Skip to content

Kernel-check the Section 9 matching-restriction product bound - #34

Open
SamPetkov wants to merge 4 commits into
mainfrom
agent/625-lean-matching-restriction-product
Open

Kernel-check the Section 9 matching-restriction product bound#34
SamPetkov wants to merge 4 commits into
mainfrom
agent/625-lean-matching-restriction-product

Conversation

@SamPetkov

@SamPetkov SamPetkov commented Jul 24, 2026

Copy link
Copy Markdown
Owner

Summary

This PR translates the finite combinatorial core of PR #30 into Lean.

For a finite even bipartite edge set and an exposed matching M, deleting M is injective. Consequently the weighted even-family sum is bounded directly by the full subset product on the cells outside M, without a cycle decomposition.

Files

  • 625/formalization/Erdos625/Section9MatchingRestrictionProduct.lean
  • 625/formalization/MATCHING_RESTRICTION_PRODUCT_AUDIT.md
  • .github/workflows/erdos625-matching-restriction-product.yml

The module proves:

  1. sdiff_matching_injective_on_bipartiteEvenEdgeSets — restriction F ↦ F \ M is injective on the finite even-edge family;

  2. weighted_evenSubgraph_ennreal_matching_product

    sum_F edgeWeightOutsideENN q M F
      <= product_{e notin M} (1 + q_e);
    
  3. residualCappedEvenFixedFSum_le_lambdaProduct_mul_matchingProduct — the existing fixed-F aggregation is bounded by the local-lambda product times the direct outside-matching product.

Why

The current formalization already contains the incidence-matrix parity encoding, matching-restriction uniqueness for supported even matrices, and the fixed-F threshold aggregation. This module composes those accepted finite ingredients into the simpler endpoint proposed in PR #30.

For this particular finite algebraic bound it removes the need for a simple-cycle decomposition, residual-walk enumeration, or marked matching-cycle encoding. Those independently useful modules remain unchanged.

Validation gates

The module has direct #print axioms checks. A focused workflow:

  • rejects sorry, admit, project-defined axioms/constants, and unsafe declarations;
  • prepares the pinned Lean/mathlib project;
  • compiles the module with -DwarningAsError=true.

The full repository Lean workflow also builds the formalization library with --wfail.

Scope boundary

This PR does not yet:

  • import the module into the root aggregate or generated self-contained checkpoint;
  • identify the fixed-F sum with the tagged actual attachment expectation;
  • prove the analytic O(U^2 + U^4/m_0) envelope;
  • prove Lemma 9.1, Proposition 9.2, or Erdos625Statement.

The audit file records the exact remaining bridges.

Copy link
Copy Markdown
Owner Author

Validation update: the corrected focused workflow, the full Lean 4.31 --wfail build, generated-source freshness check, warning-fatal self-contained compilation, and arXiv artifact synchronization all pass on head 35fa4f5aa85a12ae9ae23b7351031cd96dc6305f.

The earlier focused-workflow failure was CI plumbing only: it attempted direct compilation before building local dependencies. The theorem source did not change in that repair. PR #35 is now stacked on this green head and integrates the finite product theorem with the literal attachment numerator and analytic envelope.

@SamPetkov
SamPetkov marked this pull request as ready for review July 25, 2026 05:03
SamPetkov added a commit that referenced this pull request Jul 25, 2026
Integrate PRs #34, #35, and #37 into the head branch of audit PR #39. This is a mechanical branch-unification commit; it does not assert Proposition 9.2 or the final theorem.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant