Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Definitions for Bounded functional calculi for unbounded operators

Definition
QAlgorithms_Papers_Independent_Intakenext100_P35

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

prove2quantum

The objects the formalized statements of Bounded functional calculi for unbounded operators (arXiv:2202.03143) are stated in. Each was checked against the paper by a blind read-back and a moderator.

  • AnalyticSymbol
  • hp: The open right half-plane (source p.4).
  • closedHp: The closed right half-plane (source p.4).
  • openSectorDomain: The source's open sector, including its special angle-zero convention (p.4).
  • point: Cartesian parametrization of the right half-plane.
  • positiveVolume: Lebesgue measure on the positive alpha axis. Endpoints have measure zero.
  • Hol: Holomorphicity on the actual symbol domain.
  • supSize: Supremum in an extended lattice, never a default real supremum.
  • HInfinity: A bounded holomorphic symbol on its native domain.
  • limitValue: A guarded limit selector. Its fallback is never asserted to be a limit.
  • infinityValue: The canonical infinity value selected along the positive real axis.
  • zeroValue: The canonical zero value selected along the positive real axis.
  • extension: Continuous boundary extension, when it exists; interior values are unchanged.
  • infinityWithin: Norm infinity filter restricted to a native symbol domain.
  • SectorialLimits: Both sectorial limits, with the same values for every strict subsector (p.20).
  • bSemi: Besov seminorm, equation (3.1), without a boundedness hypothesis (p.6).
  • BesovClass: The space B as it is defined before Proposition 3.1.
  • bSize: Full Besov norm (p.7), in an extended lattice off its natural domain.
  • BZero: The ideal B0, with vanishing canonical infinity value (p.7).
  • wSize: Mixed W norm uses an essential, rather than pointwise, supremum (3.3).
  • MixedDerivativeClass: Measurable representatives of W. All W assertions below are a.e. invariant.
  • AbsoluteDouble: Absolute convergence of a double scalar integral, including measurability.
  • qKernel: The reflected resolvent kernel from (3.5), p.8.
  • besovPrimitive: Q is computed from its scalar integral; B0 membership is not assumed.
  • eSemi: The derivative dual seminorm from p.11.
  • DerivativeDualClass: The holomorphic space E.
  • besovPairing: The bilinear (not sesquilinear) partial duality (3.9).
  • LaplaceRep: A positive finite control measure and an integrable complex density represent an arbitrary finite complex measure; no positivity of the density is imposed.
  • LM: The full finite-complex-measure Laplace class LM (p.5).
  • hpSize: HP norm: infimum of total variations over representations. The empty infimum is infinity, so non-LM symbols cannot have a spurious zero norm.
  • vSize: Extended mixed weighted norm (4.3); alpha is positive in this integral.
  • WeightedDerivativeClass: Measurable representatives of V_s.
  • D: D_s imposes only the source's holomorphic derivative-integrability condition.
  • dSemi: The D_s seminorm and norm (pp.18--20).
  • dSize: The full D_s norm retains its infinity term separately from the seminorm.
  • DUnion: D_infinity is a union, not a boundedness condition.
  • DBounded: Bounded subalgebra D_s^infinity.
  • dBoundedSize: Its stronger norm (p.20).
  • powerPull: Principal complex powers on native sectors.
  • qsKernel: Kernel in (4.6); the power is on the entire denominator.
  • Qs: The explicit Q_s integral.
  • hardySize: Radial Hardy H1 norm, summing both rays as in (4.8).
  • HardySobolevClass: H_psi membership is independent of any boundary extension theorem.
  • hSize: Full Hardy-Sobolev norm (p.21).
  • NormConverges: A norm-convergent sequence in a domain-relative symbol space.
  • NormCauchy: The Cauchy condition, with no assumed normed-space instance.
  • BanachSymbols: Banach-space properties stated extensionally on the native domain.
  • BanachAlgebraSymbols: A unital Banach algebra of symbols, modulo equality on its domain.
  • NormClosure: Relative norm closure; the ambient space is supplied separately.
  • CompactClosure: Uniform approximation on each compact subset, the compact-open topology.
  • CompactConverges: Uniform convergence on compact sets for a sequence.
  • NotSeparableSymbols: Nonseparability in the norm topology, expressed using every countable family.
  • Vanishes: Source C0 on the closed half-plane, using canonical boundary values.
  • resolventSymbol: Resolvent symbols (p.5).
  • ResolventSpan: Finite sums of resolvents, compared only on the indicated domain.
  • localDerivative: Localized derivative integral in Lemma 3.5.
  • BConvergence: Exactly the three sequence assumptions of Lemma 3.5; no limit membership.
  • arccot: The holomorphic arccot branch (3.2).
  • arccotZero: The continuous right-half-plane arccot value at zero is pi/2.
  • sinhQuotient: Continuous value of sinh(zt)/t at t=0 (Example 3.3).
  • scalarCayley: Cayley symbol.
  • phiSymbol: The regularized inverse-exponential symbol preceding Proposition 3.19.
  • betaValue: Euler's beta integral on its positive-parameter domain.
  • boundaryAverage: The boundary average (4.11), defined through continuous native traces.
  • DenseClosedGraphOperator: Closed densely defined complex-linear operators, represented by their actual single-valued graph. No semigroup or dense-range hypothesis is built in (p.4).
  • InverseAt: A bounded two-sided inverse of z+A, on the actual graph domain.
  • plusResolvent: A concrete resolvent selector, with an irrelevant fallback off the resolvent set.
  • operatorSpectrum: Spectrum of A, with the sign appropriate to the z+A resolvent.
  • SectorialAtAngle: Sectoriality of angle theta uses the CLOSED sector and all larger angles.
  • SectorialBelow: The strict pi/2-minus class.
  • mSize: M_A with its genuine extended supremum (2.2).
  • gsfSlice: Scalar GSF integral for the square resolvent.
  • GSF: Exactly the pointwise GSF assumptions (3.12), without adding a supplied uniform GSF constant or semigroup as a hypothesis.
  • gammaSize: Least uniform GSF constant as an extended supremum over the two unit balls. The factor 2/pi is NOT part of gamma_A in (3.14).
  • weakKernel: Integrand of the actual weak operator formula (3.15).
  • weakBValue: The scalar formula is fixed before any operator is asserted to exist.
  • RealizesB: Representation in X, as opposed to a merely bidual-valued weak formula.
  • bOperator: Canonical value of (3.15) when it is represented by a bounded X-valued operator; existence is a conclusion of Theorem 3.13.
  • AlgebraMapOn: Algebra homomorphism on native-domain symbols, with domain extensionality.
  • BoundedMapOn: Boundedness on a source symbol space, with a uniform constant.
  • BCalculus: A B-calculus as defined on p.4, with its resolvent normalization.
  • C0SemigroupData: Bounded C0-semigroup data, reusing the accepted predicate (source pp.4--5). The subtype only packages its time map with that predicate; it adds no axioms.
  • Generates: Generator graph with the paper's negative-generator convention.
  • HPCompatible: Exact HP agreement for every finite COMPLEX measure, using a positive control measure and a complex density as above.
  • HilbertNorm: Hilbert structure for the SAME complex normed-space structure as the operators. Equality of the inherited NormedSpace also fixes complex scalars.
  • HilbertGenerator: The bounded Hilbert-semigroup regime, without imposing it on Banach clauses.
  • besovOperatorIntegrand: Operator-norm integrand of Theorem 4.1.
  • AbsoluteOperatorDouble: Absolute convergence in operator norm, rather than only weak convergence.
  • TransposeEigen: The eigenfunctional condition for the BANACH transpose: no conjugation.
  • operatorCayley: The concrete Cayley transform I-2(1+A)^(-1).
Definition code
import Mathlib
import Definitions.Def_QAlgorithms_Papers_Independent_Gap48_P07

/-!
Batty--Gomilko--Tomilov, arXiv:2202.03143v1.
All symbol predicates and comparisons are relative to the displayed open domain.
Extended nonnegative norms retain divergence. Values selected by `limitValue`
are used only when the corresponding existence assertion is established by a
stated result; no existence or regularity conclusion is assumed in a definition.
-/

noncomputable section

open Complex MeasureTheory Filter Set
open scoped Topology ENNReal

namespace QAlgorithms.Papers.Independent.Intakenext100.P35

open QAlgorithms.Papers.Independent.Renew46.P11 (HolomorphicBernstein)

attribute [local instance] Classical.propDecidable

abbrev AnalyticSymbol := ℂ → ℂ

/-- The open right half-plane (source p.4). -/
def hp : Set ℂ := {z | 0 < z.re}

/-- The closed right half-plane (source p.4). -/
def closedHp : Set ℂ := {z | 0 ≤ z.re}

/-- The source's open sector, including its special angle-zero convention (p.4). -/
def openSectorDomain (psi : ℝ) : Set ℂ :=
  if psi = 0 then {z | z.im = 0 ∧ 0 < z.re}
  else {z | z ≠ 0 ∧ |Complex.arg z| < psi}

/-- Cartesian parametrization of the right half-plane. -/
def point (a b : ℝ) : ℂ := (a : ℂ) + Complex.I * (b : ℂ)

/-- Lebesgue measure on the positive alpha axis. Endpoints have measure zero. -/
def positiveVolume : Measure ℝ := volume.restrict (Ioi 0)

/-- Holomorphicity on the actual symbol domain. -/
def Hol (S : Set ℂ) (f : AnalyticSymbol) : Prop := DifferentiableOn ℂ f S

/-- Supremum in an extended lattice, never a default real supremum. -/
def supSize (S : Set ℂ) (f : AnalyticSymbol) : ℝ≥0∞ :=
  ⨆ z : S, ENNReal.ofReal ‖f z‖

/-- A bounded holomorphic symbol on its native domain. -/
def HInfinity (S : Set ℂ) (f : AnalyticSymbol) : Prop := Hol S f ∧ supSize S f < ⊤

/-- A guarded limit selector. Its fallback is never asserted to be a limit. -/
def limitValue {D : Type*} (F : Filter D) (f : D → ℂ) : ℂ :=
  if h : ∃ c, Tendsto f F (𝓝 c) then Classical.choose h else 0

/-- The canonical infinity value selected along the positive real axis. -/
def infinityValue (f : AnalyticSymbol) : ℂ :=
  limitValue atTop (fun t : ℝ => f (t : ℂ))

/-- The canonical zero value selected along the positive real axis. -/
def zeroValue (f : AnalyticSymbol) : ℂ :=
  limitValue (𝓝[>] (0 : ℝ)) (fun t : ℝ => f (t : ℂ))

/-- Continuous boundary extension, when it exists; interior values are unchanged. -/
def extension (S : Set ℂ) (f : AnalyticSymbol) (z : ℂ) : ℂ :=
  if z ∈ S then f z else limitValue (𝓝[S] z) f

/-- Norm infinity filter restricted to a native symbol domain. -/
def infinityWithin (S : Set ℂ) : Filter ℂ :=
  Filter.comap (fun z : ℂ => ‖z‖) atTop ⊓ Filter.principal S

