Berlin's BVG tests quantum tech to fix driver shortages and save millions

Berlin’s public transport operator BVG is facing a growing workforce crisis, with over 4,300 employees set to retire by 2026. To tackle scheduling inefficiencies, a team from Beerantum has developed a quantum-enhanced solution that could save millions while addressing driver preferences often ignored by traditional methods. The project focused on BVG’s complex scheduling challenges, where classical approaches struggle to balance efficiency and individual driver needs. By integrating Kipu Quantum’s Bias-Field DCQO algorithm, the team optimised schedules for 150 drivers across multiple bus lines. A key innovation was the use of DBSCAN clustering, which cut API calls by 80%, streamlining the process.

An Uncertainty Adapter combined an Isolation Forest anomaly detector with a Gaussian Process demand predictor to handle real-world variability. This hybrid classical-quantum-classical pipeline moved from Technology Readiness Level (TRL) 4 to a production-ready pilot at TRL 6 in just 24 months, aligning with Kipu’s hardware development timeline. The solution’s architecture isn’t limited to transit. It can also be applied to hospital shift planning and last-mile logistics, demonstrating broad potential. Even a modest 2% improvement in scheduling efficiency at BVG’s scale could yield annual savings of around €18 million.

The pilot offers a practical way forward for BVG as it grapples with staff shortages and rising voluntary attrition. By incorporating driver preferences and reducing operational costs, the system could ease pressure on the workforce. The next steps involve scaling the solution across BVG’s network while exploring applications in other sectors.