Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Definitions for Operator Algebras and Conformal Field Theory III. Fusion of positive energy representations of LSU(N) using bounded operators

Definition
QAlgorithms_Papers_Independent_Intakenext100_P41

by Gohan · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

The objects the formalized statements of Operator Algebras and Conformal Field Theory III. Fusion of positive energy representations of LSU(N) using bounded operators (arXiv:math/9806031) are stated in. Each was checked against the paper by a blind read-back and a moderator.

  • AngularCircle: Angular circle, with coordinate θ modulo 2π (§4, PDF pp.10–11).
  • angularHaar: Normalized angular Haar measure, of total mass one (§14, PDF p.20).
  • CircleL2: Actual vector-valued L² classes, with the Euclidean norm on C^N (§14, p.20).
  • vectorMonomial: Continuous nonnegative Fourier monomial θ ↦ exp(inθ)v (§4, pp.10–11).
  • vectorMonomialL2: The same Fourier monomial as an L² class (§4, pp.10–11).
  • circleHardySpace: Hilbert-closed span of all z^n v, n ≥ 0, not a smooth-function space (§14, p.20).
  • angularArc: Nonempty proper open arcs are obtained from a < b < a+2π (§14, p.20). The real endpoints are retained, so arcs crossing θ=0 are also allowed.
  • VanishesOnArc: Vanishing on an arc means Haar-a.e. vanishing of an L² representative (§14, p.20).
  • linearRankOneMatrix: Rank-one operator Q(x)=φ(x)v (§18, PDF p.25). This allows an arbitrary complex-linear functional φ; the frozen Advance50.P38.rankOneMatrix instead denotes the Hermitian operator xx* and cannot represent this source's general rank-one operator.
  • resolventScalar: Resolvent scalar χ_A(t)=1+φ((tI−A)⁻¹v) (§22, PDF p.28). Theorems evaluating it require the resolvent matrix to be invertible.
  • BasicODEData: The actual normalized spectral data of the basic ODE (§18, PDF p.25). Both P and R=Q−P have simple nonresonant spectra and are in general position with Q=vφ; mu records the eigenvalues of P−Q, not those of R.
  • frobeniusPair: Recursive Frobenius coefficients and their partial sums (§22, PDF p.28). The pair is (ξ_n, ξ_0+⋯+ξ_n), constructed from the ODE recurrence. Nonresonance supplies every inverse at positive n in its applications.
  • projectedCoefficient: α_{i,n}=φ(ξ_{i,0}+⋯+ξ_{i,n}) from the constructed Frobenius solution (§22, p.28).
  • IntegerSignature: Unnormalized integer signatures, before common-shift SU(N) equivalence (§2, p.8). The frozen Dependency50.P26.Signature describes quantum-comb wire dimensions; it is unrelated to these integer highest-weight coordinates.
  • IsSignature: A signature is weakly decreasing (§2, PDF p.8).
  • IsPermissible: Level-permissibility, expressed without artificial zero-padding (§§9,34, pp.16,43).
  • AddsOneBox: Adding one actual box before renormalizing signatures (§2, p.8).
  • BoxPath: A path retains every intermediate permissible signature (§32, pp.40–42).
  • AdjacentPaths: Adjacent paths differ at exactly one interior signature (§32, PDF pp.41–42).
  • AffineWeylElement: An element of Λ₀⋊S_N, with translation (N+ℓ)m and Σm_i=0 (§34, PDF p.43).
  • affineAction: Coordinate action with the usual inverse-permutation convention (§34, p.43).
  • HasTrivialStabilizer: Trivial stabilizer in the actual affine action (§34, PDF p.43).
  • weylShift: The source's Weyl shift δ=(N−1,…,0), with zero-based Lean indices (§34, p.43).
  • InAffineRootLattice: Root-lattice membership uses zero-sum integral translations, not arbitrary shifts (§34, p.43).
  • gaugeParameterSet: The nonresonant parameter set for a holomorphic family B=zA (§19, PDF p.26). Zero eigenvalue differences are allowed; only positive integral differences are excluded.
  • gaugeDomain: Holomorphic gauge domain including z=0 (§19, PDF p.26).
  • IsNormalizedGauge: A normalized holomorphic invertible matrix gauge; its equation is imposed only for 0<|z|<R (§19, PDF p.26). Values outside the domain play no role.
  • companionMatrix: Companion matrix with superdiagonal ones and last row a (§20, PDF pp.26–27).
  • lastRowMatrix: The simultaneous rank-one form has only its last row nonzero (§20, pp.26–27).
  • companionPolynomial: Polynomial a(t)=t^d−Σa_i t^(i−1), with zero-based Fin indices (§20, p.27).
  • rowPolynomial: Polynomial b(t)=Σb_i t^(i−1) (§20, PDF p.27).
  • NoCommonRoot: No common complex root, not merely inequality of the two polynomials (§20, p.27).
  • HasSimpleSpectrum: Simple spectrum includes a full eigenbasis (§§18,20, PDF pp.25–27).
  • RankOneGeneralPosition: Both-sided general position for a simple-spectrum P and rank-one Q: Q annihilates no eigenline and its range has no zero eigenbasis coordinate (§20, p.26).
  • realAffineAction: Real affine action, on all R^N rather than a zero-sum slice (§34, PDF p.43).
  • IsRealRootTranslation: Real zero-sum root-lattice translations (§34, PDF p.43).
  • HasTrivialRealStabilizer: Trivial stabilizer in the real affine action (§34, PDF p.43).
  • translationAlcove: The symmetric closed translation domain (§34 lemma (a), PDF p.43).
  • dominantAlcove: The closed dominant affine alcove (§34 lemma (b), PDF p.43).
  • IsTranslationFundamentalDomain: Closed fundamental domain: orbit coverage and no overlap of interiors under nonidentity translations. Boundary duplicates are allowed (§34, PDF p.43).
  • IsAffineFundamentalDomain: Closed fundamental domain for the affine Weyl action (§34, PDF p.43).
  • IsCyclicFor: Cyclicity uses the norm-dense actual operator orbit MΩ (§12, PDF p.18).
  • closedSelfAdjointRealPart: K is the real closed span of self-adjoint vectors aΩ (§12, PDF p.18). This does not assume standardness or any modular identities.
  • imaginaryRealPart: The real subspace iK uses the original complex structure (§12, PDF p.18).
Definition code
import Mathlib

/-!
Statements from Antony Wassermann, arXiv:math/9806031v1.
All constructions below are local to this paper. Analytic representation-theoretic
nodes needing further substrate are explicitly inventoried in HARD.md.
-/

noncomputable section

open scoped BigOperators
open MeasureTheory Set

namespace QAlgorithms.Papers.Independent.Intakenext100.P41

/-- Angular circle, with coordinate θ modulo 2π (§4, PDF pp.10–11). -/
abbrev AngularCircle := AddCircle (2 * Real.pi)

instance angularPeriodPositive : Fact (0 < 2 * Real.pi) :=
  ⟨mul_pos (by norm_num) Real.pi_pos⟩

/-- Normalized angular Haar measure, of total mass one (§14, PDF p.20). -/
abbrev angularHaar : Measure AngularCircle := AddCircle.haarAddCircle

/-- Actual vector-valued L² classes, with the Euclidean norm on C^N (§14, p.20). -/
abbrev CircleL2 (N : ℕ) := Lp (EuclideanSpace ℂ (Fin N)) 2 angularHaar

/-- Continuous nonnegative Fourier monomial θ ↦ exp(inθ)v (§4, pp.10–11). -/
def vectorMonomial (N n : ℕ) (v : EuclideanSpace ℂ (Fin N)) :
    C(AngularCircle, EuclideanSpace ℂ (Fin N)) where
  toFun θ := fourier (T := 2 * Real.pi) (n : ℤ) θ • v
  continuous_toFun := (fourier (T := 2 * Real.pi) (n : ℤ)).continuous.smul continuous_const

/-- The same Fourier monomial as an L² class (§4, pp.10–11). -/
def vectorMonomialL2 (N n : ℕ) (v : EuclideanSpace ℂ (Fin N)) : CircleL2 N :=
  ContinuousMap.toLp 2 angularHaar ℂ (vectorMonomial N n v)

/-- Hilbert-closed span of all z^n v, n ≥ 0, not a smooth-function space (§14, p.20). -/
def circleHardySpace (N : ℕ) : Submodule ℂ (CircleL2 N) :=
  (Submodule.span ℂ {f | ∃ n v, f = vectorMonomialL2 N n v}).topologicalClosure