/-- Both sectorial limits, with the same values for every strict subsector (p.20). -/
def SectorialLimits (f : AnalyticSymbol) : Prop :=
  ∀ psi : ℝ, 0 < psi → psi < Real.pi / 2 →
    Tendsto f (𝓝[openSectorDomain psi] 0) (𝓝 (zeroValue f)) ∧
    Tendsto f (infinityWithin (openSectorDomain psi)) (𝓝 (infinityValue f))

/-- Besov seminorm, equation (3.1), without a boundedness hypothesis (p.6). -/
def bSemi (f : AnalyticSymbol) : ℝ≥0∞ :=
  ∫⁻ a, (⨆ b : ℝ, ENNReal.ofReal ‖deriv f (point a b)‖) ∂positiveVolume

/-- The space B as it is defined before Proposition 3.1. -/
def BesovClass (f : AnalyticSymbol) : Prop := Hol hp f ∧ bSemi f < ⊤

/-- Full Besov norm (p.7), in an extended lattice off its natural domain. -/
def bSize (f : AnalyticSymbol) : ℝ≥0∞ := supSize hp f + bSemi f

/-- The ideal B0, with vanishing canonical infinity value (p.7). -/
def BZero (f : AnalyticSymbol) : Prop := BesovClass f ∧ infinityValue f = 0

/-- Mixed W norm uses an essential, rather than pointwise, supremum (3.3). -/
def wSize (g : AnalyticSymbol) : ℝ≥0∞ :=
  ∫⁻ a, essSup (fun b : ℝ => ENNReal.ofReal ‖g (point a b)‖)
    volume ∂positiveVolume

/-- Measurable representatives of W. All W assertions below are a.e. invariant. -/
def MixedDerivativeClass (g : AnalyticSymbol) : Prop :=
  AEStronglyMeasurable (fun p : ℝ × ℝ => g (point p.1 p.2))
    (positiveVolume.prod volume) ∧ wSize g < ⊤

/-- Absolute convergence of a double scalar integral, including measurability. -/
def AbsoluteDouble (F : ℝ → ℝ → ℂ) : Prop :=
  AEStronglyMeasurable (fun p : ℝ × ℝ => F p.1 p.2)
    (positiveVolume.prod volume) ∧
  (∫⁻ a, ∫⁻ b, ENNReal.ofReal ‖F a b‖ ∂volume ∂positiveVolume) < ⊤

/-- The reflected resolvent kernel from (3.5), p.8. -/
def qKernel (g : AnalyticSymbol) (z : ℂ) (a b : ℝ) : ℂ :=
  (a : ℂ) * g (point a b) / (point a (-b) + z) ^ 2

/-- Q is computed from its scalar integral; B0 membership is not assumed. -/
def besovPrimitive (g : AnalyticSymbol) (z : ℂ) : ℂ :=
  (-2 / (Real.pi : ℂ)) * ∫ a, ∫ b, qKernel g z a b ∂volume ∂positiveVolume

/-- The derivative dual seminorm from p.11. -/
def eSemi (g : AnalyticSymbol) : ℝ≥0∞ :=
  ⨆ a : {a : ℝ // 0 < a}, ENNReal.ofReal a.1 *
    ∫⁻ b, ENNReal.ofReal ‖deriv g (point a.1 b)‖ ∂volume

/-- The holomorphic space E. -/
def DerivativeDualClass (g : AnalyticSymbol) : Prop := Hol hp g ∧ eSemi g < ⊤

/-- The bilinear (not sesquilinear) partial duality (3.9). -/
def besovPairing (g f : AnalyticSymbol) : ℂ :=
  ∫ a, (a : ℂ) * ∫ b, deriv g (point a (-b)) * deriv f (point a b)
    ∂volume ∂positiveVolume

/-- A positive finite control measure and an integrable complex density represent
an arbitrary finite complex measure; no positivity of the density is imposed. -/
def LaplaceRep (mu : Measure ℝ) (w : ℝ → ℂ) (f : AnalyticSymbol) : Prop :=
  IsFiniteMeasure mu ∧ mu (Iio 0) = 0 ∧ Integrable w mu ∧
    EqOn f (fun z => ∫ t, Complex.exp (-z * (t : ℂ)) * w t ∂mu) hp

/-- The full finite-complex-measure Laplace class LM (p.5). -/
def LM (f : AnalyticSymbol) : Prop := ∃ mu w, LaplaceRep mu w f

/-- HP norm: infimum of total variations over representations. The empty
infimum is infinity, so non-LM symbols cannot have a spurious zero norm. -/
def hpSize (f : AnalyticSymbol) : ℝ≥0∞ :=
  sInf {v | ∃ mu w, LaplaceRep mu w f ∧ v = ∫⁻ t, ENNReal.ofReal ‖w t‖ ∂mu}

/-- Extended mixed weighted norm (4.3); alpha is positive in this integral. -/
def vSize (s : ℝ) (g : AnalyticSymbol) : ℝ≥0∞ :=
  ∫⁻ a, ∫⁻ b, ENNReal.ofReal
    (Real.rpow a s / Real.rpow (a ^ 2 + b ^ 2) ((s + 1) / 2)) *
      ENNReal.ofReal ‖g (point a b)‖ ∂volume ∂positiveVolume

/-- Measurable representatives of V_s. -/
def WeightedDerivativeClass (s : ℝ) (g : AnalyticSymbol) : Prop :=
  AEStronglyMeasurable (fun p : ℝ × ℝ => g (point p.1 p.2))
    (positiveVolume.prod volume) ∧ vSize s g < ⊤

/-- D_s imposes only the source's holomorphic derivative-integrability condition. -/
def D (s : ℝ) (f : AnalyticSymbol) : Prop := Hol hp f ∧ WeightedDerivativeClass s (deriv f)

/-- The D_s seminorm and norm (pp.18--20). -/
def dSemi (s : ℝ) (f : AnalyticSymbol) : ℝ≥0∞ := vSize s (deriv f)

/-- The full D_s norm retains its infinity term separately from the seminorm. -/
def dSize (s : ℝ) (f : AnalyticSymbol) : ℝ≥0∞ :=
  ENNReal.ofReal ‖infinityValue f‖ + dSemi s f

/-- D_infinity is a union, not a boundedness condition. -/
def DUnion (f : AnalyticSymbol) : Prop := ∃ s : ℝ, -1 < s ∧ D s f

/-- Bounded subalgebra D_s^infinity. -/
def DBounded (s : ℝ) (f : AnalyticSymbol) : Prop := D s f ∧ HInfinity hp f

/-- Its stronger norm (p.20). -/
def dBoundedSize (s : ℝ) (f : AnalyticSymbol) : ℝ≥0∞ := supSize hp f + dSemi s f

/-- Principal complex powers on native sectors. -/
def powerPull (r : ℝ) (f : AnalyticSymbol) (z : ℂ) : ℂ := f (z ^ (r : ℂ))

/-- Kernel in (4.6); the power is on the entire denominator. -/
def qsKernel (s : ℝ) (g : AnalyticSymbol) (z : ℂ) (a b : ℝ) : ℂ :=
  (Real.rpow a s : ℂ) * g (point a b) /
    (point a (-b) + z) ^ ((s + 1 : ℝ) : ℂ)

/-- The explicit Q_s integral. -/
def Qs (s : ℝ) (g : AnalyticSymbol) (z : ℂ) : ℂ :=
  (-(Real.rpow 2 s) / Real.pi : ℝ) *
    ∫ a, ∫ b, qsKernel s g z a b ∂volume ∂positiveVolume

/-- Radial Hardy H1 norm, summing both rays as in (4.8). -/
def hardySize (psi : ℝ) (g : AnalyticSymbol) : ℝ≥0∞ :=
  ⨆ phi : {phi : ℝ // |phi| < psi},
    ∫⁻ t, ENNReal.ofReal ‖g ((t : ℂ) * Complex.exp (Complex.I * phi.1))‖ +
      ENNReal.ofReal ‖g ((t : ℂ) * Complex.exp (-Complex.I * phi.1))‖
      ∂positiveVolume

/-- H_psi membership is independent of any boundary extension theorem. -/
def HardySobolevClass (psi : ℝ) (f : AnalyticSymbol) : Prop :=
  Hol (openSectorDomain psi) f ∧ hardySize psi (deriv f) < ⊤

/-- Full Hardy-Sobolev norm (p.21). -/
def hSize (psi : ℝ) (f : AnalyticSymbol) : ℝ≥0∞ :=
  supSize (openSectorDomain psi) f + hardySize psi (deriv f)

/-- A norm-convergent sequence in a domain-relative symbol space. -/
def NormConverges (N : AnalyticSymbol → ℝ≥0∞) (fs : ℕ → AnalyticSymbol) (f : AnalyticSymbol) : Prop :=
  ∀ eps : ℝ, 0 < eps → ∃ k, ∀ n ≥ k, N (fs n - f) < ENNReal.ofReal eps

/-- The Cauchy condition, with no assumed normed-space instance. -/
def NormCauchy (N : AnalyticSymbol → ℝ≥0∞) (fs : ℕ → AnalyticSymbol) : Prop :=
  ∀ eps : ℝ, 0 < eps → ∃ k, ∀ n ≥ k, ∀ m ≥ k,
    N (fs n - fs m) < ENNReal.ofReal eps

/-- Banach-space properties stated extensionally on the native domain. -/
def BanachSymbols (S : Set ℂ) (A : AnalyticSymbol → Prop) (N : AnalyticSymbol → ℝ≥0∞) : Prop :=
  A 0 ∧ (∀ f, A f → N f < ⊤ ∧ (N f = 0 ↔ EqOn f 0 S)) ∧
  (∀ f g, A f → A g → A (f + g) ∧ N (f + g) ≤ N f + N g) ∧
  (∀ c : ℂ, ∀ f, A f → A (c • f) ∧ N (c • f) = ENNReal.ofReal ‖c‖ * N f) ∧
  ∀ fs : ℕ → AnalyticSymbol, (∀ n, A (fs n)) → NormCauchy N fs →
    ∃ f, A f ∧ NormConverges N fs f

/-- A unital Banach algebra of symbols, modulo equality on its domain. -/
def BanachAlgebraSymbols (S : Set ℂ) (A : AnalyticSymbol → Prop)
    (N : AnalyticSymbol → ℝ≥0∞) : Prop :=
  BanachSymbols S A N ∧ A 1 ∧ N 1 = 1 ∧
    ∀ f g, A f → A g → A (f * g) ∧ N (f * g) ≤ N f * N g

/-- Relative norm closure; the ambient space is supplied separately. -/
def NormClosure (A : AnalyticSymbol → Prop) (N : AnalyticSymbol → ℝ≥0∞) (f : AnalyticSymbol) : Prop :=
  ∀ eps : ℝ, 0 < eps → ∃ g, A g ∧ N (g - f) < ENNReal.ofReal eps

/-- Uniform approximation on each compact subset, the compact-open topology. -/
def CompactClosure (S : Set ℂ) (A : AnalyticSymbol → Prop) (f : AnalyticSymbol) : Prop :=
  ∀ K : Set ℂ, K ⊆ S → IsCompact K → ∀ eps : ℝ, 0 < eps →
    ∃ g, A g ∧ supSize K (g - f) < ENNReal.ofReal eps

