Hybrid quantum annealing for large-scale exam scheduling: Validation in real-world educational scenarios

This study applied the hybrid quantum annealing to address complex exam scheduling challenges in large-scale educational scenarios through hybrid quantum annealing enhanced by graph-based preprocessing. By formulating the problem as a Quadratic Unconstrained Binary Optimization (QUBO) model and leveraging the D-Wave Advantage system, our method integrates quantum annealing with classical preprocessing to resolve constraints on exam room availability, student-course conflicts, and batch synchronization. The approach was applied in a university’s 2022 pandemic-era makeup exam scheduling for 1807 students and 215 courses (2749 exam instances) with zero conflicts. Experimental results show that hybrid quantum annealing consumes merely 86 ms of Quantum Processing Unit (QPU) execution time. In contrast, classical simulated annealing requires 13547 ms of Central Processing Unit (CPU) execution time for the same problem scale. This work bridges quantum computing and educational operations, offering a comparative analysis of hybrid algorithms in multi-constraint optimization domains.