Short-Term Travel Volume Prediction and Delivery Volume Prediction of Shared Bicycles Built on Different Sites

With the progressive development of the “sharing economy”, sharing bicycles are becoming increasingly popular in cities, and citizens’ last-mile travel problems are greatly alleviated. In addition, bicycles are also a good solution for short-distance travel. However, sharing bicycle operators do not have a good approach to dealing with supply and demand relationships, and bicycles are distributed in a position where supply is greater than demand, making it convenient for considerable discounts. The study aimed to establish a predictive model for predicting the short-term travel of bicycles to be shared in a particular area of the city. Shared bicycle data from 10 to 19 May 2017 were collected from the online platform of the Moby Algorithm Challenge, selecting data from land ranging from 152Mx152M, with the number of bicycles used reaching 2484561. ADF testing is designed to ensure that the parameter input in the model maximizes the accuracy of the model. The ARIMA model is then used to match the data and obtain a highly matching parameter model. As a result, the fixed value of the model is relatively close to the distribution state of the true value. The model predicts data for the next 12 days and obtains relatively accurate predictions.