Research On Demand Forecasting And Repositioning Optimization Of Shared Bikes Considering Carbon Emissions

Shared bikes undertakes the service functions of short distance travel and connecting with public transport,which has great potential in energy conservation and emission reduction.Due to the uneven distribution of shared bikes often occurs in the operation process,the operating enterprises need to reposition them to meet the user demand.However,the widespread use of fuel vehicles in repositioning operation will undoubtedly increase energy consumptions and carbon emissions.Therefore,this paper applies electric vehicles to the repositioning of shared bikes to provide decision support for operating enterprises to realize low-carbon and efficient repositioning of shared bikes by studying the demand forecasting model and low-carbon repositioning optimization model of shared bikes.The main research contents and conclusions are as follows:Firstly,based on the multi-source data such as shared bike order data,electronic fence data and point of interest data,the macro distribution characteristics and micro evolution rules of shared bike activity patterns from the space-time dimension respectively are mined.The results show that the activity patterns of shared bikes have obvious spatio-temporal imbalance and spatial correlation during different hours in working days and rest days.Secondly,based on the spatio-temporal evolution rules of the demand for shared bikes in electronic fences,the LSTM-GRU hybrid forecasting model is constructed to forecast demand for shared bikes in electronic fences.The results show that compared with the single model of LSTM and GRU,the MSE of LSTM-GRU hybrid forecasting model decreases by 16.26%and 46.66%respectively,and R~2 increases by 3.64%and17.93%respectively,which verifies that the hybrid forecasting model can effectively improve the forecasting accuracy and it can provide data support for determining the repositioning demand of electronic fence.Then,based on the demand forecasting results,the repositioning sub-area division model is constructed according to the spatial distance between electronic fences and the complementary degree of repositioning demand and the improved K-means algorithm is used to solve the model.On this basis,considering the load constraints of vehicles,electric quantity constraints and time window constraints,the repositioning optimization model of shared bikes is constructed with the objective function of minimizing the cost composed of fixed cost,route cost,time penalty cost and carbon emission cost of vehicles,and the genetic algorithm is designed to solve the model.Finally,taking the shared bike system of Yanqing district as an example,the electronic fences are divided into three sub-area based on the repositioning sub-area division model,compared with the traditional division model,the model constructed in this paper reduces the average repositioning demand imbalance by 58.90%,which verifies that the application of the model can effectively improve the repositioning demand imbalance in the sub-area.The optimal repositioning scheme and optimal repositioning cost of shared bikes are obtained through programming,and the results are compared with the repositioning scheme by using fuel vehicles.The results show that under the same repositioning task,electric vehicles reduce carbon emissions by about 32%compared with fuel vehicles,which verifies that the use of electric vehicles can significantly reduce the cost and environmental pollution in the repositioning process of shared bikes,which is more in line with the development direction of low-carbon transportation.