Research On Site Selection Planning And Scheduling Path Optimization Of Shared-bike Electronic Fence

Under the background of the sharing economy era,the sharing bicycle system developed with the help of Internet technology has greatly alleviated various "urban diseases" such as traffic congestion,environmental pollution and energy shortage caused by the process of urban motorization,and has realized the perfect connection with the "last kilometer problem" which is difficult to be solved by other public transportation modes.However,with the increase of the number of users and the number of bicycles put in,the system also shows many disadvantages.Some users do not violate the rules of disorderly parking after completing their travel needs,which not only encroaches on the public road resources,damages the city image,aggravates the difficulty of municipal management,and even aggravates the traffic congestion.Solving the problem of standardized parking of shared bicycles has become the top priority of the development of the industry.Some bike sharing enterprises began to deploy dedicated electronic fence stations to meet users’ parking needs and regulate parking behavior.At the same time,due to the lack of accurate prediction of the demand for borrowing and returning of shared bicycle stations,the dispatchers can only use their experience to schedule work at each station,resulting in a large number of stations with "no cars to borrow" or "idle vehicles to pile up",which not only reduces user satisfaction,but also increases the operating cost of the enterprise,so it is necessary to borrow and return shared bicycle stations The prediction and scheduling path optimization strategies are studied.First of all,the paper adopts the strategy of "qualitative selection first,then quantitative determination of site location" for the site selection of shared bicycle electronic fence.On the basis of preliminary selection of excellent alternative sites by AHP,a mathematical model is established with the goal of minimum site construction cost and maximum demand coverage.Secondly,based on the travel record data provided by the 2017 Moby cup algorithm challenge,this paper analyzes the influencing factors of the demand for shared bicycle borrowing and returning from three aspects: time,space and weather.After summarizing the current demand forecasting methods,this paper puts forward the BP neural network time series forecasting method which considers the space,weather and other factors comprehensively,constructs and trains the BP neural network time series model with MATLAB software,and finally verifies the accuracy of the model through the fitting analysis of the borrowed(returned)vehicles in each time period of three representative stations.Based on the accurate prediction of the borrowing(returning)demand of each period of time,the scheduling demand of the station is determined.Taking the total cost of scheduling vehicles and user satisfaction as the optimization objective,combined with the constraints to be met in the actual operation process,the multi-objective shared single vehicle scheduling path optimization model is constructed.Finally,the paper constructs a reasonable example,analyzes the shared single vehicle scheduling path optimization model,and verifies the feasibility and practicality of the model and method.