/-- Uniform convergence on compact sets for a sequence. -/
def CompactConverges (S : Set ℂ) (fs : ℕ → AnalyticSymbol) (f : AnalyticSymbol) : Prop :=
  ∀ K : Set ℂ, K ⊆ S → IsCompact K → ∀ eps : ℝ, 0 < eps →
    ∃ k, ∀ n ≥ k, supSize K (fs n - f) < ENNReal.ofReal eps

/-- Nonseparability in the norm topology, expressed using every countable family. -/
def NotSeparableSymbols (A : AnalyticSymbol → Prop) (N : AnalyticSymbol → ℝ≥0∞) : Prop :=
  ∀ fs : ℕ → AnalyticSymbol, (∀ n, A (fs n)) →
    ∃ f, A f ∧ ∃ eps : ℝ, 0 < eps ∧ ∀ n, ENNReal.ofReal eps ≤ N (fs n - f)

/-- Source C0 on the closed half-plane, using canonical boundary values. -/
def Vanishes (f : AnalyticSymbol) : Prop :=
  Tendsto f (infinityWithin hp) (𝓝 0)

/-- Resolvent symbols (p.5). -/
def resolventSymbol (lam : ℂ) (z : ℂ) : ℂ := (lam + z)⁻¹

/-- Finite sums of resolvents, compared only on the indicated domain. -/
def ResolventSpan (S L : Set ℂ) (constants : Bool) (f : AnalyticSymbol) : Prop :=
  ∃ n : ℕ, ∃ c : Fin n → ℂ, ∃ lam : Fin n → ℂ,
    (∀ j, lam j ∈ L) ∧ ∃ c0 : ℂ, (constants = false → c0 = 0) ∧
      EqOn f (fun z => c0 + ∑ j, c j * resolventSymbol (lam j) z) S

/-- Localized derivative integral in Lemma 3.5. -/
def localDerivative (f : AnalyticSymbol) (r delta : ℝ) : ℝ≥0∞ :=
  ∫⁻ a in Ioo 0 delta, ⨆ b : {b : ℝ // |b| ≤ r},
    ENNReal.ofReal ‖deriv f (point a b.1)‖

/-- Exactly the three sequence assumptions of Lemma 3.5; no limit membership. -/
def BConvergence (fs : ℕ → AnalyticSymbol) (f : AnalyticSymbol) : Prop :=
  (∀ n, BesovClass (fs n)) ∧ (∃ C : ℝ, ∀ n, bSize (fs n) ≤ ENNReal.ofReal C) ∧
  (∀ z ∈ hp, Tendsto (fun n => fs n z) atTop (𝓝 (f z))) ∧
  ∀ r : ℝ, 0 < r → ∀ eps : ℝ, 0 < eps → ∃ d : ℝ, 0 < d ∧
    ∀ delta : ℝ, 0 < delta → delta < d → ∀ n,
      localDerivative (fs n) r delta < ENNReal.ofReal eps

/-- The holomorphic arccot branch (3.2). -/
def arccot (z : ℂ) : ℂ := (2 * Complex.I)⁻¹ *
  Complex.log ((z + Complex.I) / (z - Complex.I))

/-- The continuous right-half-plane arccot value at zero is pi/2. -/
def arccotZero (z : ℂ) : ℂ := if z = 0 then (Real.pi / 2 : ℝ) else arccot z

/-- Continuous value of sinh(zt)/t at t=0 (Example 3.3). -/
def sinhQuotient (z : ℂ) (t : ℝ) : ℂ :=
  if t = 0 then z else Complex.sinh (z * (t : ℂ)) / (t : ℂ)

/-- Cayley symbol. -/
def scalarCayley (z : ℂ) : ℂ := (z - 1) / (z + 1)

/-- The regularized inverse-exponential symbol preceding Proposition 3.19. -/
def phiSymbol (t : ℝ) (z : ℂ) : ℂ := z / (z + 1) * Complex.exp (-(t : ℂ) / z)

/-- Euler's beta integral on its positive-parameter domain. -/
def betaValue (a b : ℝ) : ℝ :=
  ∫ t in Ioo (0 : ℝ) 1, Real.rpow t (a - 1) * Real.rpow (1 - t) (b - 1)

/-- The boundary average (4.11), defined through continuous native traces. -/
def boundaryAverage (psi : ℝ) (f : AnalyticSymbol) (t : ℝ) : ℂ :=
  (extension (openSectorDomain psi) f (Complex.exp (Complex.I * psi) * (t : ℂ)) +
    extension (openSectorDomain psi) f (Complex.exp (-Complex.I * psi) * (t : ℂ))) / 2

/-- Structural sanity checks on source-domain operations and finite sums. -/
example (a b : ℝ) : (point a b).re = a := by simp [point]

example : (0 : ℂ) ∉ hp := by simp [hp]

example : (1 : ℂ) ∈ hp := by simp [hp]

example (S : Set ℂ) (f : AnalyticSymbol) (z : ℂ) (hz : z ∈ S) :
    extension S f z = f z := by simp [extension, hz]

example (z : ℂ) : sinhQuotient z 0 = z := by simp [sinhQuotient]

example : arccotZero 0 = (Real.pi / 2 : ℝ) := by simp [arccotZero]

example : supSize hp (0 : AnalyticSymbol) = 0 := by simp [supSize]

example : bSemi (0 : AnalyticSymbol) = 0 := by simp [bSemi]

example (s : ℝ) : dSemi s (0 : AnalyticSymbol) = 0 := by simp [dSemi, vSize]

example (psi : ℝ) : hardySize psi (0 : AnalyticSymbol) = 0 := by simp [hardySize]

example (s : ℝ) (z : ℂ) : Qs s 0 z = 0 := by simp [Qs, qsKernel]

example (z : ℂ) : besovPrimitive 0 z = 0 := by simp [besovPrimitive, qKernel]

universe u

section Operators

variable {X : Type u} [NormedAddCommGroup X] [NormedSpace ℂ X] [CompleteSpace X]

/-- Closed densely defined complex-linear operators, represented by their actual
single-valued graph. No semigroup or dense-range hypothesis is built in (p.4). -/
structure DenseClosedGraphOperator (X : Type u) [NormedAddCommGroup X] [NormedSpace ℂ X] where
  graph : Submodule ℂ (X × X)
  graphClosed : IsClosed (graph : Set (X × X))
  singleValued : ∀ x y y2, (x, y) ∈ graph → (x, y2) ∈ graph → y = y2
  domainDense : Dense {x | ∃ y, (x, y) ∈ graph}

/-- A bounded two-sided inverse of z+A, on the actual graph domain. -/
def InverseAt (A : DenseClosedGraphOperator X) (z : ℂ) (R : X →L[ℂ] X) : Prop :=
  (∀ x, (R x, x - z • R x) ∈ A.graph) ∧
    ∀ x y, (x, y) ∈ A.graph → R (z • x + y) = x

/-- A concrete resolvent selector, with an irrelevant fallback off the resolvent set. -/
def plusResolvent (A : DenseClosedGraphOperator X) (z : ℂ) : X →L[ℂ] X :=
  if h : ∃ R, InverseAt A z R then Classical.choose h else 0

/-- Spectrum of A, with the sign appropriate to the z+A resolvent. -/
def operatorSpectrum (A : DenseClosedGraphOperator X) : Set ℂ :=
  {z | ¬ ∃ R : X →L[ℂ] X, InverseAt A (-z) R}

/-- Sectoriality of angle theta uses the CLOSED sector and all larger angles. -/
def SectorialAtAngle (theta : ℝ) (A : DenseClosedGraphOperator X) : Prop :=
  0 ≤ theta ∧ theta < Real.pi ∧ operatorSpectrum A ⊆ closure (openSectorDomain theta) ∧
  ∀ psi : ℝ, theta < psi → psi < Real.pi → ∃ C : ℝ, 0 ≤ C ∧
    ∀ z ∈ openSectorDomain (Real.pi - psi), ‖z • plusResolvent A z‖ ≤ C

/-- The strict pi/2-minus class. -/
def SectorialBelow (A : DenseClosedGraphOperator X) : Prop :=
  ∃ theta : ℝ, 0 ≤ theta ∧ theta < Real.pi / 2 ∧ SectorialAtAngle theta A

/-- M_A with its genuine extended supremum (2.2). -/
def mSize (A : DenseClosedGraphOperator X) : ℝ≥0∞ :=
  ⨆ z : hp, ENNReal.ofReal ‖(z : ℂ) • plusResolvent A z‖

/-- Scalar GSF integral for the square resolvent. -/
def gsfSlice (A : DenseClosedGraphOperator X) (x : X) (ell : X →L[ℂ] ℂ)
    (a : ℝ) : ℝ≥0∞ :=
  ∫⁻ b, ENNReal.ofReal ‖ell ((plusResolvent A (point a b))
    ((plusResolvent A (point a b)) x))‖ ∂volume

/-- Exactly the pointwise GSF assumptions (3.12), without adding a supplied
uniform GSF constant or semigroup as a hypothesis. -/
def GSF (A : DenseClosedGraphOperator X) : Prop :=
  operatorSpectrum A ⊆ closedHp ∧
  ∀ x : X, ∀ ell : X →L[ℂ] ℂ,
    (⨆ a : {a : ℝ // 0 < a}, ENNReal.ofReal a.1 * gsfSlice A x ell a.1) < ⊤

/-- Least uniform GSF constant as an extended supremum over the two unit balls.
The factor 2/pi is NOT part of gamma_A in (3.14). -/
def gammaSize (A : DenseClosedGraphOperator X) : ℝ≥0∞ :=
  ⨆ x : {x : X // ‖x‖ ≤ 1}, ⨆ ell : {ell : X →L[ℂ] ℂ // ‖ell‖ ≤ 1},
    ⨆ a : {a : ℝ // 0 < a}, ENNReal.ofReal a.1 * gsfSlice A x.1 ell.1 a.1

/-- Integrand of the actual weak operator formula (3.15). -/
def weakKernel (A : DenseClosedGraphOperator X) (f : AnalyticSymbol) (x : X)
    (ell : X →L[ℂ] ℂ) (a b : ℝ) : ℂ :=
  (a : ℂ) * ell ((plusResolvent A (point a (-b)))
    ((plusResolvent A (point a (-b))) x)) * deriv f (point a b)

/-- The scalar formula is fixed before any operator is asserted to exist. -/
def weakBValue (A : DenseClosedGraphOperator X) (f : AnalyticSymbol) (x : X)
    (ell : X →L[ℂ] ℂ) : ℂ :=
  infinityValue f * ell x - (2 / (Real.pi : ℂ)) *
    ∫ a, ∫ b, weakKernel A f x ell a b ∂volume ∂positiveVolume

/-- Representation in X, as opposed to a merely bidual-valued weak formula. -/
def RealizesB (A : DenseClosedGraphOperator X) (f : AnalyticSymbol) (F : X →L[ℂ] X) : Prop :=
  ∀ x : X, ∀ ell : X →L[ℂ] ℂ,
    AbsoluteDouble (weakKernel A f x ell) ∧ ell (F x) = weakBValue A f x ell

