Magnitude conjecture

MagnitudeConjecture.Combinatorics.DirectedDeletionRealization

Directed deletion from the direct count and graded factor positivity #

This file combines the live manuscript's vertex--arrow--mesh deletion count with the graded translation-slice proof that the factor Euler excess is nonnegative. The remaining representation-theoretic input is visible: new meshes must inject into gained arrow occurrences.

theorem MagnitudeConjecture.DirectedDeletion.translationQuiverSurplus_difference_nonnegative_of_translationRecurrence {H : Type v} [Fintype H] (Phi : Matrix H H ℤ) (vA nA aA vB nB aB q a0 aH z newArrows newMeshes : ℤ) {L : ℕ} (vertices arrows tauInjectives tauProjectives : ℕ → ℤ) (projectives : ℤ) (vertexCount : vA = vB + q) (simpleCount : nA = nB + 1) (ambientArrowCount : aA = a0 + aH + z) (quotientArrowCount : aB = a0 + newArrows) (factorEulerCount : intrinsicEulerExcess Phi = 2 * q - aH - projectives - 1) (newMeshCount : newMeshes = z - (projectives - 1)) (source_singleton : vertices 0 = 1) (sink_singleton : vertices L = 1) (firstSlice : arrows 0 = vertices 0 + vertices 1 - 1) (arrowStep : ∀ (j : ℕ), j + 1 < L → arrows (j + 1) = arrows j - tauInjectives j + tauProjectives (j + 2)) (vertexStep : ∀ (j : ℕ), j + 1 < L → vertices (j + 2) = vertices j - tauInjectives j + tauProjectives (j + 2)) (meshEulerTotal : ARCount.matrixTotal Phi = ((GradedTreeExcess.vertexTotal fun (j : Fin (L + 1)) => vertices ↑j) - GradedTreeExcess.arrowTotal fun (j : Fin L) => arrows ↑j) + ((GradedTreeExcess.vertexTotal fun (j : Fin (L + 1)) => vertices ↑j) - projectives)) (realization_length_bound : projectives - 1 ≤ ↑L) (newMeshes_le_newArrows : newMeshes ≤ newArrows) :

The directed-deletion inequality after supplying the literal direct counts and replacing abstract factor-excess nonnegativity by the graded translation recurrence and realization bound.

theorem MagnitudeConjecture.DirectedDeletion.translationQuiverSurplus_difference_eq_zero_iff_of_translationRecurrence {H : Type v} [Fintype H] (Phi : Matrix H H ℤ) (vA nA aA vB nB aB q a0 aH z newArrows newMeshes : ℤ) {L : ℕ} (vertices arrows tauInjectives tauProjectives : ℕ → ℤ) (projectives : ℤ) (vertexCount : vA = vB + q) (simpleCount : nA = nB + 1) (ambientArrowCount : aA = a0 + aH + z) (quotientArrowCount : aB = a0 + newArrows) (factorEulerCount : intrinsicEulerExcess Phi = 2 * q - aH - projectives - 1) (newMeshCount : newMeshes = z - (projectives - 1)) (source_singleton : vertices 0 = 1) (sink_singleton : vertices L = 1) (firstSlice : arrows 0 = vertices 0 + vertices 1 - 1) (arrowStep : ∀ (j : ℕ), j + 1 < L → arrows (j + 1) = arrows j - tauInjectives j + tauProjectives (j + 2)) (vertexStep : ∀ (j : ℕ), j + 1 < L → vertices (j + 2) = vertices j - tauInjectives j + tauProjectives (j + 2)) (meshEulerTotal : ARCount.matrixTotal Phi = ((GradedTreeExcess.vertexTotal fun (j : Fin (L + 1)) => vertices ↑j) - GradedTreeExcess.arrowTotal fun (j : Fin L) => arrows ↑j) + ((GradedTreeExcess.vertexTotal fun (j : Fin (L + 1)) => vertices ↑j) - projectives)) (realization_length_bound : projectives - 1 ≤ ↑L) (newMeshes_le_newArrows : newMeshes ≤ newArrows) :
translationQuiverSurplus vA nA aA - translationQuiverSurplus vB nB aB = 0 ↔ ↑L = projectives - 1 ∧ newArrows = newMeshes

The sharp directed-deletion equality criterion under the same direct count and graded translation hypotheses. Equality separates into sharp realization length and bijectivity of the new-mesh/new-arrow injection.

theorem MagnitudeConjecture.DirectedDeletion.finrank_eq_one_of_translationQuiverSurplus_difference_eq_zero_of_translationRecurrence_and_posetSpace {H : Type v} [Fintype H] (Phi : Matrix H H ℤ) (vA nA aA vB nB aB q a0 aH z newArrows newMeshes : ℤ) {L : ℕ} (vertices arrows tauInjectives tauProjectives : ℕ → ℤ) (projectives : ℤ) {k T : Type w} [Field k] [PartialOrder T] [Fintype T] (projectives_eq : projectives = ↑(Fintype.card T) + 1) (grading : PosetSpace.PositiveGrading (PosetSpace.Obj k T) (PosetSpace.IsSchur k T) L) (vertexCount : vA = vB + q) (simpleCount : nA = nB + 1) (ambientArrowCount : aA = a0 + aH + z) (quotientArrowCount : aB = a0 + newArrows) (factorEulerCount : intrinsicEulerExcess Phi = 2 * q - aH - projectives - 1) (newMeshCount : newMeshes = z - (projectives - 1)) (source_singleton : vertices 0 = 1) (sink_singleton : vertices L = 1) (firstSlice : arrows 0 = vertices 0 + vertices 1 - 1) (arrowStep : ∀ (j : ℕ), j + 1 < L → arrows (j + 1) = arrows j - tauInjectives j + tauProjectives (j + 2)) (vertexStep : ∀ (j : ℕ), j + 1 < L → vertices (j + 2) = vertices j - tauInjectives j + tauProjectives (j + 2)) (meshEulerTotal : ARCount.matrixTotal Phi = ((GradedTreeExcess.vertexTotal fun (j : Fin (L + 1)) => vertices ↑j) - GradedTreeExcess.arrowTotal fun (j : Fin L) => arrows ↑j) + ((GradedTreeExcess.vertexTotal fun (j : Fin (L + 1)) => vertices ↑j) - projectives)) (realization_length_bound : projectives - 1 ≤ ↑L) (newMeshes_le_newArrows : newMeshes ≤ newArrows) (hzero : translationQuiverSurplus vA nA aA - translationQuiverSurplus vB nB aB = 0) (X : PosetSpace.Obj k T) (hX : PosetSpace.IsSchur k T X) :
Module.finrank k X.carrier = 1

Equality in the complete direct deletion count makes the poset-space realization sharp, so every Schur realization object is one-dimensional.