/-- Nonempty proper open arcs are obtained from a < b < a+2π (§14, p.20).
The real endpoints are retained, so arcs crossing θ=0 are also allowed. -/
def angularArc (a b : ℝ) : Set AngularCircle :=
  (fun θ : ℝ => (θ : AngularCircle)) '' Set.Ioo a b

/-- Vanishing on an arc means Haar-a.e. vanishing of an L² representative (§14, p.20). -/
def VanishesOnArc {N : ℕ} (f : CircleL2 N) (a b : ℝ) : Prop :=
  ∀ᵐ θ ∂angularHaar, θ ∈ angularArc a b → f θ = 0

example (N n : ℕ) (v : EuclideanSpace ℂ (Fin N)) :
    vectorMonomialL2 N n v ∈ circleHardySpace N := by
  exact (Submodule.le_topologicalClosure _) (Submodule.subset_span ⟨n, v, rfl⟩)

example (N : ℕ) : (0 : CircleL2 N) ∈ circleHardySpace N :=
  (circleHardySpace N).zero_mem

example (a : ℝ) : angularArc a a = ∅ := by
  simp [angularArc]

/-- Rank-one operator Q(x)=φ(x)v (§18, PDF p.25).
This allows an arbitrary complex-linear functional φ; the frozen
Advance50.P38.rankOneMatrix instead denotes the Hermitian operator xx* and
cannot represent this source's general rank-one operator. -/
def linearRankOneMatrix {d : ℕ} (v : Fin d → ℂ)
    (phi : (Fin d → ℂ) →ₗ[ℂ] ℂ) : Matrix (Fin d) (Fin d) ℂ :=
  fun i j => v i * phi (Pi.single j 1)

/-- Resolvent scalar χ_A(t)=1+φ((tI−A)⁻¹v) (§22, PDF p.28).
Theorems evaluating it require the resolvent matrix to be invertible. -/
def resolventScalar {d : ℕ} (A : Matrix (Fin d) (Fin d) ℂ)
    (v : Fin d → ℂ) (phi : (Fin d → ℂ) →ₗ[ℂ] ℂ) (t : ℂ) : ℂ :=
  1 + phi (((t • (1 : Matrix (Fin d) (Fin d) ℂ)) - A)⁻¹.mulVec v)

/-- The actual normalized spectral data of the basic ODE (§18, PDF p.25).
Both P and R=Q−P have simple nonresonant spectra and are in general position
with Q=vφ; mu records the eigenvalues of P−Q, not those of R. -/
structure BasicODEData (d : ℕ) where
  positiveDimension : 0 < d
  P : Matrix (Fin d) (Fin d) ℂ
  v : Fin d → ℂ
  phi : (Fin d → ℂ) →ₗ[ℂ] ℂ
  nonzeroTrace : phi v ≠ 0
  lambda : Fin d → ℂ
  mu : Fin d → ℂ
  xi : Module.Basis (Fin d) ℂ (Fin d → ℂ)
  zeta : Module.Basis (Fin d) ℂ (Fin d → ℂ)
  eigenP : ∀ i, P.mulVec (xi i) = lambda i • xi i
  eigenPminusQ : ∀ i, (P - linearRankOneMatrix v phi).mulVec (zeta i) = mu i • zeta i
  nonresonantLambda : ∀ i j, i ≠ j → ∀ n : ℤ, lambda i - lambda j ≠ (n : ℂ)
  nonresonantMu : ∀ i j, i ≠ j → ∀ n : ℤ, mu i - mu j ≠ (n : ℂ)
  normalizedXi : ∀ i, phi (xi i) = 1
  normalizedZeta : ∀ i, phi (zeta i) = 1
  generalPositionP : ∀ i, xi.repr v i ≠ 0
  generalPositionR : ∀ i, zeta.repr v i ≠ 0

example {d : ℕ} (phi : (Fin d → ℂ) →ₗ[ℂ] ℂ) :
    linearRankOneMatrix (0 : Fin d → ℂ) phi = 0 := by
  ext i j
  simp [linearRankOneMatrix]

/-- Recursive Frobenius coefficients and their partial sums (§22, PDF p.28).
The pair is (ξ_n, ξ_0+⋯+ξ_n), constructed from the ODE recurrence.
Nonresonance supplies every inverse at positive n in its applications. -/
def frobeniusPair {d : ℕ} (P Q : Matrix (Fin d) (Fin d) ℂ)
    (lambda : ℂ) (w : Fin d → ℂ) : ℕ → (Fin d → ℂ) × (Fin d → ℂ)
  | 0 => (w, w)
  | n + 1 =>
      let previous := frobeniusPair P Q lambda w n
      let next := (((lambda + ((n + 1 : ℕ) : ℂ)) •
        (1 : Matrix (Fin d) (Fin d) ℂ)) - P)⁻¹.mulVec (Q.mulVec previous.2)
      (next, previous.2 + next)

/-- α_{i,n}=φ(ξ_{i,0}+⋯+ξ_{i,n}) from the constructed Frobenius solution (§22, p.28). -/
def projectedCoefficient {d : ℕ} (A : BasicODEData d) (i : Fin d) (n : ℕ) : ℂ :=
  A.phi (frobeniusPair A.P (linearRankOneMatrix A.v A.phi) (A.lambda i) (A.xi i) n).2

example {d : ℕ} (P Q : Matrix (Fin d) (Fin d) ℂ) (lambda : ℂ) (w : Fin d → ℂ) :
    frobeniusPair P Q lambda w 0 = (w, w) := rfl

/-- Unnormalized integer signatures, before common-shift SU(N) equivalence (§2, p.8).
The frozen Dependency50.P26.Signature describes quantum-comb wire dimensions;
it is unrelated to these integer highest-weight coordinates. -/
abbrev IntegerSignature (N : ℕ) := Fin N → ℤ

/-- A signature is weakly decreasing (§2, PDF p.8). -/
def IsSignature {N : ℕ} (f : IntegerSignature N) : Prop :=
  ∀ i j, i ≤ j → f j ≤ f i

/-- Level-permissibility, expressed without artificial zero-padding (§§9,34, pp.16,43). -/
def IsPermissible {N : ℕ} (ell : ℕ) (f : IntegerSignature N) : Prop :=
  IsSignature f ∧ ∀ i j, f i - f j ≤ (ell : ℤ)

/-- Adding one actual box before renormalizing signatures (§2, p.8). -/
def AddsOneBox {N : ℕ} (f g : IntegerSignature N) : Prop :=
  ∃ row : Fin N, ∀ i, g i = f i + if i = row then 1 else 0

/-- A path retains every intermediate permissible signature (§32, pp.40–42). -/
structure BoxPath {N : ℕ} (ell k : ℕ) (f g : IntegerSignature N) where
  vertex : Fin (k + 1) → IntegerSignature N
  start : vertex 0 = f
  finish : vertex (Fin.last k) = g
  permissible : ∀ j, IsPermissible ell (vertex j)
  step : ∀ j : Fin k, AddsOneBox (vertex j.castSucc) (vertex j.succ)

/-- Adjacent paths differ at exactly one interior signature (§32, PDF pp.41–42). -/
def AdjacentPaths {N ell k : ℕ} {f g : IntegerSignature N}
    (P Q : BoxPath ell k f g) : Prop :=
  ∃ j : Fin (k + 1), 0 < j.val ∧ j.val < k ∧
    P.vertex j ≠ Q.vertex j ∧ ∀ i, i ≠ j → P.vertex i = Q.vertex i

example (N ell : ℕ) : IsPermissible ell (0 : IntegerSignature N) := by
  constructor
  · intro i j hij
    simp
  · intro i j
    simp

/-- An element of Λ₀⋊S_N, with translation (N+ℓ)m and Σm_i=0 (§34, PDF p.43). -/
structure AffineWeylElement (N : ℕ) where
  translation : Fin N → ℤ
  sumZero : ∑ i, translation i = 0
  permutation : Equiv.Perm (Fin N)

/-- Coordinate action with the usual inverse-permutation convention (§34, p.43). -/
def affineAction {N : ℕ} (ell : ℕ) (w : AffineWeylElement N)
    (x : Fin N → ℤ) : Fin N → ℤ :=
  fun i => x (w.permutation.symm i) + ((N + ell : ℕ) : ℤ) * w.translation i

/-- Trivial stabilizer in the actual affine action (§34, PDF p.43). -/
def HasTrivialStabilizer {N : ℕ} (ell : ℕ) (x : Fin N → ℤ) : Prop :=
  ∀ w : AffineWeylElement N, affineAction ell w x = x →
    w.translation = 0 ∧ w.permutation = Equiv.refl (Fin N)