/-- Canonical value of (3.15) when it is represented by a bounded X-valued
operator; existence is a conclusion of Theorem 3.13. -/
def bOperator (A : DenseClosedGraphOperator X) (f : AnalyticSymbol) : X →L[ℂ] X :=
  if h : ∃ F, RealizesB A f F then Classical.choose h else 0

/-- Algebra homomorphism on native-domain symbols, with domain extensionality. -/
def AlgebraMapOn (S : Set ℂ) (F : AnalyticSymbol → X →L[ℂ] X) (A : AnalyticSymbol → Prop) : Prop :=
  F 0 = 0 ∧ F 1 = ContinuousLinearMap.id ℂ X ∧
  (∀ f g, A f → A g → EqOn f g S → F f = F g) ∧
  (∀ f g, A f → A g → F (f + g) = F f + F g) ∧
  (∀ c : ℂ, ∀ f, A f → F (c • f) = c • F f) ∧
  ∀ f g, A f → A g → F (f * g) = (F f).comp (F g)

/-- Boundedness on a source symbol space, with a uniform constant. -/
def BoundedMapOn (F : AnalyticSymbol → X →L[ℂ] X) (A : AnalyticSymbol → Prop)
    (N : AnalyticSymbol → ℝ≥0∞) : Prop :=
  ∃ C : ℝ, 0 < C ∧ ∀ f, A f → ENNReal.ofReal ‖F f‖ ≤ ENNReal.ofReal C * N f

/-- A B-calculus as defined on p.4, with its resolvent normalization. -/
def BCalculus (A : DenseClosedGraphOperator X) (F : AnalyticSymbol → X →L[ℂ] X) : Prop :=
  AlgebraMapOn hp F BesovClass ∧ BoundedMapOn F BesovClass bSize ∧
    ∀ lam ∈ hp, F (resolventSymbol lam) = plusResolvent A lam

/-- Bounded C0-semigroup data, reusing the accepted predicate (source pp.4--5).
The subtype only packages its time map with that predicate; it adds no axioms. -/
abbrev C0SemigroupData (X : Type u) [NormedAddCommGroup X] [NormedSpace ℂ X]
    [CompleteSpace X] :=
  {T : ℝ → X →L[ℂ] X // QAlgorithms.Papers.Independent.Gap48.P07.BoundedC0 T}

example (T : C0SemigroupData X) :
    QAlgorithms.Papers.Independent.Gap48.P07.BoundedC0 T.val := T.property

/-- Generator graph with the paper's negative-generator convention. -/
def Generates (T : C0SemigroupData X) (A : DenseClosedGraphOperator X) : Prop :=
  ∀ x y : X, (x, y) ∈ A.graph ↔
    Tendsto (fun t : ℝ => (t : ℂ)⁻¹ • (T.val t x - x))
      (𝓝[>] 0) (𝓝 (-y))

/-- Exact HP agreement for every finite COMPLEX measure, using a positive
control measure and a complex density as above. -/
def HPCompatible (T : C0SemigroupData X) (F : AnalyticSymbol → X →L[ℂ] X) : Prop :=
  ∀ mu w f, LaplaceRep mu w f → ∀ x : X,
    Integrable (fun t => w t • T.val t x) mu ∧
      F f x = ∫ t, w t • T.val t x ∂mu

/-- Hilbert structure for the SAME complex normed-space structure as the
operators. Equality of the inherited NormedSpace also fixes complex scalars. -/
def HilbertNorm (X : Type u) [NormedAddCommGroup X] [NormedSpace ℂ X] : Prop :=
  ∃ h : InnerProductSpace ℂ X,
    h.toNormedSpace = (inferInstance : NormedSpace ℂ X)

example {Y : Type u} [NormedAddCommGroup Y] [InnerProductSpace ℂ Y] :
    HilbertNorm Y := by
  exact ⟨inferInstance, rfl⟩

/-- The bounded Hilbert-semigroup regime, without imposing it on Banach clauses. -/
def HilbertGenerator (A : DenseClosedGraphOperator X) : Prop :=
  HilbertNorm X ∧ ∃ T : C0SemigroupData X, Generates T A

/-- Operator-norm integrand of Theorem 4.1. -/
def besovOperatorIntegrand (A : DenseClosedGraphOperator X) (f : AnalyticSymbol) (a b : ℝ) : X →L[ℂ] X :=
  ((a : ℂ) * deriv f (point a b)) •
    (plusResolvent A (point a (-b))).comp (plusResolvent A (point a (-b)))

/-- Absolute convergence in operator norm, rather than only weak convergence. -/
def AbsoluteOperatorDouble (F : ℝ → ℝ → X →L[ℂ] X) : Prop :=
  AEStronglyMeasurable (fun p : ℝ × ℝ => F p.1 p.2)
    (positiveVolume.prod volume) ∧
  (∫⁻ a, ∫⁻ b, ENNReal.ofReal ‖F a b‖ ∂volume ∂positiveVolume) < ⊤

/-- The eigenfunctional condition for the BANACH transpose: no conjugation. -/
def TransposeEigen (A : DenseClosedGraphOperator X) (lam : ℂ) (ell : X →L[ℂ] ℂ) : Prop :=
  ∀ x y : X, (x, y) ∈ A.graph → ell y = lam * ell x

/-- The concrete Cayley transform I-2(1+A)^(-1). -/
def operatorCayley (A : DenseClosedGraphOperator X) : X →L[ℂ] X :=
  ContinuousLinearMap.id ℂ X - (2 : ℂ) • plusResolvent A 1

example (A : DenseClosedGraphOperator X) (z : ℂ) (h : ¬ ∃ R, InverseAt A z R) :
    plusResolvent A z = 0 := by simp [plusResolvent, h]

example (A : DenseClosedGraphOperator X) (f : AnalyticSymbol) (x : X) :
    weakKernel A f x (0 : X →L[ℂ] ℂ) = 0 := by
  funext a b
  simp [weakKernel]

end Operators

/- BEGIN PAPER STATEMENTS -/

end QAlgorithms.Papers.Independent.Intakenext100.P35
Source
Bounded functional calculi for unbounded operators, arXiv:2202.03143v1, https://arxiv.org/abs/2202.03143v1
Read-back

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

QAlgorithms.Papers.Independent.Intakenext100.P35.AbsoluteDouble

For any function F:R→R→CF:\mathbb R\to\mathbb R\to\mathbb CF:R→R→C, this property means that (a,b)↦F(a,b)(a,b)\mapsto F(a,b)(a,b)↦F(a,b) is almost everywhere strongly measurable with respect to the product of Lebesgue measure restricted to a>0a>0a>0 and Lebesgue measure in bbb, and that

∫(0,∞) ⁣−∫R ⁣−∥F(a,b)∥ db da<∞.\int_{(0,\infty)}^{\!-} \int_{\mathbb R}^{\!-}\|F(a,b)\|\,db\,da<\infty.∫(0,∞)−​∫R−​∥F(a,b)∥dbda<∞.

Both integrals are nonnegative Lebesgue integrals, with the norm embedded into the extended nonnegative reals; values with a≤0a\le0a≤0 are ignored.

QAlgorithms.Papers.Independent.Intakenext100.P35.AbsoluteOperatorDouble

For a complex normed space XXX and any function F:R×R→L(X)F:\mathbb R\times\mathbb R\to\mathcal L(X)F:R×R→L(X), this property says that FFF is almost everywhere strongly measurable for the product of Lebesgue measure restricted to (0,∞)(0,\infty)(0,∞) in the first coordinate and Lebesgue measure in the second coordinate, and that

∫(0,∞)∫R∥F(a,b)∥ db da<∞.\int_{(0,\infty)}\int_{\mathbb R}\|F(a,b)\|\,db\,da<\infty.∫(0,∞)​∫R​∥F(a,b)∥dbda<∞.

The integrals are nonnegative extended Lebesgue integrals of the operator norm. Values with a≤0a\le0a≤0 do not affect either condition.

QAlgorithms.Papers.Independent.Intakenext100.P35.AlgebraMapOn

Let XXX be a complex Banach space, let S⊆CS\subseteq\mathbb CS⊆C, let FFF assign a bounded complex linear operator on XXX to every function C→C\mathbb C\to\mathbb CC→C, and let AAA be any predicate on these functions. The assertion AlgebraMapOn⁡(S,F,A)\operatorname{AlgebraMapOn}(S,F,A)AlgebraMapOn(S,F,A) means that F(0)=0F(0)=0F(0)=0 and F(1)=IXF(1)=I_XF(1)=IX​, where 0,10,10,1 denote the constant functions; whenever A(f)A(f)A(f) and A(g)A(g)A(g), agreement f(z)=g(z)f(z)=g(z)f(z)=g(z) for every z∈Sz\in Sz∈S implies F(f)=F(g)F(f)=F(g)F(f)=F(g), and

F(f+g)=F(f)+F(g),F(fg)=F(f)∘F(g);F(f+g)=F(f)+F(g),\qquad F(fg)=F(f)\circ F(g);F(f+g)=F(f)+F(g),F(fg)=F(f)∘F(g);

and, for every c∈Cc\in\mathbb Cc∈C and every fff satisfying A(f)A(f)A(f), F(cf)=cF(f)F(cf)=cF(f)F(cf)=cF(f). Function operations are pointwise. No closure properties or membership of 0,10,10,1 are required of AAA. The set SSS may be empty, in which case agreement on SSS is automatic.

QAlgorithms.Papers.Independent.Intakenext100.P35.AnalyticSymbol

An analytic symbol is any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C; no regularity condition is imposed.

