Insurance
Home›Insurance›Industry & Deals›Allstate and IBM test quantum methods to model homeown…
Allstate and IBM test quantum methods to model homeowners portfolios
The researchers used IBM Quantum Heron with classical computing, testing problems up to 75 items and finding the quantum-classical method matched optimal results at that scale.
Allstate and IBM are exploring how quantum computing could help insurers build homeowners insurance portfolios that account for correlated catastrophe risk, including wildfires, hurricanes and tornadoes that can hit many properties at once. The companies said the work focuses on a portfolio version of the knapsack problem, choosing valuable policy combinations while keeping potential losses within an insurer’s risk tolerance.
The research, published on arXiv in May, uses IBM Quantum Heron to generate potential portfolio combinations, then applies classical computing to correct solutions that exceed the loss budget and guide additional rounds. Allstate said it currently runs around 100,000 simulations to estimate portfolio outcomes, but rare events can be difficult to model accurately.
In testing, the method was compared with four classical approximation approaches, with each given 30 minutes to solve a problem. For instances with up to 75 items, the quantum-classical and classical approaches reached the optimal answer, while at larger scales the quantum approach stayed competitive and slightly outperformed the best classical method under tight constraints.
Coverager reports the companies said the workflow is not yet ready for large-scale commercial use, but could help position them to use quantum computing as the hardware improves. As part of the deal, Allstate became the title sponsor of the ACC Women’s Softball Championship and ACC Men’s Lacrosse Championship.