Research On Demand Forecasting And Rebalancing Strategy Of Free-floating Bike Sharing System

In recent years,with the continuous development of shared mobility,represented by bike sharing and car sharing,new vitality has been injected into the urban transportation system,which makes the urban transportation services more diversified and intelligent.As a popular short-distance travel tool,sharing-bike has significantly inproved the "last kilometer" connection with urban transit system,effectively alleviated urban traffic congestion,and made the concept of green,low-carbon and healthy travel popular.Nevertheless,the application status of sharing bike in China highlights a series of problems,which are mainly reflected in the lack of rational planning of sharing-bike parking points,excessive release of sharing-bike,unbalanced distribution of sharing-bike in time and space,lack of credit management,etc.Therefore,theoretical research on sharing-bike has far-reaching significance.Based on the research of sharing-bike demand forecasting and rebalancing method,this paper focuses on the function and development orientation of bike sharing system.Based on the analysis of the sharing-bike history data,the travel characteristics of sharing-bike users are analyzed,the construction method of sharing-bike rebalancing network is proposed,and the demand forecasting model of sharing-bike rebalancing is established.User-based and operation-based enterprises are proposed respectively.Sharing-bike rebalancing model is established and the algorithm is designed.Finally,an example is given.The main work of this paper are as follows:(1)Considering that the bike sharing system is a discrete,strongly coupled,non-linear and stochastic complex system,based on the idea of traffic zone partition and clustering analysis,a method of region partition and spatio-temporal clustering for bike sharing system is proposed to realize the division of sharing-bike rebalancing stations and simplify the network of bike sharing system.(2)The construction principle of sharing-bike rebalancing network is put forward.Based on complex network theory,the topology structure and four basic centrality indexes of sharing-bike rebalancing network are analyzed.On this basis,a centrality measurement method based on modularity is proposed to evaluate the importance of sharing-bike rebalancing stations in the network and determine the network dispatching center.(3)Forcasting the short-term rent/return demand of sharing-bike by using wavelet neural network,the priority indexes of rebalancing is proposed to apply different rebalancing strategies:user-based rebalancing strategy,price incentive and credit penalty measures are used to guide users to rent and return bikes from different places to achieve the purpose of rebalancing,which conforms to the operation characteristics of bike sharing system and can significantly reduce operating costs;as for rebalancing strategy based on operators,use dispatching vehicles to redistribute sharing-bike to the network to meet user needs.Aiming at minimizing rebalancing cost,this paper constructs a sharing-bike dispatching path optimization model with soft time-window constraints,and designs a genetic algorithm to solve it.