The field of transportation research is witnessing significant advancements in traffic flow optimization and logistics prediction. Current developments are focused on leveraging deep learning models, graph neural networks, and attention mechanisms to improve the accuracy and robustness of traffic prediction and logistics management. Notably, researchers are exploring the integration of spatial and temporal dependencies to capture complex traffic patterns and optimize traffic signal control. Additionally, there is a growing interest in developing models that can adapt to anomalous conditions, such as epidemics, and accurately predict delivery times in mixed logistics scenarios. These innovative approaches have the potential to revolutionize urban mobility planning and logistics management. Noteworthy papers include:

• TGDT, which proposes a scalable framework for dynamic traffic modeling and assessment at urban corridors.
• HetGL2R, which introduces an attributed heterogeneous graph learning approach for ranking node importance in road networks.
• TransPDT, which proposes a Transformer-based multi-task package delivery time prediction model for mixed logistics scenarios.
• DeepSTA, which presents a deep spatial-temporal attention model for logistics delivery timely rate prediction in anomaly conditions.