QAlgorithms.Papers.Independent.Intakenext100.P35.BCalculus

Let XXX be a complex Banach space, let AAA have a closed complex linear, single-valued graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain, and let FFF assign a bounded complex linear operator on XXX to every function C→C\mathbb C\to\mathbb CC→C. Put H={z:Re⁡z>0}H=\{z:\operatorname{Re}z>0\}H={z:Rez>0}, and let B\mathcal BB consist of all functions holomorphic on HHH for which

β(f)=∫0∞−sup⁡b∈R∣f′(a+ib)∣ da<∞;B(f)=sup⁡z∈H∣f(z)∣+β(f),\beta(f)=\int_0^\infty{}^{-}\sup_{b\in\mathbb R}|f'(a+ib)|\,da<\infty; \qquad B(f)=\sup_{z\in H}|f(z)|+\beta(f),β(f)=∫0∞​−b∈Rsup​∣f′(a+ib)∣da<∞;B(f)=z∈Hsup​∣f(z)∣+β(f),

with these quantities evaluated in [0,∞][0,\infty][0,∞]. Then BCalculus⁡(A,F)\operatorname{BCalculus}(A,F)BCalculus(A,F) means that F(0)=0F(0)=0F(0)=0, F(1)=IXF(1)=I_XF(1)=IX​, and, for all f,g∈Bf,g\in\mathcal Bf,g∈B, agreement on HHH implies equality of their assigned operators, F(f+g)=F(f)+F(g)F(f+g)=F(f)+F(g)F(f+g)=F(f)+F(g), and F(fg)=F(f)∘F(g)F(fg)=F(f)\circ F(g)F(fg)=F(f)∘F(g); also F(cf)=cF(f)F(cf)=cF(f)F(cf)=cF(f) for every c∈Cc\in\mathbb Cc∈C and f∈Bf\in\mathcal Bf∈B. Furthermore, there is a real C>0C>0C>0 such that ∥F(f)∥≤CB(f)\|F(f)\|\le C B(f)∥F(f)∥≤CB(f) in [0,∞][0,\infty][0,∞] for every f∈Bf\in\mathcal Bf∈B, and, for every λ∈H\lambda\in Hλ∈H, F(z↦(λ+z)−1)=RA(λ)F(z\mapsto(\lambda+z)^{-1})=R_A(\lambda)F(z↦(λ+z)−1)=RA​(λ). Here RA(λ)R_A(\lambda)RA​(λ) is a classically chosen bounded complex linear map RRR satisfying (Ru,u−λRu)∈GA(Ru,u-\lambda Ru)\in G_A(Ru,u−λRu)∈GA​ for every uuu and R(λu+v)=uR(\lambda u+v)=uR(λu+v)=u whenever (u,v)∈GA(u,v)\in G_A(u,v)∈GA​, or zero if no such map exists. The reciprocal symbol is defined on all of C\mathbb CC, with value zero at z=−λz=-\lambdaz=−λ.

QAlgorithms.Papers.Independent.Intakenext100.P35.BConvergence

For a sequence of arbitrary functions fn:C→Cf_n:\mathbb C\to\mathbb Cfn​:C→C, indexed by n∈Nn\in\mathbb Nn∈N, and a function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, write H={z:Re⁡z>0}H=\{z:\operatorname{Re}z>0\}H={z:Rez>0}, J(h)=∫(0,∞)−sup⁡b∈R∥h′(a+ib)∥ daJ(h)=\int_{(0,\infty)}^{-}\sup_{b\in\mathbb R}\lVert h'(a+ib)\rVert\,daJ(h)=∫(0,∞)−​supb∈R​∥h′(a+ib)∥da, and B(h)=sup⁡z∈H∥h(z)∥+J(h)B(h)=\sup_{z\in H}\lVert h(z)\rVert+J(h)B(h)=supz∈H​∥h(z)∥+J(h), with all suprema and nonnegative Lebesgue integrals valued in [0,∞][0,\infty][0,∞]. This property means that every fnf_nfn​ is complex differentiable on HHH and has J(fn)<∞J(f_n)<\inftyJ(fn​)<∞; there exists a real CCC such that B(fn)≤max⁡(C,0)B(f_n)\le\max(C,0)B(fn​)≤max(C,0) for every nnn; for every z∈Hz\in Hz∈H, fn(z)→f(z)f_n(z)\to f(z)fn​(z)→f(z) as n→∞n\to\inftyn→∞; and for every real r>0r>0r>0 and ε>0\varepsilon>0ε>0, there exists a real d>0d>0d>0 such that, for every real 0<δ<d0<\delta<d0<δ<d and every nnn, ∫(0,δ)−sup⁡∣b∣≤r∥fn′(a+ib)∥ da<ε\int_{(0,\delta)}^{-}\sup_{|b|\le r}\lVert f_n'(a+ib)\rVert\,da<\varepsilon∫(0,δ)−​sup∣b∣≤r​∥fn′​(a+ib)∥da<ε. The same ddd must work for all nnn. The definition places no separate differentiability or size hypothesis on fff.

QAlgorithms.Papers.Independent.Intakenext100.P35.BZero

For a function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this property means that fff is complex differentiable on {z:Re⁡z>0}\{z:\operatorname{Re}z>0\}{z:Rez>0}, that

∫(0,∞) ⁣−sup⁡b∈R∥f′(a+ib)∥ da<∞,\int_{(0,\infty)}^{\!-} \sup_{b\in\mathbb R}\|f'(a+ib)\|\,da<\infty,∫(0,∞)−​b∈Rsup​∥f′(a+ib)∥da<∞,

and that the chosen limit of f(t)f(t)f(t) as real t→+∞t\to+\inftyt→+∞, with the value defined to be 000 when no such limit exists, equals 000. Thus the last condition, considered by itself, allows either a positive-real-axis limit equal to 000 or the absence of such a limit.

QAlgorithms.Papers.Independent.Intakenext100.P35.BanachAlgebraSymbols

For any set S⊆CS\subseteq\mathbb CS⊆C, any predicate AAA on functions C→C\mathbb C\to\mathbb CC→C, and any function N:(C→C)→[0,∞]N:(\mathbb C\to\mathbb C)\to[0,\infty]N:(C→C)→[0,∞], this predicate means that the zero function satisfies AAA; every fff satisfying AAA has finite N(f)N(f)N(f), with N(f)=0N(f)=0N(f)=0 exactly when fff vanishes everywhere on SSS; functions satisfying AAA are closed under pointwise addition and multiplication by every complex scalar, with N(f+g)≤N(f)+N(g)N(f+g)\le N(f)+N(g)N(f+g)≤N(f)+N(g) and N(cf)=∣c∣N(f)N(cf)=|c|N(f)N(cf)=∣c∣N(f); and every sequence (fn)n∈N(f_n)_{n\in\mathbb N}(fn​)n∈N​ satisfying AAA termwise and satisfying, for every real ε>0\varepsilon>0ε>0, N(fn−fm)<εN(f_n-f_m)<\varepsilonN(fn​−fm​)<ε for all sufficiently large n,mn,mn,m, has some limit fff satisfying AAA for which N(fn−f)<εN(f_n-f)<\varepsilonN(fn​−f)<ε for all sufficiently large nnn, for each real ε>0\varepsilon>0ε>0. In addition, the constant function 111 satisfies AAA and N(1)=1N(1)=1N(1)=1, and for every f,gf,gf,g satisfying AAA, their pointwise product satisfies AAA and N(fg)≤N(f)N(g)N(fg)\le N(f)N(g)N(fg)≤N(f)N(g). All size values and arithmetic are in [0,∞][0,\infty][0,∞]. These requirements cannot hold when SSS is empty, since the zero-value equivalence would then force N(1)=0N(1)=0N(1)=0.

QAlgorithms.Papers.Independent.Intakenext100.P35.BanachSymbols

For any set S⊆CS\subseteq\mathbb CS⊆C, any predicate AAA on functions C→C\mathbb C\to\mathbb CC→C, and any function N:(C→C)→[0,∞]N:(\mathbb C\to\mathbb C)\to[0,\infty]N:(C→C)→[0,∞], this predicate asserts all of the following: the zero function satisfies AAA; every function fff satisfying AAA has N(f)<∞N(f)<\inftyN(f)<∞, and N(f)=0N(f)=0N(f)=0 if and only if f(z)=0f(z)=0f(z)=0 for every z∈Sz\in Sz∈S; whenever f,gf,gf,g satisfy AAA, their pointwise sum satisfies AAA and N(f+g)≤N(f)+N(g)N(f+g)\le N(f)+N(g)N(f+g)≤N(f)+N(g); for every c∈Cc\in\mathbb Cc∈C and every fff satisfying AAA, the pointwise scalar multiple cfcfcf satisfies AAA and N(cf)=∣c∣N(f)N(cf)=|c|N(f)N(cf)=∣c∣N(f), including c=0c=0c=0; and for every sequence (fn)n∈N(f_n)_{n\in\mathbb N}(fn​)n∈N​ satisfying AAA termwise, if for every real ε>0\varepsilon>0ε>0 there is k∈Nk\in\mathbb Nk∈N such that N(fn−fm)<εN(f_n-f_m)<\varepsilonN(fn​−fm​)<ε for all n,m≥kn,m\ge kn,m≥k, then there exists a function fff satisfying AAA such that for every real ε>0\varepsilon>0ε>0 there is k∈Nk\in\mathbb Nk∈N with N(fn−f)<εN(f_n-f)<\varepsilonN(fn​−f)<ε for all n≥kn\ge kn≥k. All nonnegative real quantities are interpreted in [0,∞][0,\infty][0,∞], and the asserted limit need not be unique as a function on all of C\mathbb CC. The set SSS may be empty, in which case the zero-value equivalence requires N(f)=0N(f)=0N(f)=0 for every fff satisfying AAA.

QAlgorithms.Papers.Independent.Intakenext100.P35.BesovClass

For a function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this property means that fff is complex differentiable on the open right half-plane {z:Re⁡z>0}\{z:\operatorname{Re}z>0\}{z:Rez>0} and

∫(0,∞) ⁣−sup⁡b∈R∥f′(a+ib)∥ da<∞,\int_{(0,\infty)}^{\!-} \sup_{b\in\mathbb R}\|f'(a+ib)\|\,da<\infty,∫(0,∞)−​b∈Rsup​∥f′(a+ib)∥da<∞,

where the integral is the nonnegative Lebesgue integral and the supremum is taken in the extended nonnegative reals. The property imposes no condition on the values of fff outside the open right half-plane.

QAlgorithms.Papers.Independent.Intakenext100.P35.BoundedMapOn

Let XXX be a complex Banach space, let FFF assign a bounded complex linear operator on XXX to every function C→C\mathbb C\to\mathbb CC→C, let AAA be any predicate on such functions, and let NNN assign each such function a value in [0,∞][0,\infty][0,∞]. The assertion BoundedMapOn⁡(F,A,N)\operatorname{BoundedMapOn}(F,A,N)BoundedMapOn(F,A,N) means that there exists a real constant C>0C>0C>0 such that, for every f:C→Cf:\mathbb C\to\mathbb Cf:C→C satisfying A(f)A(f)A(f),

∥F(f)∥≤C N(f),\|F(f)\|\le C\,N(f),∥F(f)∥≤CN(f),

where the operator norm on the left and CCC are embedded in [0,∞][0,\infty][0,∞], and multiplication is extended nonnegative real multiplication. No norm axioms or finiteness assumptions are imposed on NNN; N(f)=+∞N(f)=+\inftyN(f)=+∞ is permitted, and if no function satisfies AAA, the bound is vacuous.

QAlgorithms.Papers.Independent.Intakenext100.P35.C0SemigroupData

For any type XXX equipped with a complete normed additive commutative group structure and a complex normed vector space structure, C0SemigroupData⁡(X)\operatorname{C0SemigroupData}(X)C0SemigroupData(X) is the type of families T:R→LC(X)T:\mathbb R\to\mathcal L_{\mathbb C}(X)T:R→LC​(X) of bounded complex linear operators satisfying T(0)=IXT(0)=I_XT(0)=IX​, T(s+t)=T(s)∘T(t)T(s+t)=T(s)\circ T(t)T(s+t)=T(s)∘T(t) for every strictly positive s,ts,ts,t, and ∥T(t)∥≤M\|T(t)\|\le M∥T(t)∥≤M for every t≥0t\ge0t≥0 for some real M≥0M\ge0M≥0, such that t↦T(t)xt\mapsto T(t)xt↦T(t)x is continuous on (0,∞)(0,\infty)(0,∞) and on [0,∞)[0,\infty)[0,∞) for every x∈Xx\in Xx∈X. Each element consists of the family together with evidence of these properties. The family is defined at negative times, but these conditions impose no requirements there.

