Development status: FREEZE_CANDIDATE
Locked protocol status: LOCKED_EVALUATED
Locked decision: LOCKED_SUCCESS
Evidence level: exact synthetic evaluation + feasible-subspace statevector simulation
The frozen 32-instance held-out protocol has been executed once on seeds 101, 211, 307, and 401. Development seeds 11 and 23 were excluded.
- model, service, chance, capacity, collaboration, ALNS, retained-core, QAOA, and sample-budget parameters frozen before execution;
- held-out seeds disjoint from development seeds;
- 32 locked instances reported;
- paired 95% bootstrap confidence intervals reported;
- environment and result digests recorded;
- null and failed adoption cases retained;
- zero final hard violations;
- raw CSV, summary, report, and figures committed.
| Metric | Result |
|---|---|
| Strict feasible-core rate | 0.938 |
| ALNS win rate | 0.781 |
| Mean ALNS gain | 17.019 |
| Ambiguity activation rate | 0.781 |
| Adoption-gate rate | 0.281 |
| QAOA / uniform optimum-probability ratio | 6.831 |
| QAOA / uniform time-to-optimum ratio | 0.354 |
| Final hard violations | 0 |
| Decision | LOCKED_SUCCESS |
- Locked report
- Locked summary
- Locked records CSV
- Environment manifest
- Locked dashboard
- Fancy quantum circuit
- Quantum-state distribution
- Quantum-state blooms
- Ambiguity map
- Trusted certificate
The locked result supports a statevector-level selective hybrid advantage candidate on the frozen synthetic benchmark. It does not establish real-QPU utility, wall-clock or asymptotic advantage, real-workplace safety, or superiority over every advanced classical method. Q-OPS is LOCKED_EVALUATED, not automatically the project-wide ACTIVE track.