Documentation

Anderson4D.PermSum.SingleScaleFiberKernelAlignment

Alignment of the located kernel ledger with the fiber CPS kernel #

The kernel schedule keeps original positions, while the finite-Fubini/CPS schedule keeps only analytic (N,X) blocks and consumes a finite tuple. This file records the missing pointwise alignment. The central certificate LocatedRunConsecutive says exactly that a located run starts at a given position and then visits consecutive positions away from the anchor.

Rough marking is harmless here: .pair and .roughPair retain the same two positions and both kernel definitions use the same lambda * strongLambda monomial.

Assignment and consecutiveness certificates #

Read a fixed position assignment in the exact tuple order consumed by a located ledger.

Equations
Instances For

    The analogous assignment for two outward runs. It is defined in the same recursion pattern as nxAnchoredScheduleKernel, so no dependent cast is hidden at the run boundary.

    Equations
    Instances For

      The anchored assignment reads the two located ledgers in order. The explicit arity casts below isolate the otherwise fragile dependent normalization at the left ++ right boundary.

      inductive Anderson4D.XYCluster.LocatedRunConsecutive {t : PlaneTree} {m : ℕ} {Nm : t.HeppMarking} {mu : t.Multiplicities} (anchor : Fin m) :
      Fin m → List (LocatedNXParityBlock Nm mu) → Prop

      A located run is consecutive away from anchor, beginning just outside previous. The endpoint of each block becomes the preceding position for the next block.

      Instances For
        def Anderson4D.XYCluster.OutwardPositionChain {m : ℕ} (anchor : Fin m) :
        Fin m → List (Fin m) → Prop

        Position-only version of the same consecutive-run condition.

        Equations
        Instances For

          Any located block partition of a consecutive position list inherits the kernel alignment certificate. This is the reusable raw pairPositionRun interface; it is insensitive to rough marking.

          theorem Anderson4D.XYCluster.locatedRunConsecutive_of_map_positionBlock_eq_pairPositionRun {t : PlaneTree} {m : ℕ} {Nm : t.HeppMarking} {mu : t.Multiplicities} (anchor previous : Fin m) (bs : List (LocatedNXParityBlock Nm mu)) (phase : Bool) (xs : List (Fin m)) (hblocks : List.map LocatedNXParityBlock.positionBlock bs = pairPositionRun phase xs) (hchain : OutwardPositionChain anchor previous xs) :
          LocatedRunConsecutive anchor previous bs

          Raw pairPositionRun wrapper. A rough-marked located list aligns whenever erasing its rough metadata recovers the raw pairing schedule.

          Rough marking and the canonical split retain the raw left position pairing schedule exactly.

          Rough marking and the canonical split retain the raw right position pairing schedule exactly.

          The sole position-only condition still needed to instantiate the canonical alignment. It contains no classes, skip flags, rough marking, or kernels.

          Equations
          • One or more equations did not get rendered due to their size.
          Instances For

            The canonical left/right position schedules really are consecutive outward runs.

            theorem Anderson4D.XYCluster.finAnchorNXLocatedCoarseRuns_consecutive_of_chainCondition {t : PlaneTree} {m : ℕ} (Nm : t.HeppMarking) (mu : t.Multiplicities) (leftPhase rightPhase : Bool) (anchor : Fin m) (cls : Fin m → ActiveNXClass Nm mu) (O : Finset (AdjacentIndex m)) (hchain : FinAnchorOutwardChainCondition anchor) :
            LocatedRunConsecutive anchor anchor (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).left ∧ LocatedRunConsecutive anchor anchor (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).right

            The position-only chain condition produces both actual coarse-run certificates.

            Exact kernel equalities from the certificate #

            theorem Anderson4D.XYCluster.locatedLedgerEliminationKernelProduct_eq_scheduleKernel {t : PlaneTree} {m : ℕ} (ht : t.isValid = true) (hroot : t.rootV ∈ t.BranchNodes) (Nm : t.HeppMarking) (mu : t.Multiplicities) (z : HeppLeaf t → Fin 4 → ℤ) (hz : IsSeparatedEmbedding Nm z) (R : ℝ) (anchor previous : Fin m) (σ : Fin m → HeppLabeledCopy mu) (bs : List (LocatedNXParityBlock Nm mu)) (halign : LocatedRunConsecutive anchor previous bs) :

            For a consecutive located run, the product of located elimination monomials is exactly the recursive CPS assignment monomial.