QAlgorithms.Papers.Independent.Intakenext100.P35.CompactClosure

For any set S⊆CS\subseteq\mathbb CS⊆C, predicate AAA on functions C→C\mathbb C\to\mathbb CC→C, and function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this property means that for every compact set K⊆SK\subseteq SK⊆S and every real ε>0\varepsilon>0ε>0, there exists a function g:C→Cg:\mathbb C\to\mathbb Cg:C→C satisfying A(g)A(g)A(g) such that sup⁡z∈K∥g(z)−f(z)∥<ε\sup_{z\in K}\lVert g(z)-f(z)\rVert<\varepsilonsupz∈K​∥g(z)−f(z)∥<ε, with the supremum taken in [0,∞][0,\infty][0,∞]. The approximant may depend on both KKK and ε\varepsilonε. For K=∅K=\varnothingK=∅, this supremum is zero, but an approximant satisfying AAA must still exist; thus even when S=∅S=\varnothingS=∅, the property requires that AAA have a member.

QAlgorithms.Papers.Independent.Intakenext100.P35.CompactConverges

For any set S⊆CS\subseteq\mathbb CS⊆C, sequence of functions fn:C→Cf_n:\mathbb C\to\mathbb Cfn​:C→C indexed by n∈Nn\in\mathbb Nn∈N, and function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this property means that for every compact set K⊆SK\subseteq SK⊆S and every real ε>0\varepsilon>0ε>0, there exists k∈Nk\in\mathbb Nk∈N such that sup⁡z∈K∥fn(z)−f(z)∥<ε\sup_{z\in K}\lVert f_n(z)-f(z)\rVert<\varepsilonsupz∈K​∥fn​(z)−f(z)∥<ε for every n≥kn\ge kn≥k. The supremum is taken in [0,∞][0,\infty][0,∞], with value zero on the empty set. In particular, when S=∅S=\varnothingS=∅, the property holds for every sequence and every fff.

QAlgorithms.Papers.Independent.Intakenext100.P35.D

For every real number sss and every function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this predicate says that fff is complex differentiable throughout the open right half-plane, that (a,b)↦f′(a+ib)(a,b)\mapsto f'(a+ib)(a,b)↦f′(a+ib) is almost everywhere strongly measurable for the product of Lebesgue measure restricted to (0,∞)(0,\infty)(0,∞) and Lebesgue measure on R\mathbb RR, and that

∫(0,∞)∫Ras(a2+b2)(s+1)/2 ∣f′(a+ib)∣ db da<+∞.\int_{(0,\infty)}\int_{\mathbb R} \frac{a^{s}}{(a^{2}+b^{2})^{(s+1)/2}}\,|f'(a+ib)|\,db\,da<+\infty.∫(0,∞)​∫R​(a2+b2)(s+1)/2as​∣f′(a+ib)∣dbda<+∞.

The integrals are nonnegative Lebesgue integrals and the powers are real powers. No restriction on sss or on the values of fff outside that half-plane is included.

QAlgorithms.Papers.Independent.Intakenext100.P35.DBounded

For any real number sss and any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this predicate means that fff is complex differentiable on the open right half-plane H={z:Re⁡z>0}\mathbb H=\{z:\operatorname{Re}z>0\}H={z:Rez>0}, the function (a,b)↦f′(a+ib)(a,b)\mapsto f'(a+ib)(a,b)↦f′(a+ib) is almost everywhere strongly measurable for the product of Lebesgue measure restricted to (0,∞)(0,\infty)(0,∞) and Lebesgue measure on R\mathbb RR, and both

∫(0,∞)∫Ras(a2+b2)(s+1)/2 ∣f′(a+ib)∣ db da<∞,sup⁡z∈H∣f(z)∣<∞.\int_{(0,\infty)}\int_{\mathbb R} \frac{a^s}{(a^2+b^2)^{(s+1)/2}}\,|f'(a+ib)|\,db\,da<\infty, \qquad \sup_{z\in\mathbb H}|f(z)|<\infty.∫(0,∞)​∫R​(a2+b2)(s+1)/2as​∣f′(a+ib)∣dbda<∞,z∈Hsup​∣f(z)∣<∞.

The integral and supremum here are taken in [0,∞][0,\infty][0,∞]. There is no restriction on sss, in particular no assumption that s>−1s>-1s>−1.

QAlgorithms.Papers.Independent.Intakenext100.P35.DUnion

For any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this predicate says that there exists a real number s>−1s>-1s>−1 such that fff is complex differentiable throughout the open right half-plane, the map (a,b)↦f′(a+ib)(a,b)\mapsto f'(a+ib)(a,b)↦f′(a+ib) is almost everywhere strongly measurable for the product of Lebesgue measure restricted to (0,∞)(0,\infty)(0,∞) and Lebesgue measure on R\mathbb RR, and

∫(0,∞)∫Ras(a2+b2)(s+1)/2 ∣f′(a+ib)∣ db da<+∞.\int_{(0,\infty)}\int_{\mathbb R} \frac{a^{s}}{(a^{2}+b^{2})^{(s+1)/2}}\,|f'(a+ib)|\,db\,da<+\infty.∫(0,∞)​∫R​(a2+b2)(s+1)/2as​∣f′(a+ib)∣dbda<+∞.

The powers are real powers and the integrals are nonnegative Lebesgue integrals. The exponent need not be unique, s=−1s=-1s=−1 is excluded, and no condition is imposed on fff outside the open right half-plane.

QAlgorithms.Papers.Independent.Intakenext100.P35.DenseClosedGraphOperator

For any type XXX equipped with a normed additive commutative group structure and a complex normed vector space structure, this structure consists of a complex linear subspace G⊆X×XG\subseteq X\times XG⊆X×X, together with the requirements that GGG is closed in the product topology, that

(x,y)∈G and (x,y′)∈G⟹y=y′(x,y)\in G\ \text{and}\ (x,y')\in G\quad\Longrightarrow\quad y=y'(x,y)∈G and (x,y′)∈G⟹y=y′

for all x,y,y′∈Xx,y,y'\in Xx,y,y′∈X, and that its domain {x∈X:∃y∈X, (x,y)∈G}\{x\in X:\exists y\in X,\ (x,y)\in G\}{x∈X:∃y∈X, (x,y)∈G} is dense in XXX. Being a linear subspace includes containing (0,0)(0,0)(0,0) and being closed under addition and complex scalar multiplication. Completeness of XXX is not a field or requirement of this structure.

QAlgorithms.Papers.Independent.Intakenext100.P35.DerivativeDualClass

For any function g:C→Cg:\mathbb C\to\mathbb Cg:C→C, this predicate says that ggg is complex differentiable at every point of the open right half-plane {z:Re⁡z>0}\{z:\operatorname{Re}z>0\}{z:Rez>0} and that

sup⁡a>0a∫R∣g′(a+ib)∣ db<+∞,\sup_{a>0}a\int_{\mathbb R}|g'(a+ib)|\,db<+\infty,a>0sup​a∫R​∣g′(a+ib)∣db<+∞,

where the integral is the nonnegative Lebesgue integral. It imposes no condition on the values of ggg outside that half-plane.

QAlgorithms.Papers.Independent.Intakenext100.P35.GSF

Let XXX be a complex Banach space, and let AAA have a closed complex linear graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X, with dense domain and at most one output for each input. For z∈Cz\in\mathbb Cz∈C, call a bounded complex linear map R:X→XR:X\to XR:X→X an inverse at zzz if (Ru,u−zRu)∈GA(Ru,u-zRu)\in G_A(Ru,u−zRu)∈GA​ for every u∈Xu\in Xu∈X, and R(zx+y)=xR(zx+y)=xR(zx+y)=x whenever (x,y)∈GA(x,y)\in G_A(x,y)∈GA​. Write RA(z)R_A(z)RA​(z) for a classically chosen such map if one exists, and for the zero map otherwise. Define σA={z∈C:no inverse at −z exists}\sigma_A=\{z\in\mathbb C:\text{no inverse at }-z\text{ exists}\}σA​={z∈C:no inverse at −z exists}. Then GSF⁡(A)\operatorname{GSF}(A)GSF(A) means that σA⊆{z:Re⁡z≥0}\sigma_A\subseteq\{z:\operatorname{Re}z\ge0\}σA​⊆{z:Rez≥0} and that, for every x∈Xx\in Xx∈X and every bounded complex linear functional ℓ:X→C\ell:X\to\mathbb Cℓ:X→C,

sup⁡a>0a∫R−∣ℓ ⁣(RA(a+ib)(RA(a+ib)x))∣ db<∞.\sup_{a>0}a\int_{\mathbb R}^{-} \left|\ell\!\left(R_A(a+ib)\bigl(R_A(a+ib)x\bigr)\right)\right|\,db<\infty.a>0sup​a∫R−​​ℓ(RA​(a+ib)(RA​(a+ib)x))​db<∞.

Here the integral and supremum take values in [0,∞][0,\infty][0,∞]. Finiteness is required separately for every x,ℓx,\ellx,ℓ; no common bound is part of this definition.

QAlgorithms.Papers.Independent.Intakenext100.P35.Generates

Let XXX be a complex Banach space, let T:R→LC(X)T:\mathbb R\to\mathcal L_{\mathbb C}(X)T:R→LC​(X) satisfy T(0)=IXT(0)=I_XT(0)=IX​, T(s+t)=T(s)∘T(t)T(s+t)=T(s)\circ T(t)T(s+t)=T(s)∘T(t) for strictly positive s,ts,ts,t, a common bound ∥T(t)∥≤M\|T(t)\|\le M∥T(t)∥≤M for all t≥0t\ge0t≥0 with some M≥0M\ge0M≥0, and continuity of t↦T(t)xt\mapsto T(t)xt↦T(t)x on [0,∞)[0,\infty)[0,∞) for every x∈Xx\in Xx∈X. Let AAA have a closed complex linear, single-valued graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain. The assertion Generates⁡(T,A)\operatorname{Generates}(T,A)Generates(T,A) means that, for every x,y∈Xx,y\in Xx,y∈X,

(x,y)∈GA⟺lim⁡t↓0T(t)x−xt=−y(x,y)\in G_A \quad\Longleftrightarrow\quad \lim_{t\downarrow0}\frac{T(t)x-x}{t}=-y(x,y)∈GA​⟺t↓0lim​tT(t)x−x​=−y

in the norm topology of XXX. The limit is taken through strictly positive real times, so it does not use the value of the quotient at zero or values of TTT at negative times.

QAlgorithms.Papers.Independent.Intakenext100.P35.HInfinity

For any set S⊆CS\subseteq\mathbb CS⊆C and any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, HInfinity(S,f)\mathrm{HInfinity}(S,f)HInfinity(S,f) means that fff is complex differentiable within SSS at every point of SSS and that sup⁡z∈S∥f(z)∥<∞\sup_{z\in S}\|f(z)\|<\inftysupz∈S​∥f(z)∥<∞, with the supremum taken in [0,∞][0,\infty][0,∞]. No openness assumption on SSS is imposed; every function satisfies this condition when SSS is empty.

