Research on Cargo Volume Prediction of Logistics Networks Based on Time Series Processing and Directed Acyclic Graph Model

With the rapid development of e-commerce and the logistics industry, accurately predicting the cargo volume at logistics sorting centers to improve the efficiency and service quality of the logistics system has become an important research topic. In response to this, this paper conducts a comprehensive study. First, logistics operation data from the platform are obtained through the competition's official website, followed by data preprocessing, including time series conversion, wavelet transform denoising, white noise detection, differencing and smoothing, and stationarity tests, to transform the cargo volume data of the logistics sorting center into a stationary series. Next, the ARIMA model is applied to forecast the cargo volume for the next 30 days. Furthermore, a directed acyclic graph model of the logistics sorting center is constructed, and historical data is used to calculate node strength. Future transportation route changes are also considered to update node strength and solve the state transition weight matrix. Ultimately, by using the model, the overall prediction accuracy of the cargo volume reached over 80%, with an average predicted cargo volume of 50677.8kg for December, and a total predicted cargo volume of 1520334kg for the entire month of December. This paper holds significant importance in the study of transportation and cargo volume forecasting within logistics networks.