Predictive analytics for traffic flow optimization in urban logistics: A transformer-based time series approach.

Area of Science:

• Urban logistics and traffic flow analysis.
• Environmental impact of urbanization.
• Time series forecasting methodologies.

Background:

• Growing urbanization presents challenges for managing complex urban logistics traffic.
• Existing forecasting methods struggle with large, noisy datasets.
• Accurate traffic flow prediction is vital for real-time management and planning.

Purpose of the Study:

• To propose a novel composite network model for enhanced urban logistics traffic flow prediction.
• To improve the accuracy and efficiency of traffic flow forecasting.
• To address limitations of current methods in handling complex, noisy data.

Main Methods:

• Development of a composite network model integrating Gated Recurrent Unit (GRU), Autoregressive Integrated Moving Average (ARIMA), and Temporal Fusion Transformer (TFT).
• Application and experimental analysis of the GRU-ARIMA-TFT model on diverse datasets.
• Evaluation of prediction accuracy and efficiency against existing methods.

Main Results:

• The GRU-ARIMA-TFT model demonstrated significant advantages in prediction accuracy.
• The model effectively captures and analyzes complex urban traffic patterns.
• Improved efficiency in processing large and complex traffic datasets was observed.

Conclusions:

• The proposed GRU-ARIMA-TFT model offers a robust solution for urban logistics traffic flow prediction.
• This research provides theoretical advancements and practical tools for optimizing urban traffic and logistics.
• The study offers new methodologies for future urban traffic management and logistics system planning.