QAlgorithms.Papers.Independent.Intakenext100.P35.HPCompatible

Let XXX be a complex Banach space, let T:R→LC(X)T:\mathbb R\to\mathcal L_{\mathbb C}(X)T:R→LC​(X) satisfy T(0)=IXT(0)=I_XT(0)=IX​, T(s+t)=T(s)∘T(t)T(s+t)=T(s)\circ T(t)T(s+t)=T(s)∘T(t) for strictly positive s,ts,ts,t, a common bound ∥T(t)∥≤M\|T(t)\|\le M∥T(t)∥≤M for t≥0t\ge0t≥0 with some M≥0M\ge0M≥0, and continuity of t↦T(t)xt\mapsto T(t)xt↦T(t)x on [0,∞)[0,\infty)[0,∞) for every x∈Xx\in Xx∈X. Let FFF assign a bounded complex linear operator on XXX to every function C→C\mathbb C\to\mathbb CC→C. The assertion HPCompatible⁡(T,F)\operatorname{HPCompatible}(T,F)HPCompatible(T,F) means that, for every finite measure μ\muμ on R\mathbb RR with μ((−∞,0))=0\mu((-\infty,0))=0μ((−∞,0))=0, every μ\muμ-integrable function w:R→Cw:\mathbb R\to\mathbb Cw:R→C, and every f:C→Cf:\mathbb C\to\mathbb Cf:C→C satisfying

f(z)=∫Re−ztw(t) dμ(t)for every z with Re⁡z>0,f(z)=\int_{\mathbb R}e^{-zt}w(t)\,d\mu(t) \quad\text{for every }z\text{ with }\operatorname{Re}z>0,f(z)=∫R​e−ztw(t)dμ(t)for every z with Rez>0,

the function t↦w(t)T(t)xt\mapsto w(t)T(t)xt↦w(t)T(t)x is Bochner integrable for every x∈Xx\in Xx∈X, and

F(f)x=∫Rw(t)T(t)x dμ(t).F(f)x=\int_{\mathbb R}w(t)T(t)x\,d\mu(t).F(f)x=∫R​w(t)T(t)xdμ(t).

This requirement applies to every such representation, including the zero measure, and imposes no condition on the values of fff outside the open right half-plane.

QAlgorithms.Papers.Independent.Intakenext100.P35.HardySobolevClass

For every real number ψ\psiψ and every function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this predicate means that fff is complex differentiable within the set

Sψ={{z:Im⁡z=0, Re⁡z>0},ψ=0,{z:z≠0, ∣arg⁡z∣<ψ},ψ≠0,S_\psi= \begin{cases} \{z:\operatorname{Im}z=0,\ \operatorname{Re}z>0\},&\psi=0,\\ \{z:z\ne0,\ |\arg z|<\psi\},&\psi\ne0, \end{cases}Sψ​={{z:Imz=0, Rez>0},{z:z=0, ∣argz∣<ψ},​ψ=0,ψ=0,​

and that

sup⁡ϕ∈R∣ϕ∣<ψ∫(0,∞)(∣f′(teiϕ)∣+∣f′(te−iϕ)∣) dt<∞.\sup_{\substack{\phi\in\mathbb R\\|\phi|<\psi}} \int_{(0,\infty)} \left( |f'(te^{i\phi})|+|f'(te^{-i\phi})| \right)\,dt<\infty.ϕ∈R∣ϕ∣<ψ​sup​∫(0,∞)​(∣f′(teiϕ)∣+∣f′(te−iϕ)∣)dt<∞.

The integral is a nonnegative Lebesgue integral, and f′f'f′ is the total complex derivative, zero where complex differentiability fails. No range for ψ\psiψ is assumed: for ψ<0\psi<0ψ<0 both the differentiability set and angle set are empty, so the predicate holds for every fff; for ψ=0\psi=0ψ=0 the integral condition is automatic and only differentiability within the positive real ray remains.

QAlgorithms.Papers.Independent.Intakenext100.P35.HilbertGenerator

Let XXX be a complex Banach space, and let AAA be a complex linear subspace of X×XX\times XX×X that is closed, single valued, and has dense domain {x:∃y, (x,y)∈A}\{x:\exists y,\ (x,y)\in A\}{x:∃y, (x,y)∈A}. This property says that XXX admits a complex inner product space structure inducing its specified complex normed space structure, and that there exists a family T:R→L(X)T:\mathbb R\to\mathcal L(X)T:R→L(X) of bounded complex linear operators satisfying all of the following: T(0)=IT(0)=IT(0)=I; T(s+t)=T(s)T(t)T(s+t)=T(s)T(t)T(s+t)=T(s)T(t) whenever s,t>0s,t>0s,t>0; there is a real M≥0M\ge0M≥0 with ∥T(t)∥≤M\|T(t)\|\le M∥T(t)∥≤M for every t≥0t\ge0t≥0; for every x∈Xx\in Xx∈X, the map t↦T(t)xt\mapsto T(t)xt↦T(t)x is continuous on both (0,∞)(0,\infty)(0,∞) and [0,∞)[0,\infty)[0,∞); and, for every x,y∈Xx,y\in Xx,y∈X,

(x,y)∈A⟺lim⁡t↓0T(t)x−xt=−y.(x,y)\in A \quad\Longleftrightarrow\quad \lim_{t\downarrow0}\frac{T(t)x-x}{t}=-y.(x,y)∈A⟺t↓0lim​tT(t)x−x​=−y.

The family is defined at every real time, although these conditions impose no requirements at negative times.

QAlgorithms.Papers.Independent.Intakenext100.P35.HilbertNorm

For a type XXX with a specified normed additive commutative group structure and a specified complex normed space structure, this property says that there exists a complex inner product space structure on XXX whose induced complex normed space structure is equal to the specified one. Completeness is not part of this definition.

QAlgorithms.Papers.Independent.Intakenext100.P35.Hol

For any set S⊆CS\subseteq\mathbb CS⊆C and any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, Hol(S,f)\mathrm{Hol}(S,f)Hol(S,f) means that fff is complex differentiable within SSS at every point of SSS. The set SSS need not be open; when SSS is empty, this condition holds vacuously.

QAlgorithms.Papers.Independent.Intakenext100.P35.InverseAt

Let XXX be a complex normed vector space, let AAA have a closed complex linear graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X that is single-valued and has dense domain, let z∈Cz\in\mathbb Cz∈C, and let R:X→XR:X\to XR:X→X be a continuous complex-linear map. This predicate means that both

∀x∈X,(Rx, x−zRx)∈GA,\forall x\in X,\quad (Rx,\ x-zRx)\in G_A,∀x∈X,(Rx, x−zRx)∈GA​,

and

∀x,y∈X,(x,y)∈GA ⟹ R(zx+y)=x.\forall x,y\in X,\quad (x,y)\in G_A\ \Longrightarrow\ R(zx+y)=x.∀x,y∈X,(x,y)∈GA​ ⟹ R(zx+y)=x.

There is no restriction on zzz, including when z=0z=0z=0.

QAlgorithms.Papers.Independent.Intakenext100.P35.LM

For any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this predicate says that there exist a finite measure μ\muμ on R\mathbb RR and a μ\muμ-Bochner-integrable function w:R→Cw:\mathbb R\to\mathbb Cw:R→C such that μ((−∞,0))=0\mu((-\infty,0))=0μ((−∞,0))=0 and

f(z)=∫Re−ztw(t) dμ(t)(Re⁡z>0).f(z)=\int_{\mathbb R}e^{-zt}w(t)\,d\mu(t) \qquad(\operatorname{Re}z>0).f(z)=∫R​e−ztw(t)dμ(t)(Rez>0).

The measure may have mass at 000, and no values of fff outside the open right half-plane are prescribed.

QAlgorithms.Papers.Independent.Intakenext100.P35.LaplaceRep

For a measure μ\muμ on R\mathbb RR, a function w:R→Cw:\mathbb R\to\mathbb Cw:R→C, and a function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this predicate says that μ\muμ is finite, μ((−∞,0))=0\mu((-\infty,0))=0μ((−∞,0))=0, www is Bochner integrable with respect to μ\muμ, and

f(z)=∫Re−ztw(t) dμ(t)for every z∈C with Re⁡z>0.f(z)=\int_{\mathbb R}e^{-zt}w(t)\,d\mu(t) \qquad\text{for every }z\in\mathbb C\text{ with }\operatorname{Re}z>0.f(z)=∫R​e−ztw(t)dμ(t)for every z∈C with Rez>0.

Mass at t=0t=0t=0 is allowed. The equality imposes no condition on fff outside the open right half-plane.

QAlgorithms.Papers.Independent.Intakenext100.P35.MixedDerivativeClass

For a function g:C→Cg:\mathbb C\to\mathbb Cg:C→C, this property means that (a,b)↦g(a+ib)(a,b)\mapsto g(a+ib)(a,b)↦g(a+ib) is almost everywhere strongly measurable with respect to the product of Lebesgue measure restricted to a>0a>0a>0 and Lebesgue measure in bbb, and that

∫(0,∞) ⁣−ess sup⁡b∈R∥g(a+ib)∥ da<∞.\int_{(0,\infty)}^{\!-} \operatorname*{ess\,sup}_{b\in\mathbb R}\|g(a+ib)\|\,da<\infty.∫(0,∞)−​b∈Resssup​∥g(a+ib)∥da<∞.

The essential supremum uses Lebesgue measure in bbb, and the integral is the extended nonnegative Lebesgue integral. No requirement that ggg be a derivative or be holomorphic is included.

QAlgorithms.Papers.Independent.Intakenext100.P35.NormCauchy

For any function N:(C→C)→[0,∞]N:(\mathbb C\to\mathbb C)\to[0,\infty]N:(C→C)→[0,∞] and any sequence of functions (fn)n∈N(f_n)_{n\in\mathbb N}(fn​)n∈N​ from C\mathbb CC to C\mathbb CC, this predicate means that for every real ε>0\varepsilon>0ε>0 there is a natural number kkk such that

N(fn−fm)<εfor all natural numbers n,m≥k,N(f_n-f_m)<\varepsilon \qquad\text{for all natural numbers }n,m\ge k,N(fn​−fm​)<εfor all natural numbers n,m≥k,

where subtraction is pointwise and ε\varepsilonε is embedded into [0,∞][0,\infty][0,∞]. The indices include zero; no symmetry, triangle inequality, or other norm axiom for NNN is assumed.

QAlgorithms.Papers.Independent.Intakenext100.P35.NormClosure

For any predicate AAA on functions C→C\mathbb C\to\mathbb CC→C, any map N:(C→C)→[0,∞]N:(\mathbb C\to\mathbb C)\to[0,\infty]N:(C→C)→[0,∞], and any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this property means that for every real ε>0\varepsilon>0ε>0, there exists a function g:C→Cg:\mathbb C\to\mathbb Cg:C→C satisfying A(g)A(g)A(g) and N(g−f)<εN(g-f)<\varepsilonN(g−f)<ε, where subtraction is pointwise and ε\varepsilonε is regarded as an extended nonnegative real. No norm axioms for NNN or analyticity assumptions on the functions are required.