            theorem Anderson4D.XYCluster.locatedOutwardRunsEliminationKernelProduct_eq_anchoredScheduleKernel {t : PlaneTree} {m : ℕ} (ht : t.isValid = true) (hroot : t.rootV ∈ t.BranchNodes) (Nm : t.HeppMarking) (mu : t.Multiplicities) (z : HeppLeaf t → Fin 4 → ℤ) (hz : IsSeparatedEmbedding Nm z) (R : ℝ) (anchor : Fin m) (σ : Fin m → HeppLabeledCopy mu) (left right : List (LocatedNXParityBlock Nm mu)) (previous : Fin m) (hleft : LocatedRunConsecutive anchor previous left) (hright : LocatedRunConsecutive anchor anchor right) :
            locatedLedgerEliminationKernelProduct ht hroot Nm mu z hz R anchor σ left * locatedLedgerEliminationKernelProduct ht hroot Nm mu z hz R anchor σ right = nxAnchoredScheduleKernel ht hroot Nm mu z hz R (List.map LocatedNXParityBlock.analyticBlock left) (List.map LocatedNXParityBlock.analyticBlock right) (labeledCopyPoint z (σ anchor)) (labeledCopyPoint z (σ previous)) (locatedAnchoredScheduleAssignment σ left right)

            Two consecutive outward runs give exactly nxAnchoredScheduleKernel: the right run starts from the anchor point after the left tuple has been consumed, with no cross-run factor.

            theorem Anderson4D.XYCluster.edgeKernelProduct_le_finAnchorNXAnchoredScheduleKernel_of_consecutive {t : PlaneTree} {m : ℕ} (ht : t.isValid = true) (hroot : t.rootV ∈ t.BranchNodes) (Nm : t.HeppMarking) (mu : t.Multiplicities) (z : HeppLeaf t → Fin 4 → ℤ) (hz : IsSeparatedEmbedding Nm z) (R : ℝ) (O : Finset (AdjacentIndex m)) (leftPhase rightPhase : Bool) (anchor : Fin m) (cls : Fin m → ActiveNXClass Nm mu) (σ : Fin m → HeppLabeledCopy mu) (hleft : LocatedRunConsecutive anchor anchor (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).left) (hright : LocatedRunConsecutive anchor anchor (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).right) :
            ∏ edge : AdjacentIndex m, finAnchorNXArrangementEdgeKernel ht hroot Nm mu z hz R O leftPhase rightPhase anchor cls σ edge ≤ nxAnchoredScheduleKernel ht hroot Nm mu z hz R (List.map LocatedNXParityBlock.analyticBlock (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).left) (List.map LocatedNXParityBlock.analyticBlock (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).right) (labeledCopyPoint z (σ anchor)) (labeledCopyPoint z (σ anchor)) (locatedAnchoredScheduleAssignment σ (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).left (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).right)

            Pointwise bridge ready for the finite-Fubini monotonicity theorem. The only structural inputs are the two consecutive-run certificates; all analytic kernel comparison has already been discharged by KernelSchedule.

