Workshop Description
Quantum Amplitude Estimation offers a theoretical quadratic speed-up for Monte Carlo simulation. For catastrophe modelling, this translates to reducing a 10-hour overnight batch to approximately 3 hours with equivalent confidence intervals, or achieving tighter confidence bounds in the same run time. The algorithm encodes loss distributions as quantum states and uses quantum interference to estimate expected values with fewer samples than classical random sampling requires.
QAOA (Quantum Approximate Optimisation Algorithm) addresses the combinatorial problems that actuaries face in reinsurance treaty structuring and Solvency II capital allocation. Optimising excess-of-loss layer attachment points, quota share percentages, and aggregate deductibles across a multi-line book is NP-hard. Classical solvers use heuristics that find good solutions but rarely the optimal one. QAOA explores the solution space more efficiently for problem sizes that fit on current hardware, though production-scale reinsurance portfolios remain out of reach on NISQ devices. This workshop maps both QAE and QAOA against real actuarial workflows, gives honest hardware assessments, and builds a vendor evaluation framework specific to insurance.
What participants cover
- Quantum Amplitude Estimation for Monte Carlo: encoding catastrophe loss distributions, estimating expected annual loss, and reducing cat model batch run times
- QAOA for reinsurance portfolio optimisation: formulating treaty structures as QUBO problems, Solvency II SCR optimisation across business lines
- NISQ hardware assessment: IBM and Quantinuum demonstrations at 50-100 qubit scale achieve benchmark-specific performance comparisons, not general advantage
- D-Wave quantum annealing: practical experiments in financial portfolio optimisation and direct applicability to insurance capital allocation
- Vendor evaluation framework: IBM Quantum Network, AWS Braket, Azure Quantum, D-Wave access models and pricing for insurance pilot programmes
- Timeline and investment planning: NISQ-era pilots (2025-2028), early fault-tolerant applications (2029-2032), production deployment (2032+)