Research on Quantum Approximate Optimization Algorithm Enhanced Low-Altitude Traffic Monitoring for Urban Congestion Prediction and Dynamic Path Optimization

Urban traffic congestion has become increasingly severe, and traditional prediction and optimization methods struggle to meet the demands of real-time dynamic decision-making. This paper proposes a hybrid computational framework that integrates the Quantum Approximate Optimization Algorithm with a spatiotemporal graph attention network, achieving end-to-end joint solving of urban congestion prediction and dynamic path optimization. In the congestion prediction module, a multi-scale spatiotemporal graph attention encoder is designed, which models differentiated influence weights between nodes through a spatial graph attention mechanism, captures dependency relationships across three temporal scales—recent moments, daily periodicity, and weekly periodicity—using multi-scale causal convolution, and introduces adaptive graph structure learning to discover implicit spatial correlations.