Research On The Demand Forecast And Scheduling Of Bike-sharing

In the world,the traditional public bicycle service has many disadvantages,and the pain point of the "last mile" is hard to be solved.The sharing bike which was initiated by Chinese enterprises,and has developed for several years,proposed a new solution for the world to solve the problem of "last mile" and changes the way of urban travel.Bike sharing is in line with the development concept of innovation,harmony,green,open,sharing,widely accepted by the society.The update speed of hardware is far beyond expectations,sharing bike has become one of the main forms of urban traffic.However,the research on the operation and management of sharing bike is lagging behind the technological progress and market demand,which has become a bottleneck of restricting the development of sharing bike.The study of sharing bike demand forecasting and scheduling is beneficial to the series problems of the operation and management of sharing bike,and help to share continuous,stable and healthy development of sharing bike.Consulting the literature at home and abroad,analyzing the existing travel demand prediction and vehicle scheduling research,it is found that the existing travel demand prediction research is based on single time dimension or single spatial dimension,and cannot reflect the unique temporal and spatial characteristics of travel demand.Besides,the existing vehicle scheduling studies usually consider the characteristics of the scheduling problems,such as dynamics and time windows,ignoring the research of scheduling problems with multiple scheduling subjects and various participation modes.The prediction method based on grid theory,time series,constructs a sharing bike demand prediction model,to reveal the spatial and temporal characteristics of bike sharing demand distribution;By forecasting sharing bike demand and construction of vehicle routing optimization model,in order to minimize the total cost of sharing bike scheduling,to study scheduling problems with operators as the main body.Based on the schedule of those cars that will send red packets,this paper analyzes the mechanism of sing-vehicle dispatch for what operators induce the user to do,what's more,this paper summarize the induction results into the starting point drift and the end drift phenomenon,and establish a sharing bike scheduling model for the users;considering operator and user independent scheduling have shortcomings,this paper design the collaborative scheduling mechanism for operators and users,selecting the reward area,appropriate reward efforts in the setting of induced reward,which can guarantee the distribution and prediction of the bicycle demand distribution phase on the basis of matching,and realize the lowest total scheduling cost.Using the v-mobile bicycle travel data,research on the “grid division-demand forecast-single subject sharing bike scheduling-multi subject sharing bike collaborative scheduling "for empirical analysis,results show that the reasonable choice of regional reward and reward efforts,the multi-agent collaborative sharing scheduling strategy can achieve a dual optimal of the total scheduling cost and scheduling efficiency.