. At

last, some insights are put forward according to the characteristics of the models

and the test results.