F3 Certificate Passes Pairwise Rank-Order Extension But Not Full Closure¶
Claim/Theorem¶
The finite checker in docs/project_QEM-QEC/tmp/scripts/route_d_operation_extension.py does not refute the reconstructed f_3 certificate as a \Gamma_{\mathrm{sub},2} operation-level obstruction. It sharpens the obstruction side to the full rank-order extension problem.
For the f_3 certificate:
- the observed seven input-output columns are conservative;
- all observed pairwise Hamming inequalities hold;
- the minimum observed pairwise slack is
0; - coordinate-permutation extensions are infeasible;
- nested canonical rank-order extensions are infeasible;
- the pairwise meet/join rank-order closure ILP is feasible with
51closure points,2550binary variables, and20700monotonicity constraints; - full meet/join closure exceeds the configured cap, reaching
3001closure points before completion.
The six-qubit hidden-vertex-representable control behaves similarly:
- the observed support test passes;
- the pairwise meet/join rank-order closure ILP is feasible with
38closure points; - full meet/join closure also exceeds the
3000cap.
The toy meet/join submodularity certificate behaves differently:
- the direct observed test passes;
- the nested canonical subclass is feasible;
- the full closure ILP solves optimally on closure size
4.
Thus the current finite tests preserve the intended calibration:
\[
\text{toy positive case is solved, while } f_3 \text{ and the representable six-qubit control remain indistinguished.}
\]
The current f_3 certificate is therefore not killed by any implemented operation-extension necessary condition. The next exact finite task is the full rank-preserving order-preserving extension problem, not another pairwise Hamming check.
Dependencies¶
- [[conservative-hamming-operation-extension-is-rank-order-lattice-extension.md]]
- [[f3-fan-certificate-passes-necessary-multimorphism-tests-but-not-operation-lift.md]]
- [[fan-cone-certificates-are-not-hidden-vertex-invariants.md]]
- [[six-qubit-witness-is-hidden-vertex-graph-cut-representable.md]]
Conflicts/Gaps¶
- Pairwise meet/join closure is a sound necessary restriction, not a sufficient total-operation proof.
- The full closure cap is a computational barrier, not an infeasibility theorem.
- Because the six-qubit representable control also passes the implemented tests, the present finite checks are not yet hidden-vertex invariants.
Sources¶
10.1016/j.dam.2009.07.001- local computation:
docs/project_QEM-QEC/tmp/certificates/operation_extension_report.json - local script:
docs/project_QEM-QEC/tmp/scripts/route_d_operation_extension.py