Traffic Demand

Traffic demand prediction aims to forecast the movement of people or vehicles across a transportation network, crucial for optimizing resource allocation and improving transportation efficiency. Current research heavily utilizes graph neural networks (GNNs), often incorporating innovative graph construction methods to handle diverse data sources like ride-sharing and taxi services, and employing techniques like semi-decentralized inference to improve scalability and computational efficiency. These advancements are improving the accuracy and speed of demand forecasting, leading to better urban planning, more effective traffic management, and enhanced service provision for transportation providers and users.