Research On The Balance Problem Of Modern Bike Sharing System

In recent years,a new economic model has emerged in the wave of mobile Internet:the sharing economy.The sharing economy aims to provide some shared resources for residents in the town and bring some convenience to people,and the modern bike sharing system is a concrete example of sharing economy.In the increasingly congested cities of the 21 st century,the modern bike sharing system has risen rapidly in the city,greatly facilitates the short-distance travel of urban residents.Therefore,it is very popular among urban residents.Although bike sharing system is convenient for people's lives,how to manage a large number of bikes in a city reasonably and efficiently is a very challenging problem.Because the user's behavior will inevitably lead to the imbalance of supply and demand in different areas.Different from the traditional site-based bike sharing system,the modern bike sharing system is more in line with the definition of “sharing”.In the modern bike sharing system,the bike can be parked anywhere in the city,and its parking position depends on the last user who uses the bike.The modern bike sharing system is more friendly to users,but it greatly increases the difficulty of distribution and management.To this end,this paper deeply studies the problem and proposes a two-step solution.First,by predicting the bike flow in different areas in the next period of time,analyze the user demand in each area,and then use the bike service provider's carriers to carry the extra bikes from the small demand area to the area which need more bikes.Two-step solutions can be modeled separately as flow prediction and path planning problem.For the first step of the bike flow prediction problem,this paper proposes a Spatial-Temporal Bike Flow Prediction Model(ST-BFP)to predict the bike flow in each area.ST-BFP captures spatial associations between regions over traditional time series predictions,while using additional factors that influence user usage to simulate user behavior.In the path planning problem of the second step,in order to restore the bike sharing system to balance state as soon as possible,this paper proposes a Multiple-Carriers Improved Local Search Algorithm(MC-ILSA).MC-ILSA is able to plan a relatively short path for the carriers.Finally,this paper uses real data to carry out experiments and verify the two-step solution.ST-BFP has higher prediction accuracy than some flow prediction methods,and The MC-ILSA also significantly reduces the time spent on the carriers to complete the handling task.