            Canonical finite-Fubini assignment #

            theorem Anderson4D.XYCluster.flatMap_positionEntries_finAnchorNXLocatedCoarseRuns_left {t : PlaneTree} {m : ℕ} (Nm : t.HeppMarking) (mu : t.Multiplicities) (leftPhase rightPhase : Bool) (anchor : Fin m) (cls : Fin m → ActiveNXClass Nm mu) (O : Finset (AdjacentIndex m)) :
            List.flatMap (fun (b : LocatedNXParityBlock Nm mu) => b.positionBlock.entries) (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).left = leftAnchorPositions anchor
            theorem Anderson4D.XYCluster.flatMap_positionEntries_finAnchorNXLocatedCoarseRuns_right {t : PlaneTree} {m : ℕ} (Nm : t.HeppMarking) (mu : t.Multiplicities) (leftPhase rightPhase : Bool) (anchor : Fin m) (cls : Fin m → ActiveNXClass Nm mu) (O : Finset (AdjacentIndex m)) :
            List.flatMap (fun (b : LocatedNXParityBlock Nm mu) => b.positionBlock.entries) (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).right = rightAnchorPositions anchor
            theorem Anderson4D.XYCluster.append_flatMap_positionEntries_finAnchorNXLocatedCoarseRuns {t : PlaneTree} {m : ℕ} (Nm : t.HeppMarking) (mu : t.Multiplicities) (leftPhase rightPhase : Bool) (anchor : Fin m) (cls : Fin m → ActiveNXClass Nm mu) (O : Finset (AdjacentIndex m)) :
            List.flatMap (fun (b : LocatedNXParityBlock Nm mu) => b.positionBlock.entries) (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).left ++ List.flatMap (fun (b : LocatedNXParityBlock Nm mu) => b.positionBlock.entries) (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).right = List.flatMap PositionBlock.entries (finAnchorPositionScheduleWithPhases leftPhase rightPhase anchor)
            theorem Anderson4D.XYCluster.map_paperFinAnchorAssignmentFromTail_outwardOrder {t : PlaneTree} {mu : t.Multiplicities} (leftPhase rightPhase : Bool) (anchor : Fin (totalMultiplicity mu)) (b : HeppLabeledCopy mu) (tail : Fin (paperTailLength mu) → HeppLabeledCopy mu) :
            List.map (paperFinAnchorAssignmentFromTail leftPhase rightPhase anchor b tail) (paperFinAnchorOutwardOrderWithPhases leftPhase rightPhase anchor) = List.ofFn tail
            theorem Anderson4D.XYCluster.locatedAnchoredScheduleAssignment_finAnchor_fromTail {t : PlaneTree} (Nm : t.HeppMarking) (mu : t.Multiplicities) (leftPhase rightPhase : Bool) (anchor : Fin (totalMultiplicity mu)) (cls : Fin (totalMultiplicity mu) → ActiveNXClass Nm mu) (O : Finset (AdjacentIndex (totalMultiplicity mu))) (b : HeppLabeledCopy mu) (tail : Fin (paperTailLength mu) → HeppLabeledCopy mu) :
            locatedAnchoredScheduleAssignment (paperFinAnchorAssignmentFromTail leftPhase rightPhase anchor b tail) (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).left (finAnchorNXLocatedCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).right = tail ∘ Fin.cast ⋯

            The assignment reconstructed by finite Fubini is consumed by the located left/right ledgers as exactly the supplied tail tuple.

            theorem Anderson4D.XYCluster.nxAnchoredScheduleKernel_congr_cast {t : PlaneTree} (ht : t.isValid = true) (hroot : t.rootV ∈ t.BranchNodes) (Nm : t.HeppMarking) (mu : t.Multiplicities) (z : HeppLeaf t → Fin 4 → ℤ) (hz : IsSeparatedEmbedding Nm z) (R : ℝ) (left right left' right' : List (NXParityBlock Nm mu)) (hleft : left = left') (hright : right = right') (anchorPoint u : Fin 4 → ℤ) (n : ℕ) (hArity : nxParityScheduleArity (left ++ right) = n) (hArity' : nxParityScheduleArity (left' ++ right') = n) (f : Fin n → HeppLabeledCopy mu) :
            nxAnchoredScheduleKernel ht hroot Nm mu z hz R left right anchorPoint u (f ∘ Fin.cast hArity) = nxAnchoredScheduleKernel ht hroot Nm mu z hz R left' right' anchorPoint u (f ∘ Fin.cast hArity')
            theorem Anderson4D.XYCluster.edgeKernelProduct_le_finAnchorNXAnchoredScheduleKernel_fromTail {t : PlaneTree} (ht : t.isValid = true) (hroot : t.rootV ∈ t.BranchNodes) (Nm : t.HeppMarking) (mu : t.Multiplicities) (z : HeppLeaf t → Fin 4 → ℤ) (hz : IsSeparatedEmbedding Nm z) (R : ℝ) (O : Finset (AdjacentIndex (totalMultiplicity mu))) (leftPhase rightPhase : Bool) (anchor : Fin (totalMultiplicity mu)) (cls : Fin (totalMultiplicity mu) → ActiveNXClass Nm mu) (b : HeppLabeledCopy mu) (tail : Fin (paperTailLength mu) → HeppLabeledCopy mu) :
            ∏ edge : AdjacentIndex (totalMultiplicity mu), finAnchorNXArrangementEdgeKernel ht hroot Nm mu z hz R O leftPhase rightPhase anchor cls (paperFinAnchorAssignmentFromTail leftPhase rightPhase anchor b tail) edge ≤ nxAnchoredScheduleKernel ht hroot Nm mu z hz R (finAnchorNXCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).left (finAnchorNXCoarseRunsWithPhases Nm mu leftPhase rightPhase anchor cls O).right (labeledCopyPoint z b) (labeledCopyPoint z b) (tail ∘ Fin.cast ⋯)

            Canonical pointwise bound for the fixed-fiber CPS theorem. Consecutiveness, anchor reconstruction, analytic erasure, and the two arity casts are all discharged internally.