/-- The source's Weyl shift δ=(N−1,…,0), with zero-based Lean indices (§34, p.43). -/
def weylShift (N : ℕ) : Fin N → ℤ := fun i => (N : ℤ) - 1 - (i.val : ℤ)

/-- Root-lattice membership uses zero-sum integral translations, not arbitrary shifts (§34, p.43). -/
def InAffineRootLattice {N : ℕ} (ell : ℕ) (x : Fin N → ℤ) : Prop :=
  ∃ m : Fin N → ℤ, (∑ i, m i) = 0 ∧
    ∀ i, x i = ((N + ell : ℕ) : ℤ) * m i

example : weylShift 2 = ![1, 0] := by
  ext i
  fin_cases i <;> norm_num [weylShift]

example (N ell : ℕ) : InAffineRootLattice (N := N) ell 0 := by
  refine ⟨0, by simp, ?_⟩
  intro i
  simp

/-- The nonresonant parameter set for a holomorphic family B=zA (§19, PDF p.26).
Zero eigenvalue differences are allowed; only positive integral differences are excluded. -/
def gaugeParameterSet {d m : ℕ}
    (B : (Fin m → ℂ) × ℂ → Matrix (Fin d) (Fin d) ℂ) : Set (Fin m → ℂ) :=
  {t | ∀ lambda mu : ℂ,
    (B (t, 0)).charpoly.eval lambda = 0 → (B (t, 0)).charpoly.eval mu = 0 →
    ∀ n : ℕ, 0 < n → lambda - mu ≠ (n : ℂ)}

/-- Holomorphic gauge domain including z=0 (§19, PDF p.26). -/
def gaugeDomain {d m : ℕ} (R : ℝ)
    (B : (Fin m → ℂ) × ℂ → Matrix (Fin d) (Fin d) ℂ) :
    Set ((Fin m → ℂ) × ℂ) :=
  gaugeParameterSet B ×ˢ Metric.ball 0 R

/-- A normalized holomorphic invertible matrix gauge; its equation is imposed
only for 0<|z|<R (§19, PDF p.26). Values outside the domain play no role. -/
def IsNormalizedGauge {d m : ℕ} (R : ℝ)
    (B g : (Fin m → ℂ) × ℂ → Matrix (Fin d) (Fin d) ℂ) : Prop :=
  DifferentiableOn ℂ g (gaugeDomain R B) ∧
  (∀ p ∈ gaugeDomain R B, (g p).det ≠ 0) ∧
  (∀ t ∈ gaugeParameterSet B, g (t, 0) = 1) ∧
  ∀ t ∈ gaugeParameterSet B, ∀ z : ℂ, ‖z‖ < R → z ≠ 0 →
    (g (t, z))⁻¹ * (z⁻¹ • B (t, z)) * g (t, z) -
      (g (t, z))⁻¹ * deriv (fun w : ℂ => g (t, w)) z = z⁻¹ • B (t, 0)

example {d m : ℕ} (B : (Fin m → ℂ) × ℂ → Matrix (Fin d) (Fin d) ℂ) :
    gaugeDomain 0 B = ∅ := by
  simp [gaugeDomain, Metric.ball_zero]

/-- Companion matrix with superdiagonal ones and last row a (§20, PDF pp.26–27). -/
def companionMatrix {d : ℕ} (a : Fin d → ℂ) : Matrix (Fin d) (Fin d) ℂ :=
  fun i j => if i.val + 1 = d then a j else if j.val = i.val + 1 then 1 else 0

/-- The simultaneous rank-one form has only its last row nonzero (§20, pp.26–27). -/
def lastRowMatrix {d : ℕ} (b : Fin d → ℂ) : Matrix (Fin d) (Fin d) ℂ :=
  fun i j => if i.val + 1 = d then b j else 0

/-- Polynomial a(t)=t^d−Σa_i t^(i−1), with zero-based Fin indices (§20, p.27). -/
def companionPolynomial {d : ℕ} (a : Fin d → ℂ) : Polynomial ℂ :=
  Polynomial.X ^ d - ∑ i, Polynomial.C (a i) * Polynomial.X ^ i.val

/-- Polynomial b(t)=Σb_i t^(i−1) (§20, PDF p.27). -/
def rowPolynomial {d : ℕ} (b : Fin d → ℂ) : Polynomial ℂ :=
  ∑ i, Polynomial.C (b i) * Polynomial.X ^ i.val

/-- No common complex root, not merely inequality of the two polynomials (§20, p.27). -/
def NoCommonRoot (a b : Polynomial ℂ) : Prop :=
  ∀ z : ℂ, ¬(a.eval z = 0 ∧ b.eval z = 0)

/-- Simple spectrum includes a full eigenbasis (§§18,20, PDF pp.25–27). -/
def HasSimpleSpectrum {d : ℕ} (P : Matrix (Fin d) (Fin d) ℂ) : Prop :=
  ∃ (e : Module.Basis (Fin d) ℂ (Fin d → ℂ)) (lambda : Fin d → ℂ),
    Function.Injective lambda ∧ ∀ i, P.mulVec (e i) = lambda i • e i

/-- Both-sided general position for a simple-spectrum P and rank-one Q:
Q annihilates no eigenline and its range has no zero eigenbasis coordinate (§20, p.26). -/
def RankOneGeneralPosition {d : ℕ} (P Q : Matrix (Fin d) (Fin d) ℂ) : Prop :=
  ∃ (e : Module.Basis (Fin d) ℂ (Fin d → ℂ)) (lambda : Fin d → ℂ),
    Function.Injective lambda ∧ (∀ i, P.mulVec (e i) = lambda i • e i) ∧
    (∀ i, Q.mulVec (e i) ≠ 0) ∧ ∀ i, ∃ x, e.repr (Q.mulVec x) i ≠ 0

example (a : Fin 1 → ℂ) : companionMatrix a 0 0 = a 0 := by
  simp [companionMatrix]

example (b : Fin 1 → ℂ) : lastRowMatrix b 0 0 = b 0 := by
  simp [lastRowMatrix]

/-- Real affine action, on all R^N rather than a zero-sum slice (§34, PDF p.43). -/
def realAffineAction {N : ℕ} (ell : ℕ) (w : AffineWeylElement N)
    (x : Fin N → ℝ) : Fin N → ℝ :=
  fun i => x (w.permutation.symm i) + ((N + ell : ℕ) : ℝ) * (w.translation i : ℝ)

/-- Real zero-sum root-lattice translations (§34, PDF p.43). -/
def IsRealRootTranslation {N : ℕ} (ell : ℕ) (x : Fin N → ℝ) : Prop :=
  ∃ m : Fin N → ℤ, (∑ i, m i) = 0 ∧
    ∀ i, x i = ((N + ell : ℕ) : ℝ) * (m i : ℝ)

/-- Trivial stabilizer in the real affine action (§34, PDF p.43). -/
def HasTrivialRealStabilizer {N : ℕ} (ell : ℕ) (x : Fin N → ℝ) : Prop :=
  ∀ w : AffineWeylElement N, realAffineAction ell w x = x →
    w.translation = 0 ∧ w.permutation = Equiv.refl (Fin N)

/-- The symmetric closed translation domain (§34 lemma (a), PDF p.43). -/
def translationAlcove (N ell : ℕ) : Set (Fin N → ℝ) :=
  {x | ∀ i j, |x i - x j| ≤ ((N + ell : ℕ) : ℝ)}

/-- The closed dominant affine alcove (§34 lemma (b), PDF p.43). -/
def dominantAlcove (N ell : ℕ) : Set (Fin N → ℝ) :=
  {x | (∀ i j, i ≤ j → x j ≤ x i) ∧
    ∀ i j, x i - x j ≤ ((N + ell : ℕ) : ℝ)}

/-- Closed fundamental domain: orbit coverage and no overlap of interiors under
nonidentity translations. Boundary duplicates are allowed (§34, PDF p.43). -/
def IsTranslationFundamentalDomain {N : ℕ} (ell : ℕ) (D : Set (Fin N → ℝ)) : Prop :=
  IsClosed D ∧
  (∀ x : Fin N → ℝ, ∃ a : Fin N → ℝ, IsRealRootTranslation ell a ∧ x + a ∈ D) ∧
  ∀ x a : Fin N → ℝ, IsRealRootTranslation ell a →
    x ∈ interior D → x + a ∈ interior D → a = 0

