New Version

Allstate insurance portfolio — QPC on IBM Heron

Chance-constrained insurance knapsack in the Allstate × IBM problem class (arXiv:2605.01169): premium, primary risk, and multi-constraint blocks run as parallel contextures in one polycontextural IBM Heron submission.

Plain English

What this is: Insurance portfolio selection under chance constraints — maximize premium while respecting stochastic loss limits (ε=0.2).
What QPC computed: n=20 policies · M=5 chance constraints · 156Q Heron156 layout · premium / risk / constraint contextures · 4096 shots on ibm_fez.
QPC result: Polycontextural decoder premium score 578.10 · one coupled 156-qubit IBM job · d94nfuuvtlqs73fuhkj0.
What that means: An insurance book plan — which policies to write — earning premium 578.10 while staying inside the chance-constrained loss budget, from one polycontextural quantum job.

Lane B · Industry pilot. Headline premium 578.10 uses QPC’s declared problem decoder on quantum samples from one coupled 156-qubit job — an insurance book metric, not an ICC architecture witness.

Architecture coupling proofs on raw IBM counts are Lane A: Joint Structure Challenge · two-lane overview.

IBM Fez — Heron156 (July 2026)

Instance: n=20 · M=5 chance constraints · ε=0.2 · 4096 shots

QPC execution	Decoder score	Qubits	IBM job ID
QPC polycontextural (one run)	578.104	156	`d94nfuuvtlqs73fuhkj0`

Three contextures — premium · primary risk · multi-constraint — coupled with transjunction bridges in one 156-qubit Fez submission. Full Heron co-residence for insurance portfolio structure on gate hardware.

QPC polycontextural workflow

Premium, risk, and constraint blocks encoded as parallel contextures on the Heron register — submitted as one gate-native IBM job with auditable job IDs.

Series: Overview · Cerrado · Traffic · Allstate

Full overview → Reference paper IBM blog