Hybrid Quantum Risk Minimisation for Spatial Flood Allocation
A QRU–CVaR–QUBO–QAOA pipeline that links hydrological forecasting to correlation-aware spatial tail-risk allocation and evaluates both simulator and current-hardware constraints.
Forecast-to-decision pipeline
This project couples a single-qubit Quantum Re-Uploading Unit (QRU) forecaster with a downstream spatial allocation problem. Daily precipitation and river-level windows are converted into predictive scenarios and bounded flood-risk indicators. A composite interaction model combines hydrological connectivity, geographic proximity, and social exposure.
The allocation layer uses a tail-risk objective based on Conditional Value-at-Risk (CVaR), a QUBO representation, and a QAOA solver for small portfolio instances.
Evaluation principles
The project does not claim quantum advantage. Its evaluation instead asks:
- whether QRU and classical forecasters are compared under controlled capacity;
- how forecast errors propagate into tail-risk decisions;
- whether QAOA samples recover or approach known classical solutions;
- how hardware noise, compilation, Hamming-weight leakage, and finite shots affect feasibility.
Recognition and publication
The underlying flood-risk prototype originated in the DelphiQ project, winner of the World Bank Climate Risk Challenge track of the 2025 Global Industry Challenge. The extended technical paper was accepted in the Quantum End-to-End Hybrid Case Studies (QECS) track at IEEE Quantum Week 2026.
Conference website · Code and reproducibility artifact · Publication entry