/-- Closed fundamental domain for the affine Weyl action (§34, PDF p.43). -/
def IsAffineFundamentalDomain {N : ℕ} (ell : ℕ) (D : Set (Fin N → ℝ)) : Prop :=
  IsClosed D ∧
  (∀ x : Fin N → ℝ, ∃ w : AffineWeylElement N, realAffineAction ell w x ∈ D) ∧
  ∀ (w : AffineWeylElement N) (x : Fin N → ℝ),
    x ∈ interior D → realAffineAction ell w x ∈ interior D →
      w.translation = 0 ∧ w.permutation = Equiv.refl (Fin N)

example (N ell : ℕ) : (0 : Fin N → ℝ) ∈ dominantAlcove N ell := by
  constructor
  · intro i j hij
    simp
  · intro i j
    simpa using (Nat.cast_nonneg (N + ell) : (0 : ℝ) ≤ ((N + ell : ℕ) : ℝ))

section StandardRealPart

variable {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℂ H] [CompleteSpace H]

local instance : InnerProductSpace ℝ H := InnerProductSpace.complexToReal

/-- Cyclicity uses the norm-dense actual operator orbit MΩ (§12, PDF p.18). -/
def IsCyclicFor (M : VonNeumannAlgebra H) (omega : H) : Prop :=
  Dense {x : H | ∃ a : H →L[ℂ] H, a ∈ M ∧ x = a omega}

/-- K is the real closed span of self-adjoint vectors aΩ (§12, PDF p.18).
This does not assume standardness or any modular identities. -/
def closedSelfAdjointRealPart (M : VonNeumannAlgebra H) (omega : H) : Submodule ℝ H :=
  (Submodule.span ℝ {x : H | ∃ a : H →L[ℂ] H,
    a ∈ M ∧ IsSelfAdjoint a ∧ x = a omega}).topologicalClosure

/-- The real subspace iK uses the original complex structure (§12, PDF p.18). -/
def imaginaryRealPart (K : Submodule ℝ H) : Submodule ℝ H :=
  K.map ((Complex.I • (LinearMap.id : H →ₗ[ℂ] H)).restrictScalars ℝ)

example : imaginaryRealPart (⊥ : Submodule ℝ H) = ⊥ := by
  simp [imaginaryRealPart]

end StandardRealPart

end QAlgorithms.Papers.Independent.Intakenext100.P41
Source
Operator Algebras and Conformal Field Theory III. Fusion of positive energy representations of LSU(N) using bounded operators, arXiv:math/9806031v1, https://arxiv.org/abs/math/9806031v1
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

QAlgorithms.Papers.Independent.Intakenext100.P41.AddsOneBox

For every natural number NNN and every two integer-valued functions f,gf,gf,g on {0,…,N−1}\{0,\ldots,N-1\}{0,…,N−1}, this predicate means that there exists an index rrr in that index set such that, for every index iii,

gi={fi+1,i=r,fi,i≠r.g_i= \begin{cases} f_i+1,& i=r,\\ f_i,& i\ne r. \end{cases}gi​={fi​+1,fi​,​i=r,i=r.​

No ordering, sign, or bound on the differences of entries is required of either function. When N=0N=0N=0, the predicate is false because there is no index rrr.

QAlgorithms.Papers.Independent.Intakenext100.P41.AdjacentPaths

For natural numbers N,ℓ,kN,\ell,kN,ℓ,k, write IN={0,…,N−1}I_N=\{0,\ldots,N-1\}IN​={0,…,N−1}. An integer signature is a vector in ZIN\mathbb Z^{I_N}ZIN​; it is ℓ\ellℓ-permissible if its coordinates are weakly decreasing and fi−fj≤ℓf_i-f_j\leq\ellfi​−fj​≤ℓ for every i,j∈INi,j\in I_Ni,j∈IN​. Given integer signatures f,gf,gf,g and two paths P,QP,QP,Q, each consisting of k+1k+1k+1 ℓ\ellℓ-permissible signatures indexed by 0,…,k0,\ldots,k0,…,k, starting at fff, ending at ggg, and increasing exactly one coordinate by 111 at each step, PPP and QQQ are called adjacent precisely when there exists an index jjj with 0<j<k0<j<k0<j<k such that Pj≠QjP_j\neq Q_jPj​=Qj​ and Pi=QiP_i=Q_iPi​=Qi​ at every other index iii. Thus they differ at exactly one interior vertex. The definition allows zero values of N,ℓ,kN,\ell,kN,ℓ,k; when k=0k=0k=0 or k=1k=1k=1, no such interior index exists, so no pair is adjacent.

QAlgorithms.Papers.Independent.Intakenext100.P41.AffineWeylElement

For every natural number NNN, an affine Weyl element consists of an integer vector a=(ai)i∈INa=(a_i)_{i\in I_N}a=(ai​)i∈IN​​, where IN={0,…,N−1}I_N=\{0,\ldots,N-1\}IN​={0,…,N−1}, satisfying ∑i∈INai=0\sum_{i\in I_N}a_i=0∑i∈IN​​ai​=0, together with an arbitrary permutation σ\sigmaσ of INI_NIN​. These are the only required data and constraint. The case N=0N=0N=0 is included: the vector is empty, its sum is 000, and the permutation is the unique permutation of the empty set.

QAlgorithms.Papers.Independent.Intakenext100.P41.AngularCircle

AngularCircle\mathrm{AngularCircle}AngularCircle denotes the additive quotient group R/(2πZ)\mathbb R/(2\pi\mathbb Z)R/(2πZ): two real numbers represent the same point exactly when their difference is an integer multiple of 2π2\pi2π.

QAlgorithms.Papers.Independent.Intakenext100.P41.BasicODEData

For a natural number ddd, this structure consists of a proof that d>0d>0d>0, a complex d×dd\times dd×d matrix PPP, a vector v∈Cdv\in\mathbb C^dv∈Cd, a complex-linear functional ϕ:Cd→C\phi:\mathbb C^d\to\mathbb Cϕ:Cd→C with ϕ(v)≠0\phi(v)\ne0ϕ(v)=0, two indexed families λi,μi∈C\lambda_i,\mu_i\in\mathbb Cλi​,μi​∈C, and two bases (ξi)i=0d−1(\xi_i)_{i=0}^{d-1}(ξi​)i=0d−1​ and (ζi)i=0d−1(\zeta_i)_{i=0}^{d-1}(ζi​)i=0d−1​ of Cd\mathbb C^dCd. Define the matrix QQQ by Qij=viϕ(ej)Q_{ij}=v_i\phi(e_j)Qij​=vi​ϕ(ej​), where eje_jej​ is the standard coordinate vector. The data must satisfy, for every iii,

Pξi=λiξi,(P−Q)ζi=μiζi,ϕ(ξi)=ϕ(ζi)=1,P\xi_i=\lambda_i\xi_i,\qquad (P-Q)\zeta_i=\mu_i\zeta_i,\qquad \phi(\xi_i)=\phi(\zeta_i)=1,Pξi​=λi​ξi​,(P−Q)ζi​=μi​ζi​,ϕ(ξi​)=ϕ(ζi​)=1,

and the coefficient of vvv at position iii in each of the bases (ξi)(\xi_i)(ξi​) and (ζi)(\zeta_i)(ζi​) must be nonzero. For every pair of distinct indices i,ji,ji,j and every integer nnn, the data must also satisfy λi−λj≠n\lambda_i-\lambda_j\ne nλi​−λj​=n and μi−μj≠n\mu_i-\mu_j\ne nμi​−μj​=n, with integers viewed as complex numbers. There are no such data when d=0d=0d=0; when d=1d=1d=1, the conditions involving distinct indices are vacuous.

QAlgorithms.Papers.Independent.Intakenext100.P41.BoxPath

For every natural number NNN, natural numbers ℓ,k\ell,kℓ,k, and integer-valued functions f,gf,gf,g on {0,…,N−1}\{0,\ldots,N-1\}{0,…,N−1}, this structure consists of a sequence of integer-valued functions h(0),…,h(k)h^{(0)},\ldots,h^{(k)}h(0),…,h(k) on that same index set, together with the requirements that h(0)=fh^{(0)}=fh(0)=f and h(k)=gh^{(k)}=gh(k)=g; that for every stage r∈{0,…,k}r\in\{0,\ldots,k\}r∈{0,…,k} and all indices i,ji,ji,j, i≤ji\le ji≤j implies hj(r)≤hi(r)h^{(r)}_j\le h^{(r)}_ihj(r)​≤hi(r)​, and hi(r)−hj(r)≤ℓh^{(r)}_i-h^{(r)}_j\le\ellhi(r)​−hj(r)​≤ℓ; and that for every r∈{0,…,k−1}r\in\{0,\ldots,k-1\}r∈{0,…,k−1} there exists an index qqq such that hq(r+1)=hq(r)+1h^{(r+1)}_q=h^{(r)}_q+1hq(r+1)​=hq(r)​+1 and hi(r+1)=hi(r)h^{(r+1)}_i=h^{(r)}_ihi(r+1)​=hi(r)​ for every i≠qi\ne qi=q. All entries are integers and may be negative. When k=0k=0k=0, there are no step requirements, and such data exist precisely when f=gf=gf=g and this common function satisfies the stated ordering and difference bounds. When N=0N=0N=0, such data exist for k=0k=0k=0, while for k>0k>0k>0 the required choice of a row at each step is impossible.

