Research On Scheduling Of Shared Bicycles Based On Mobike Travel Data

Since 2016,shared bicycles with shared economic attributes have been born.Shared bicycles have generated tremendous energy in building a green transportation system and promoting the transformation and upgrading of traditional industries,and more effectively solve the problem of short-distance travel in urban transportation.At the same time,sharing bicycles has also become the focus of risk capital chasing,and various brands of bicycles have been put into the market.Over time,users use shared bicycles in different spatial locations,causing them to shift in space,resulting in uneven spatial distribution.The root of this problem is that there is no scientific and reasonable scheduling of shared bicycles,so scientific research on the scheduling of shared bicycles is needed.Based on SQL Server database,data mining,data visualization and other technologies,this paper studies the actual travel order data of Mobike bicycles from May 10 to May 24,2017 in Beijing,and extracts the use of shared bicycles with different attributes of weeks and hours.Intensity and travel trajectory data,time and space visualization analysis,the distribution of shared bicycle usage in the working day is characterized by early,middle and late peaks.The peak period duration is about one hour,and the morning and evening peaks are concentrated in rail transit.Around the site,the travel trajectory is a state in which a part of the trajectory is concentrated and spreads from the inside to the outside.Combining the above-mentioned shared bicycle time and space distribution characteristics,the GeoHash algorithm is used to extract the virtual scheduling area,and the shared bicycle demand forecasting model based on BP neural network is established to predict the scheduling amount of the scheduling area.The average absolute error of the prediction result is 0.0530,the average absolute percentage error.For the 5.6768%,the model prediction fits well.According to the predicted results of bicycle demand in the peak period of each dispatching area,the problem of shared bicycle supply and demand matching and scheduling optimization is further discussed,reasonable model assumptions and constraints are proposed,and the shared bicycle peak with minimum fixed scheduling cost is constructed.Period scheduling model.Next,the genetic algorithm and the improved genetic algorithm are designed to solve the above models separately.Finally,the effectiveness of the model is verified by the scheduling of Mobei bicycles in Beijing.A shared bicycle scheduling model based on spatio-temporal distribution features is constructed.Two intelligent algorithms are used to solve the model,and the optimization speed of genetic algorithm is improved compared with traditional genetic algorithm.The improvement is nearly 40%,and the scheduling cost is saved by 8.7%,which verifies the feasibility of solving the algorithm and improves the advancement of the genetic algorithm.