Graph neural networks with flow conservation constraints for real-time origin-destination matrix completion
Abstract: Highlights•We present a novel framework to tackle real-time OD matrix incompleteness.•Our GNN-based model reconstructs complete OD matrices by integrating historical data.•We incorporate origin flow constraints to refine model outputs and enhance estimation accuracy.•The OD matrix inferred by our model improves real-time OD forecasting.
External IDs:dblp:journals/pr/ZhangYCW26
Loading