QAlgorithms.Papers.Independent.Intakenext100.P41.CircleL2

For each natural number NNN, CircleL2(N)\mathrm{CircleL2}(N)CircleL2(N) denotes the L2L^2L2 space of functions from R/(2πZ)\mathbb R/(2\pi\mathbb Z)R/(2πZ) to the complex Euclidean space CN\mathbb C^NCN, with coordinates indexed by 0,…,N−10,\ldots,N-10,…,N−1, using the designated additive Haar measure on the circle. Its elements are equivalence classes, modulo equality almost everywhere, of measurable functions with integrable squared Euclidean norm. The definition includes N=0N=0N=0, when the target space and this L2L^2L2 space are both zero spaces.

QAlgorithms.Papers.Independent.Intakenext100.P41.HasSimpleSpectrum

For every natural number ddd, including 000, and every matrix P∈Md(C)P\in M_d(\mathbb C)P∈Md​(C), this property means that there exist a complex basis (ei)0≤i<d(e_i)_{0\le i<d}(ei​)0≤i<d​ of Cd\mathbb C^dCd and complex numbers (λi)0≤i<d(\lambda_i)_{0\le i<d}(λi​)0≤i<d​ such that the map i↦λii\mapsto\lambda_ii↦λi​ is injective and

Pei=λieifor every 0≤i<d.Pe_i=\lambda_i e_i\qquad\text{for every }0\le i<d.Pei​=λi​ei​for every 0≤i<d.

For d=0d=0d=0, the empty basis and empty family of eigenvalues satisfy these conditions.

QAlgorithms.Papers.Independent.Intakenext100.P41.HasTrivialRealStabilizer

For every natural number NNN, every natural number ℓ\ellℓ, and every real vector x∈RNx\in\mathbb R^Nx∈RN, this property means that, for every integer vector m∈ZNm\in\mathbb Z^Nm∈ZN with ∑i=0N−1mi=0\sum_{i=0}^{N-1}m_i=0∑i=0N−1​mi​=0 and every permutation σ\sigmaσ of {0,…,N−1}\{0,\ldots,N-1\}{0,…,N−1}, the implication

[∀i<N, xσ−1(i)+(N+ℓ)mi=xi] ⟹ [m=0 ∧ σ=id]\bigl[\forall i<N,\ x_{\sigma^{-1}(i)}+(N+\ell)m_i=x_i\bigr] \ \Longrightarrow\ \bigl[m=0\ \land\ \sigma=\mathrm{id}\bigr][∀i<N, xσ−1(i)​+(N+ℓ)mi​=xi​] ⟹ [m=0 ∧ σ=id]

holds. Both N=0N=0N=0 and ℓ=0\ell=0ℓ=0 are included; for N=0N=0N=0, the property holds because the integer vector and the permutation are uniquely the empty vector and the identity permutation.

QAlgorithms.Papers.Independent.Intakenext100.P41.HasTrivialStabilizer

For every natural number NNN, natural number ℓ\ellℓ, and integer vector xxx indexed by IN={0,…,N−1}I_N=\{0,\ldots,N-1\}IN​={0,…,N−1}, the assertion that xxx has trivial stabilizer means the following: for every integer vector aaa with ∑iai=0\sum_i a_i=0∑i​ai​=0 and every permutation σ\sigmaσ of INI_NIN​, if xσ−1(i)+(N+ℓ)ai=xix_{\sigma^{-1}(i)}+(N+\ell)a_i=x_ixσ−1(i)​+(N+ℓ)ai​=xi​ for all i∈INi\in I_Ni∈IN​, then ai=0a_i=0ai​=0 for every iii and σ\sigmaσ is the identity permutation. The case N=0N=0N=0 is included, and the condition then holds because the empty vector and identity permutation are the only possibilities.

QAlgorithms.Papers.Independent.Intakenext100.P41.InAffineRootLattice

For every natural number NNN, natural number ℓ\ellℓ, and integer vector xxx indexed by IN={0,…,N−1}I_N=\{0,\ldots,N-1\}IN​={0,…,N−1}, membership in the specified affine root lattice means that there exists an integer vector aaa with ∑i∈INai=0\sum_{i\in I_N}a_i=0∑i∈IN​​ai​=0 such that xi=(N+ℓ)aix_i=(N+\ell)a_ixi​=(N+ℓ)ai​ for every i∈INi\in I_Ni∈IN​. Zero values of NNN and ℓ\ellℓ are allowed. In particular, when N=0N=0N=0, the sole empty vector satisfies this condition, including when ℓ=0\ell=0ℓ=0.

QAlgorithms.Papers.Independent.Intakenext100.P41.IntegerSignature

For every natural number NNN, an integer signature is a function from the index set {0,…,N−1}\{0,\ldots,N-1\}{0,…,N−1} to Z\mathbb ZZ, equivalently an indexed tuple (f0,…,fN−1)(f_0,\ldots,f_{N-1})(f0​,…,fN−1​) of arbitrary integers. This abbreviation imposes no ordering or sign condition. When N=0N=0N=0, it is the unique function on the empty index set.

QAlgorithms.Papers.Independent.Intakenext100.P41.IsAffineFundamentalDomain

For natural numbers N,ℓN,\ellN,ℓ and a subset D⊆RND\subseteq\mathbb R^ND⊆RN, this property means that DDD is closed; for every x∈RNx\in\mathbb R^Nx∈RN, there exist a permutation σ\sigmaσ of the coordinate indices and an integer vector mmm with ∑imi=0\sum_i m_i=0∑i​mi​=0 such that the vector with coordinates xσ−1(i)+(N+ℓ)mix_{\sigma^{-1}(i)}+(N+\ell)m_ixσ−1(i)​+(N+ℓ)mi​ belongs to DDD; and, for every such pair (m,σ)(m,\sigma)(m,σ) and every x∈RNx\in\mathbb R^Nx∈RN, if both xxx and that transformed vector belong to the interior of DDD, then m=0m=0m=0 and σ\sigmaσ is the identity permutation. The interior is taken in RN\mathbb R^NRN. Zero values of NNN and ℓ\ellℓ are allowed; for N=0N=0N=0, there is only the empty integer vector and the identity permutation, and coverage requires DDD to contain the unique empty vector.

QAlgorithms.Papers.Independent.Intakenext100.P41.IsCyclicFor

For a von Neumann algebra MMM of bounded complex-linear operators on the ambient space HHH, and a vector ω∈H\omega\in Hω∈H, this property means that the set {aω:a∈M}\{a\omega:a\in M\}{aω:a∈M} is dense in HHH in its norm topology. The vector ω\omegaω is not separately required to be nonzero; if ω=0\omega=0ω=0, this density condition holds exactly when HHH is the zero space.

QAlgorithms.Papers.Independent.Intakenext100.P41.IsNormalizedGauge

For arbitrary natural numbers d,md,md,m, a real number RRR, and functions B,g:Cm×C→Mat⁡d×d(C)B,g:\mathbb C^m\times\mathbb C\to\operatorname{Mat}_{d\times d}(\mathbb C)B,g:Cm×C→Matd×d​(C), let UUU consist of those parameters ttt for which no ordered pair of roots λ,μ\lambda,\muλ,μ of the characteristic polynomial of B(t,0)B(t,0)B(t,0) satisfies λ−μ=n\lambda-\mu=nλ−μ=n for any positive integer nnn, and let D={(t,z):t∈U, ∣z∣<R}D=\{(t,z):t\in U,\ |z|<R\}D={(t,z):t∈U, ∣z∣<R}. The assertion that ggg is a normalized gauge means that ggg is complex differentiable on DDD, det⁡g(t,z)≠0\det g(t,z)\neq0detg(t,z)=0 at every point of DDD, g(t,0)=Idg(t,0)=I_dg(t,0)=Id​ for every t∈Ut\in Ut∈U, and, for every t∈Ut\in Ut∈U and every nonzero z∈Cz\in\mathbb Cz∈C with ∣z∣<R|z|<R∣z∣<R, it satisfies g(t,z)−1(z−1B(t,z))g(t,z)−g(t,z)−1ddzg(t,z)=z−1B(t,0)g(t,z)^{-1}\bigl(z^{-1}B(t,z)\bigr)g(t,z)-g(t,z)^{-1}\frac{d}{dz}g(t,z)=z^{-1}B(t,0)g(t,z)−1(z−1B(t,z))g(t,z)−g(t,z)−1dzd​g(t,z)=z−1B(t,0), where the derivative holds ttt fixed and multiplication between matrices is matrix multiplication. Zero dimensions are allowed; when d=0d=0d=0, U=CmU=\mathbb C^mU=Cm. When R≤0R\leq0R≤0, DDD is empty and the differentiability, determinant, and differential-equation requirements are vacuous, but the requirement g(t,0)=Idg(t,0)=I_dg(t,0)=Id​ for every t∈Ut\in Ut∈U remains. If UUU is empty, all four requirements are vacuous.

