Research on Freight Volume Forecasting and Personnel Scheduling at Logistics Transfer Stations Based on the ARIMA Algorithm

This paper focuses on the prediction of daily freight volumes at logistics transfer stations and the scheduling of personnel based on the ARIMA algorithm. The study begins with data preprocessing, including outlier handling and linear imputation for missing values. An ARIMA model is then established to forecast the daily freight volumes for the next thirty days. Leveraging historical data, the average proportion of daily freight volumes for each sorting centre is computed to derive future daily volume predictions. Finally, the proportion of hourly freight volumes within daily freight volumes is calculated to forecast hourly freight volumes for the upcoming thirty days. In addition, adjustments in certain transportation routes are considered during the prediction process to account for changes in inbound and outbound shipments to each sorting centre. Subsequently, the new forecast data is iterated through each day, and the transfer volumes are computed to obtain the predicted results. The final phase involves the allocation of 200 full-time employees based on the predicted daily freight volume proportions obtained in the second phase. An equation is established based on the relationship between the proportion of full-time employees resting on a given day out of the seven days in a week and the proportion of freight volume on that day out of the total freight volume. This equation is solved to determine the number of full-time employees resting on each of the seven days in a week.