Logistics Cargo Volume Prediction Model Based on Combined ARIMA-LSTM Prediction Methods

In the new era, the rapid development of information technology, the continuous introduction of policies, and the frequent occurrence of natural disasters, such as earthquakes and typhoons, have made the prediction of cargo volume in sorting centers extremely unstable. Therefore, improving the management efficiency of sorting centers has become particularly important. This paper aims to address this issue by establishing a mathematical model based on historical data to predict future cargo volumes, thereby optimizing personnel allocation and enhancing the overall management efficiency of various sorting centers. First, the study utilizes historical cargo volume data from each sorting center to forecast daily and hourly cargo volumes for the next 30 days. After cleaning and integrating the data, the complexity of cargo volume data, which includes both linear and nonlinear relationships, is considered. Traditional single models often struggle to capture this complexity adequately. Hence, this paper adopts a hybrid ARIMA-LSTM model for daily cargo volume prediction. For hourly cargo volume prediction, the ARIMA model, known for its suitability for short-term forecasting, is employed. The results are then visualized to provide clear insights. The findings reveal significant differences in logistics volumes across different seasons and time periods at each sorting center. Additionally, a dramatic increase in logistics volume is observed during certain shopping festivals. These variations significantly impact management decisions and resource allocation. Accurate cargo volume predictions, therefore, are essential for effectively allocating resources, optimizing personnel deployment, and improving the overall management efficiency of sorting centers to meet the dynamic demands of the logistics industry.