QAlgorithms.Papers.Independent.Intakenext100.P41.IsPermissible

For every natural number NNN, every natural number ℓ\ellℓ, and every integer-valued function fff on {0,…,N−1}\{0,\ldots,N-1\}{0,…,N−1}, this predicate means that both of the following hold: for all indices i,ji,ji,j with i≤ji\le ji≤j, one has fj≤fif_j\le f_ifj​≤fi​; and for all indices i,ji,ji,j, one has fi−fj≤ℓf_i-f_j\le\ellfi​−fj​≤ℓ, where ℓ\ellℓ is viewed as an integer. The subtraction is integer subtraction. When N=0N=0N=0, both conditions are vacuous; when N=1N=1N=1, every such function satisfies them. When ℓ=0\ell=0ℓ=0, the conditions require all entries to be equal.

QAlgorithms.Papers.Independent.Intakenext100.P41.IsRealRootTranslation

For every natural number NNN, every natural number ℓ\ellℓ, and every vector x∈RNx\in\mathbb R^Nx∈RN, this property means that there exists an integer vector m∈ZNm\in\mathbb Z^Nm∈ZN such that

∑i=0N−1mi=0andxi=(N+ℓ)mifor every 0≤i<N.\sum_{i=0}^{N-1}m_i=0 \qquad\text{and}\qquad x_i=(N+\ell)m_i\quad\text{for every }0\le i<N.i=0∑N−1​mi​=0andxi​=(N+ℓ)mi​for every 0≤i<N.

The parameters may be zero. For N=0N=0N=0, the unique empty vector satisfies the property, with the empty sum equal to zero.

QAlgorithms.Papers.Independent.Intakenext100.P41.IsSignature

For every natural number NNN and every integer-valued function fff on {0,…,N−1}\{0,\ldots,N-1\}{0,…,N−1}, this predicate means that for all indices i,ji,ji,j, if i≤ji\le ji≤j then fj≤fif_j\le f_ifj​≤fi​. Thus the entries are weakly decreasing in their index, with no restriction on their signs. For N=0N=0N=0 or N=1N=1N=1, the predicate holds for every such function.

QAlgorithms.Papers.Independent.Intakenext100.P41.IsTranslationFundamentalDomain

For natural numbers N,ℓN,\ellN,ℓ and a subset D⊆RND\subseteq\mathbb R^ND⊆RN, this property means that DDD is closed; for every x∈RNx\in\mathbb R^Nx∈RN, there exist a∈RNa\in\mathbb R^Na∈RN and integers m0,…,mN−1m_0,\ldots,m_{N-1}m0​,…,mN−1​ with ∑imi=0\sum_i m_i=0∑i​mi​=0 such that ai=(N+ℓ)mia_i=(N+\ell)m_iai​=(N+ℓ)mi​ for every iii and x+a∈Dx+a\in Dx+a∈D; and, for every x,a∈RNx,a\in\mathbb R^Nx,a∈RN, if aaa has such an integer representation and both xxx and x+ax+ax+a belong to the interior of DDD, then a=0a=0a=0. The interior is taken in the usual topology on the whole space RN\mathbb R^NRN. No positivity assumption on NNN or ℓ\ellℓ is imposed; when N=0N=0N=0, all coordinate conditions and the zero-sum condition are vacuous, and the coverage requirement forces DDD to contain the unique empty vector.

QAlgorithms.Papers.Independent.Intakenext100.P41.NoCommonRoot

For any two polynomials a,b∈C[X]a,b\in\mathbb C[X]a,b∈C[X], this property means that there is no complex number at which both polynomials vanish:

∀z∈C,¬(a(z)=0 ∧ b(z)=0).\forall z\in\mathbb C,\qquad \neg\bigl(a(z)=0\ \land\ b(z)=0\bigr).∀z∈C,¬(a(z)=0 ∧ b(z)=0).

The polynomials may be zero polynomials; if one is zero, the property requires the other to have no complex root, and if both are zero, the property fails.

QAlgorithms.Papers.Independent.Intakenext100.P41.RankOneGeneralPosition

For every natural number ddd, including 000, and any matrices P,Q∈Md(C)P,Q\in M_d(\mathbb C)P,Q∈Md​(C), this property means that there exist a complex basis (ei)0≤i<d(e_i)_{0\le i<d}(ei​)0≤i<d​ of Cd\mathbb C^dCd and an injective family of complex numbers (λi)0≤i<d(\lambda_i)_{0\le i<d}(λi​)0≤i<d​ such that, for every index iii, Pei=λieiPe_i=\lambda_i e_iPei​=λi​ei​, Qei≠0Qe_i\ne0Qei​=0, and there exists a vector x∈Cdx\in\mathbb C^dx∈Cd for which the iii-th coordinate of QxQxQx in the basis eee is nonzero. The vector xxx may depend on iii. For d=0d=0d=0, all the indexed conditions are vacuous, and the property holds.

QAlgorithms.Papers.Independent.Intakenext100.P41.VanishesOnArc

For any natural number NNN, any element fff of L2(R/(2πZ);CN)L^2(\mathbb R/(2\pi\mathbb Z);\mathbb C^N)L2(R/(2πZ);CN) with respect to the designated additive Haar measure, and any real numbers a,ba,ba,b, this predicate means that, for almost every circle point θ\thetaθ, if θ\thetaθ has a real representative in (a,b)(a,b)(a,b), then f(θ)=0f(\theta)=0f(θ)=0. Equivalently, fff is zero almost everywhere on the image of (a,b)(a,b)(a,b) in the quotient circle. No ordering or length assumption is imposed on a,ba,ba,b; when a≥ba\ge ba≥b, the predicate holds vacuously.

QAlgorithms.Papers.Independent.Intakenext100.P41.affineAction

For every natural number NNN, every natural number ℓ\ellℓ, every pair w=(a,σ)w=(a,\sigma)w=(a,σ) consisting of an integer vector a∈ZINa\in\mathbb Z^{I_N}a∈ZIN​ with ∑iai=0\sum_i a_i=0∑i​ai​=0 and a permutation σ\sigmaσ of IN={0,…,N−1}I_N=\{0,\ldots,N-1\}IN​={0,…,N−1}, and every integer vector x∈ZINx\in\mathbb Z^{I_N}x∈ZIN​, the affine action produces the integer vector whose iiith coordinate is xσ−1(i)+(N+ℓ)aix_{\sigma^{-1}(i)}+(N+\ell)a_ixσ−1(i)​+(N+ℓ)ai​. There are no positivity assumptions on NNN or ℓ\ellℓ. When N=0N=0N=0, the resulting vector is empty.

QAlgorithms.Papers.Independent.Intakenext100.P41.angularArc

For any real numbers a,ba,ba,b, this defines the subset of R/(2πZ)\mathbb R/(2\pi\mathbb Z)R/(2πZ) consisting of all classes having a real representative strictly between aaa and bbb:

{θ mod 2π:a<θ<b}.\{\theta\bmod 2\pi:a<\theta<b\}.{θmod2π:a<θ<b}.

There is no restriction on the endpoints or the interval length; in particular, the set is empty when a≥ba\ge ba≥b, and representatives are identified modulo 2π2\pi2π.

QAlgorithms.Papers.Independent.Intakenext100.P41.angularHaar

angularHaar\mathrm{angularHaar}angularHaar denotes the designated additive Haar measure on the circle R/(2πZ)\mathbb R/(2\pi\mathbb Z)R/(2πZ).

QAlgorithms.Papers.Independent.Intakenext100.P41.angularPeriodPositive

This declaration supplies the fact that the real period 2π2\pi2π is strictly positive: 0<2π0<2\pi0<2π.

