| Nowadays,bike-sharing systems have been widely adopted and deployed in major cities around the world,bringing great convenience to people’s lives.However,as a public facility,bicycles are easily damaged and require frequent replacement which will lead to high system maintenance costs.One of the fundamental causes of bicycle damage is the severe unbalance load of bicycle use.In order to inprove the quality of services,and inplement better system planning,it is crucial to obtain accurate user demand for bike-sharing system.Through data mining and data analysis of usage record data in the bike-sharing system,this paper focuses on the research of load balancing and demand estimation,and takes Washington D.C.bike-sharing system as the specific study object.The main research content of this paper consists of three parts:(1)According to the analysis of usage record dataset of the existing bike-sharing system,this paper firstly finds and verifies the load-unbalance phenomenon ofbicycle use in the bike-sharing system(2)Aiming at the research of load balancing in the bike-sharing system,this paper proposes a hybrid bicycle allocation strategy for usage load balancing and lifetime optimization,which is evaluated on Washington D.C.bike-sharing system to verify the feasibility and effectiveness using four metrics.And the results show that the hybrid bicycle allocation strategy could effectively reduce imbalance of system load,at the same time also can greatly reduce the proportion of bicycles need to be replaced in the lifetime optimization of bicycles.(3)Aiming at the research of demand estimation in the bike-sharing system,this paper considers the situations that users are not served by the system occur in real world.Combined with the bicycle usage record dataset,station status dataset and station location dataset,this paper analyzies and verifies the correlation between user demand in time and space,and puts forward a demand estimation model based on this,which is verified on Washington D.C.bike-sharing syste. |