Volume forecast of e-commerce sorting center based on adaptive hybrid ARIMA-LSTM and fuzzy logic

Abstract

With the rapid development of e-commerce, the logistics sorting center plays an important role. It not only realizes the rapid circulation of goods, but also is the key to improve customer satisfaction and supply chain efficiency. This research is devoted to constructing an accurate forecasting system, which can improve the efficiency of operation, promote the accurate planning of resources and save the cost by forecasting the daily and real-time cargo volume of the Sorting Center, and ensure the timely delivery of goods. This not only enhances the consumer experience, but also optimizes inventory management to prevent inventory shortages, while strengthening risk management and strategic decision-making to promote sustainable development and avoid waste of resources, and gain an edge in the fierce market competition.

In order to improve the accuracy of the prediction model, LSTM long-term and short-term memory network is introduced into the existing Arima model. Arima model is good at processing time series data with obvious trend and seasonality, while LSTM is good at capturing complex nonlinear relationships and long-term dependencies. By combining these two methods, we create an adaptive hybrid ARIMA-LSTM prediction model.

The results show that compared with Arima model alone, the adaptive hybrid ARIMA-LSTM model has a significant improvement in the prediction effect, and the prediction accuracy has been greatly enhanced. By introducing LSTM, the model can better understand and deal with the non-linear characteristics and long-term dependencies in time series data, thus making the prediction more accurate, this is of great value in guiding the operational strategy and resource planning of the Sorting Center.

By comparing the regression fitting between D1 and D2 to D5, it is found that the adaptive hybrid ARIMA-LSTM model is more advantageous than the single Arima model in the prediction of cargo volume. Thus, the model can be used to predict the future 24 hours of any day at every moment of the volume, to provide more detailed data guidance. However, because of the complexity and uncertainty of such predictions, we also need to further modify them using fuzzy logic to address possible inaccuracies or ambiguities.