QAlgorithms.Papers.Independent.Intakenext100.P41.circleHardySpace

For every natural number NNN, this defines a complex linear subspace of L2(R/(2πZ);CN)L^2(\mathbb R/(2\pi\mathbb Z);\mathbb C^N)L2(R/(2πZ);CN) with respect to the designated additive Haar measure: it is the closure, in the L2L^2L2 norm, of the complex linear span of all equivalence classes represented by θ↦einθv\theta\mapsto e^{in\theta}vθ↦einθv, for natural numbers nnn and arbitrary vectors v∈CNv\in\mathbb C^Nv∈CN. Equivalently, its elements are L2L^2L2 limits of finite sums of these functions. The frequency n=0n=0n=0 is included, and when N=0N=0N=0 this subspace is the zero space.

QAlgorithms.Papers.Independent.Intakenext100.P41.closedSelfAdjointRealPart

For a von Neumann algebra MMM of bounded complex-linear operators on the ambient space HHH, and any vector ω∈H\omega\in Hω∈H, this defines a real linear subspace of HHH: the norm closure of the real linear span of all vectors aωa\omegaaω with a∈Ma\in Ma∈M self-adjoint, meaning a∗=aa^*=aa∗=a. Equivalently, it is the closure of the set of finite real linear combinations of such vectors. No cyclicity or nonzero assumption on ω\omegaω is imposed; when ω=0\omega=0ω=0, the resulting subspace is {0}\{0\}{0}.

QAlgorithms.Papers.Independent.Intakenext100.P41.companionMatrix

For every natural number ddd, including 000, and every vector a=(a0,…,ad−1)∈Cda=(a_0,\ldots,a_{d-1})\in\mathbb C^da=(a0​,…,ad−1​)∈Cd, this defines the d×dd\times dd×d complex matrix M(a)M(a)M(a), with rows and columns indexed from 000 to d−1d-1d−1, by

M(a)ij={aj,i+1=d,1,i+1≠d and j=i+1,0,otherwise.M(a)_{ij}= \begin{cases} a_j,&i+1=d,\\ 1,&i+1\ne d\ \text{and}\ j=i+1,\\ 0,&\text{otherwise}. \end{cases}M(a)ij​=⎩⎨⎧​aj​,1,0,​i+1=d,i+1=d and j=i+1,otherwise.​

Thus the last row is aaa, and each preceding row has a single 111 immediately above the diagonal. For d=0d=0d=0, this is the empty matrix; for d=1d=1d=1, it is the matrix (a0)(a_0)(a0​).

QAlgorithms.Papers.Independent.Intakenext100.P41.companionPolynomial

For every natural number ddd, including 000, and every vector a∈Cda\in\mathbb C^da∈Cd, this defines the complex polynomial

pa(X)=Xd−∑i=0d−1aiXi.p_a(X)=X^d-\sum_{i=0}^{d-1}a_iX^i.pa​(X)=Xd−i=0∑d−1​ai​Xi.

When d=0d=0d=0, the sum is empty and the polynomial is 111.

QAlgorithms.Papers.Independent.Intakenext100.P41.dominantAlcove

For natural numbers NNN and ℓ\ellℓ, this defines the subset of RN\mathbb R^NRN, with coordinates indexed by 0,…,N−10,\ldots,N-10,…,N−1, consisting of all vectors xxx such that xj≤xix_j\le x_ixj​≤xi​ whenever i≤ji\le ji≤j, and xi−xj≤N+ℓx_i-x_j\le N+\ellxi​−xj​≤N+ℓ for every pair of coordinate indices i,ji,ji,j. Thus the coordinates are weakly decreasing and every ordered coordinate difference is at most N+ℓN+\ellN+ℓ. Both N=0N=0N=0 and ℓ=0\ell=0ℓ=0 are allowed; when N=0N=0N=0, the coordinate conditions are vacuous and the set contains the unique empty vector.

QAlgorithms.Papers.Independent.Intakenext100.P41.frobeniusPair

For every natural number ddd, every pair of complex d×dd\times dd×d matrices P,QP,QP,Q, every λ∈C\lambda\in\mathbb Cλ∈C, and every w∈Cdw\in\mathbb C^dw∈Cd, this definition gives a sequence of ordered pairs (an,sn)∈Cd×Cd(a_n,s_n)\in\mathbb C^d\times\mathbb C^d(an​,sn​)∈Cd×Cd, indexed by all natural numbers nnn, through

a0=w,s0=w,an+1=((λ+n+1)Id−P)−1Qsn,sn+1=sn+an+1.a_0=w,\qquad s_0=w,\qquad a_{n+1}=\bigl((\lambda+n+1)I_d-P\bigr)^{-1}Qs_n,\qquad s_{n+1}=s_n+a_{n+1}.a0​=w,s0​=w,an+1​=((λ+n+1)Id​−P)−1Qsn​,sn+1​=sn​+an+1​.

The natural number n+1n+1n+1 in the matrix expression is viewed as a complex number. There are no invertibility hypotheses: the matrix inverse is total and is the zero matrix when its argument is singular. The definition includes d=0d=0d=0, in which case both components are the unique vector in the zero-dimensional space.

QAlgorithms.Papers.Independent.Intakenext100.P41.gaugeDomain

For arbitrary natural numbers d,md,md,m, a real number RRR, and a function B:Cm×C→Mat⁡d×d(C)B:\mathbb C^m\times\mathbb C\to\operatorname{Mat}_{d\times d}(\mathbb C)B:Cm×C→Matd×d​(C), the gauge domain is the set of pairs (t,z)(t,z)(t,z) such that ∣z∣<R|z|<R∣z∣<R and, for every two roots λ,μ\lambda,\muλ,μ of the characteristic polynomial of B(t,0)B(t,0)B(t,0) and every positive integer nnn, λ−μ≠n\lambda-\mu\neq nλ−μ=n. No positivity assumption on RRR or regularity assumption on BBB is imposed. If R≤0R\leq0R≤0, this domain is empty. If d=0d=0d=0, the characteristic polynomial has no roots, so the domain is all of Cm\mathbb C^mCm times the open disk of radius RRR; if m=0m=0m=0, there is just one possible parameter tuple.

QAlgorithms.Papers.Independent.Intakenext100.P41.gaugeParameterSet

For arbitrary natural numbers d,md,md,m and any function B:Cm×C→Mat⁡d×d(C)B:\mathbb C^m\times\mathbb C\to\operatorname{Mat}_{d\times d}(\mathbb C)B:Cm×C→Matd×d​(C), the gauge parameter set consists of those t∈Cmt\in\mathbb C^mt∈Cm such that, for every pair λ,μ∈C\lambda,\mu\in\mathbb Cλ,μ∈C that are roots of the characteristic polynomial of B(t,0)B(t,0)B(t,0) and every positive natural number nnn, one has λ−μ≠n\lambda-\mu\neq nλ−μ=n. The roots need not be distinct; this condition does not exclude equal roots. Since the pair is ordered and universally quantified, it also excludes differences that are negative nonzero integers. No regularity of BBB is assumed. When d=0d=0d=0, the characteristic polynomial is 111, so the root hypotheses never hold and every parameter belongs to the set; when m=0m=0m=0, the parameter space consists of a single empty tuple.

QAlgorithms.Papers.Independent.Intakenext100.P41.imaginaryRealPart

For any real linear subspace KKK of the ambient complex vector space HHH, this defines the real linear subspace

iK={iv:v∈K}⊆H,iK=\{iv:v\in K\}\subseteq H,iK={iv:v∈K}⊆H,

where iii is the imaginary unit and multiplication by iii is regarded as a real-linear map. No closedness assumption on KKK is imposed, and no closure is taken in this definition; in particular, K={0}K=\{0\}K={0} gives iK={0}iK=\{0\}iK={0}.

QAlgorithms.Papers.Independent.Intakenext100.P41.lastRowMatrix

For every natural number ddd, including 000, and every vector b=(b0,…,bd−1)∈Cdb=(b_0,\ldots,b_{d-1})\in\mathbb C^db=(b0​,…,bd−1​)∈Cd, this defines the d×dd\times dd×d complex matrix L(b)L(b)L(b) by

L(b)ij={bj,i+1=d,0,otherwise,0≤i,j<d.L(b)_{ij}= \begin{cases} b_j,&i+1=d,\\ 0,&\text{otherwise}, \end{cases} \qquad 0\le i,j<d.L(b)ij​={bj​,0,​i+1=d,otherwise,​0≤i,j<d.

For positive ddd, only its last row can be nonzero, and that row is bbb. For d=0d=0d=0, this is the empty matrix.

QAlgorithms.Papers.Independent.Intakenext100.P41.linearRankOneMatrix

For every natural number ddd, every vector v∈Cdv\in\mathbb C^dv∈Cd, and every complex linear functional ϕ:Cd→C\phi:\mathbb C^d\to\mathbb Cϕ:Cd→C, this defines the d×dd\times dd×d complex matrix QQQ whose entries are

Qij=vi ϕ(ej)(0≤i,j<d),Q_{ij}=v_i\,\phi(e_j)\qquad(0\le i,j<d),Qij​=vi​ϕ(ej​)(0≤i,j<d),

where eje_jej​ has coordinate 111 at jjj and 000 elsewhere. Thus Qx=ϕ(x)vQx=\phi(x)vQx=ϕ(x)v. Neither vvv nor ϕ\phiϕ is assumed nonzero, so the matrix may be zero; d=0d=0d=0 is also allowed and gives the empty matrix.

QAlgorithms.Papers.Independent.Intakenext100.P41.projectedCoefficient

Let ddd be a natural number and let AAA consist of a positive dimension d>0d>0d>0, a matrix P∈Md(C)P\in M_d(\mathbb C)P∈Md​(C), a vector v∈Cdv\in\mathbb C^dv∈Cd, a complex-linear functional ϕ\phiϕ with ϕ(v)≠0\phi(v)\ne0ϕ(v)=0, families (λi)(\lambda_i)(λi​) and (μi)(\mu_i)(μi​), and bases (ξi)(\xi_i)(ξi​) and (ζi)(\zeta_i)(ζi​) of Cd\mathbb C^dCd. Define Qij=viϕ(ej)Q_{ij}=v_i\phi(e_j)Qij​=vi​ϕ(ej​) using the standard coordinate vectors. Require that Pξi=λiξiP\xi_i=\lambda_i\xi_iPξi​=λi​ξi​, (P−Q)ζi=μiζi(P-Q)\zeta_i=\mu_i\zeta_i(P−Q)ζi​=μi​ζi​, and ϕ(ξi)=ϕ(ζi)=1\phi(\xi_i)=\phi(\zeta_i)=1ϕ(ξi​)=ϕ(ζi​)=1 for every iii; that every coordinate of vvv in both bases is nonzero; and that for distinct i,ji,ji,j neither λi−λj\lambda_i-\lambda_jλi​−λj​ nor μi−μj\mu_i-\mu_jμi​−μj​ equals any integer viewed in C\mathbb CC. For every index i∈{0,…,d−1}i\in\{0,\ldots,d-1\}i∈{0,…,d−1} and natural number nnn, define vectors by

si,0=ξi,ai,r+1=((λi+r+1)Id−P)−1Qsi,r,si,r+1=si,r+ai,r+1(r≥0).s_{i,0}=\xi_i,\qquad a_{i,r+1}=\bigl((\lambda_i+r+1)I_d-P\bigr)^{-1}Qs_{i,r}, \qquad s_{i,r+1}=s_{i,r}+a_{i,r+1}\quad(r\ge0).si,0​=ξi​,ai,r+1​=((λi​+r+1)Id​−P)−1Qsi,r​,si,r+1​=si,r​+ai,r+1​(r≥0).

The projected coefficient at (i,n)(i,n)(i,n) is the complex number ϕ(si,n)\phi(s_{i,n})ϕ(si,n​). Matrix inversion uses its total convention, assigning the zero matrix to a singular matrix. In particular, at n=0n=0n=0 the coefficient is ϕ(ξi)=1\phi(\xi_i)=1ϕ(ξi​)=1.

QAlgorithms.Papers.Independent.Intakenext100.P41.realAffineAction

For every natural number NNN, every natural number ℓ\ellℓ, every pair w=(m,σ)w=(m,\sigma)w=(m,σ) consisting of an integer vector m∈ZNm\in\mathbb Z^Nm∈ZN with ∑i=0N−1mi=0\sum_{i=0}^{N-1}m_i=0∑i=0N−1​mi​=0 and a permutation σ\sigmaσ of {0,…,N−1}\{0,\ldots,N-1\}{0,…,N−1}, and every real vector x∈RNx\in\mathbb R^Nx∈RN, this defines a real vector by

(w⋅ℓx)i=xσ−1(i)+(N+ℓ)mi(0≤i<N).(w\cdot_\ell x)_i=x_{\sigma^{-1}(i)}+(N+\ell)m_i \qquad(0\le i<N).(w⋅ℓ​x)i​=xσ−1(i)​+(N+ℓ)mi​(0≤i<N).

Both N=0N=0N=0 and ℓ=0\ell=0ℓ=0 are allowed; when N=0N=0N=0, the resulting vector is empty.

QAlgorithms.Papers.Independent.Intakenext100.P41.resolventScalar

For every natural number ddd, every d×dd\times dd×d complex matrix AAA, every vector v∈Cdv\in\mathbb C^dv∈Cd, every complex linear functional ϕ:Cd→C\phi:\mathbb C^d\to\mathbb Cϕ:Cd→C, and every t∈Ct\in\mathbb Ct∈C, this defines the complex number

1+ϕ((tId−A)−1v).1+\phi\bigl((tI_d-A)^{-1}v\bigr).1+ϕ((tId​−A)−1v).

There is no assumption that tId−AtI_d-AtId​−A is invertible: the matrix inverse is a total operation that returns the zero matrix when its determinant is zero, so the defined scalar is 111 in that case. The definition also includes d=0d=0d=0, when the vector space is zero and the scalar is 111.

QAlgorithms.Papers.Independent.Intakenext100.P41.rowPolynomial

For every natural number ddd, including 000, and every vector b∈Cdb\in\mathbb C^db∈Cd, this defines the complex polynomial

rb(X)=∑i=0d−1biXi.r_b(X)=\sum_{i=0}^{d-1}b_iX^i.rb​(X)=i=0∑d−1​bi​Xi.

When d=0d=0d=0, this is the empty sum, hence the zero polynomial.

QAlgorithms.Papers.Independent.Intakenext100.P41.translationAlcove

For every pair of natural numbers N,ℓN,\ellN,ℓ, this defines the subset

{x∈RN: ∀ 0≤i,j<N,∣xi−xj∣≤N+ℓ}.\left\{x\in\mathbb R^N:\ \forall\,0\le i,j<N,\quad |x_i-x_j|\le N+\ell\right\}.{x∈RN: ∀0≤i,j<N,∣xi​−xj​∣≤N+ℓ}.

The bound is non-strict, so equality is allowed. The parameters may be zero; when N=0N=0N=0, the condition is vacuous and the set contains the unique empty vector, and when N=1N=1N=1, it contains every real vector.

QAlgorithms.Papers.Independent.Intakenext100.P41.vectorMonomial

For every pair of natural numbers N,nN,nN,n and every vector v∈CNv\in\mathbb C^Nv∈CN with its Euclidean norm, this defines the continuous function on R/(2πZ)\mathbb R/(2\pi\mathbb Z)R/(2πZ) given by θ↦einθv\theta\mapsto e^{in\theta}vθ↦einθv, where any real representative of θ\thetaθ gives the same value. Only nonnegative frequencies nnn occur; n=0n=0n=0 gives the constant function vvv, and N=0N=0N=0 or v=0v=0v=0 gives the zero function.

QAlgorithms.Papers.Independent.Intakenext100.P41.vectorMonomialL2

For every pair of natural numbers N,nN,nN,n and every vector v∈CNv\in\mathbb C^Nv∈CN, this defines the element of L2(R/(2πZ);CN)L^2(\mathbb R/(2\pi\mathbb Z);\mathbb C^N)L2(R/(2πZ);CN), measured using the designated additive Haar measure, represented by the continuous function θ↦einθv\theta\mapsto e^{in\theta}vθ↦einθv. Thus the result is an equivalence class modulo equality almost everywhere. The definition includes n=0n=0n=0, giving the class of the constant function vvv, and N=0N=0N=0, giving the zero element.

QAlgorithms.Papers.Independent.Intakenext100.P41.weylShift

For every natural number NNN, the Weyl shift is the integer vector ρ\rhoρ indexed by IN={0,…,N−1}I_N=\{0,\ldots,N-1\}IN​={0,…,N−1} given by ρi=N−1−i\rho_i=N-1-iρi​=N−1−i. Both subtractions are performed in the integers. For N>0N>0N>0 this vector is (N−1,N−2,…,0)(N-1,N-2,\ldots,0)(N−1,N−2,…,0); for N=0N=0N=0 it is the empty vector, so the formula has no coordinate at which to be evaluated.

Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 31, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me