QAlgorithms.Papers.Independent.Intakenext100.P35.NormConverges

For any function N:(C→C)→[0,∞]N:(\mathbb C\to\mathbb C)\to[0,\infty]N:(C→C)→[0,∞], any sequence of functions (fn)n∈N(f_n)_{n\in\mathbb N}(fn​)n∈N​ from C\mathbb CC to C\mathbb CC, and any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this predicate means that for every real ε>0\varepsilon>0ε>0 there is a natural number kkk such that

N(fn−f)<εfor every natural number n≥k,N(f_n-f)<\varepsilon\qquad\text{for every natural number }n\ge k,N(fn​−f)<εfor every natural number n≥k,

where subtraction is pointwise and ε\varepsilonε is embedded into [0,∞][0,\infty][0,∞]. The indices include zero, and no norm axioms for NNN or regularity assumptions on the functions are required.

QAlgorithms.Papers.Independent.Intakenext100.P35.NotSeparableSymbols

For any predicate AAA on functions C→C\mathbb C\to\mathbb CC→C and any map N:(C→C)→[0,∞]N:(\mathbb C\to\mathbb C)\to[0,\infty]N:(C→C)→[0,∞], this property means that for every sequence (fn)n∈N(f_n)_{n\in\mathbb N}(fn​)n∈N​ whose members all satisfy AAA, there exist a function fff satisfying AAA and a real ε>0\varepsilon>0ε>0 such that ε≤N(fn−f)\varepsilon\le N(f_n-f)ε≤N(fn​−f) for every nnn, where ε\varepsilonε is viewed in [0,∞][0,\infty][0,∞]. No norm axioms for NNN are assumed. If no function satisfies AAA, the property holds vacuously because there is no such sequence.

QAlgorithms.Papers.Independent.Intakenext100.P35.Qs

For every real number sss, every function g:C→Cg:\mathbb C\to\mathbb Cg:C→C, and every z∈Cz\in\mathbb Cz∈C, this defines

−2sπ∫(0,∞)(∫Ras g(a+ib)(a−ib+z)s+1 db) da.-\frac{2^s}{\pi} \int_{(0,\infty)} \left( \int_{\mathbb R} \frac{a^s\,g(a+ib)}{(a-ib+z)^{s+1}}\,db \right)\,da.−π2s​∫(0,∞)​(∫R​(a−ib+z)s+1asg(a+ib)​db)da.

Here 2s2^s2s and asa^sas are real powers embedded into C\mathbb CC, the denominator uses complex exponentiation, and the integrals are iterated complex-valued Bochner integrals with respect to Lebesgue measure, with the outer measure restricted to a>0a>0a>0. No integrability or regularity assumptions are required to define this value: each Bochner integral uses its total convention of returning zero when its integrand is not integrable, and division by zero returns zero.

QAlgorithms.Papers.Independent.Intakenext100.P35.RealizesB

Let XXX be a complex Banach space, let AAA have a closed complex linear, single-valued graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain, let f:C→Cf:\mathbb C\to\mathbb Cf:C→C be any function, and let F:X→XF:X\to XF:X→X be bounded and complex linear. Define RA(z)R_A(z)RA​(z) by classically choosing a bounded complex linear map RRR satisfying (Ru,u−zRu)∈GA(Ru,u-zRu)\in G_A(Ru,u−zRu)∈GA​ for every uuu and R(zu+v)=uR(zu+v)=uR(zu+v)=u whenever (u,v)∈GA(u,v)\in G_A(u,v)∈GA​, using zero if none exists; define f∞f_\inftyf∞​ as the limit of f(t)f(t)f(t) as real t→+∞t\to+\inftyt→+∞, using zero if no limit exists. For x∈Xx\in Xx∈X and a bounded complex linear functional ℓ\ellℓ, put Kx,ℓ(a,b)=a ℓ(RA(a−ib)(RA(a−ib)x))f′(a+ib)K_{x,\ell}(a,b)=a\,\ell(R_A(a-ib)(R_A(a-ib)x))f'(a+ib)Kx,ℓ​(a,b)=aℓ(RA​(a−ib)(RA​(a−ib)x))f′(a+ib), with derivative zero where complex differentiability fails. The assertion RealizesB⁡(A,f,F)\operatorname{RealizesB}(A,f,F)RealizesB(A,f,F) means that, for every such x,ℓx,\ellx,ℓ, the function (a,b)↦Kx,ℓ(a,b)(a,b)\mapsto K_{x,\ell}(a,b)(a,b)↦Kx,ℓ​(a,b) is almost everywhere strongly measurable for the product of Lebesgue measure restricted to (0,∞)(0,\infty)(0,∞) and Lebesgue measure on R\mathbb RR, and

∫0∞−∫R−∣Kx,ℓ(a,b)∣ db da<∞,ℓ(Fx)=f∞ℓ(x)−2π∫0∞∫RKx,ℓ(a,b) db da.\int_0^\infty{}^{-}\int_{\mathbb R}^{-}|K_{x,\ell}(a,b)|\,db\,da<\infty, \qquad \ell(Fx)=f_\infty\ell(x)-\frac2\pi \int_0^\infty\int_{\mathbb R}K_{x,\ell}(a,b)\,db\,da.∫0∞​−∫R−​∣Kx,ℓ​(a,b)∣dbda<∞,ℓ(Fx)=f∞​ℓ(x)−π2​∫0∞​∫R​Kx,ℓ​(a,b)dbda.

These requirements include x=0x=0x=0 and ℓ=0\ell=0ℓ=0; no holomorphy assumption on fff is included.

QAlgorithms.Papers.Independent.Intakenext100.P35.ResolventSpan

For any sets S,L⊆CS,L\subseteq\mathbb CS,L⊆C, Boolean parameter constants\mathrm{constants}constants, and function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this property means that there exist a natural number nnn, complex coefficients c0,c1,…,cnc_0,c_1,\ldots,c_nc0​,c1​,…,cn​, and parameters λ1,…,λn∈L\lambda_1,\ldots,\lambda_n\in Lλ1​,…,λn​∈L such that f(z)=c0+∑j=1ncj(λj+z)−1f(z)=c_0+\sum_{j=1}^{n}c_j(\lambda_j+z)^{-1}f(z)=c0​+∑j=1n​cj​(λj​+z)−1 for every z∈Sz\in Sz∈S, with the additional requirement c0=0c_0=0c0​=0 when constants=false\mathrm{constants}=\mathrm{false}constants=false. When constants=true\mathrm{constants}=\mathrm{true}constants=true, c0c_0c0​ is unrestricted. The case n=0n=0n=0 is allowed and gives an empty sum equal to zero; parameters need not be distinct. Each inverse is zero when its argument is zero, and no condition is imposed on fff outside SSS. If S=∅S=\varnothingS=∅, the property holds for every fff and either Boolean value.

QAlgorithms.Papers.Independent.Intakenext100.P35.SectorialAtAngle

Let XXX be a complex normed vector space, let AAA have a closed, single-valued complex linear graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain, and let θ∈R\theta\in\mathbb Rθ∈R. For any real α\alphaα, put S0={z:Im⁡z=0, Re⁡z>0}S_0=\{z:\operatorname{Im}z=0,\ \operatorname{Re}z>0\}S0​={z:Imz=0, Rez>0} and, when α≠0\alpha\ne0α=0, put Sα={z:z≠0, ∣arg⁡z∣<α}S_\alpha=\{z:z\ne0,\ |\arg z|<\alpha\}Sα​={z:z=0, ∣argz∣<α}. For each z∈Cz\in\mathbb Cz∈C, let RzR_zRz​ be a classically chosen continuous complex-linear map satisfying (Rzx,x−zRzx)∈GA(R_zx,x-zR_zx)\in G_A(Rz​x,x−zRz​x)∈GA​ for every xxx and Rz(zx+y)=xR_z(zx+y)=xRz​(zx+y)=x for every (x,y)∈GA(x,y)\in G_A(x,y)∈GA​, or the zero map if none exists. Let ΣA\Sigma_AΣA​ consist of those w∈Cw\in\mathbb Cw∈C for which no continuous complex-linear map RRR satisfies (Rx,x+wRx)∈GA(Rx,x+wRx)\in G_A(Rx,x+wRx)∈GA​ for every xxx and R(y−wx)=xR(y-wx)=xR(y−wx)=x for every (x,y)∈GA(x,y)\in G_A(x,y)∈GA​. The predicate means

0≤θ<π,ΣA⊆Sθ‾,0\le\theta<\pi,\qquad \Sigma_A\subseteq\overline{S_\theta},0≤θ<π,ΣA​⊆Sθ​​,

and, for every real ψ\psiψ with θ<ψ<π\theta<\psi<\piθ<ψ<π, there exists a real C≥0C\ge0C≥0 such that ∥zRz∥≤C\|zR_z\|\le C∥zRz​∥≤C for every z∈Sπ−ψz\in S_{\pi-\psi}z∈Sπ−ψ​, where the norm is the operator norm. The bound uses the zero fallback wherever the defining inverse does not exist; it does not separately require existence there. The case θ=0\theta=0θ=0 is included.

QAlgorithms.Papers.Independent.Intakenext100.P35.SectorialBelow

Let XXX be a complex normed vector space and let AAA have a closed, single-valued complex linear graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain. For real α\alphaα, define S0={z:Im⁡z=0, Re⁡z>0}S_0=\{z:\operatorname{Im}z=0,\ \operatorname{Re}z>0\}S0​={z:Imz=0, Rez>0} and Sα={z:z≠0, ∣arg⁡z∣<α}S_\alpha=\{z:z\ne0,\ |\arg z|<\alpha\}Sα​={z:z=0, ∣argz∣<α} when α≠0\alpha\ne0α=0. For each zzz, let RzR_zRz​ be a classically chosen continuous complex-linear map satisfying (Rzx,x−zRzx)∈GA(R_zx,x-zR_zx)\in G_A(Rz​x,x−zRz​x)∈GA​ for every xxx and Rz(zx+y)=xR_z(zx+y)=xRz​(zx+y)=x for every (x,y)∈GA(x,y)\in G_A(x,y)∈GA​, or zero if none exists. Let ΣA\Sigma_AΣA​ be the set of www for which no continuous complex-linear map RRR satisfies (Rx,x+wRx)∈GA(Rx,x+wRx)\in G_A(Rx,x+wRx)∈GA​ for every xxx and R(y−wx)=xR(y-wx)=xR(y−wx)=x for every (x,y)∈GA(x,y)\in G_A(x,y)∈GA​. This predicate means that there exists a real θ\thetaθ with 0≤θ<π/20\le\theta<\pi/20≤θ<π/2 such that also 0≤θ<π0\le\theta<\pi0≤θ<π, ΣA⊆Sθ‾\Sigma_A\subseteq\overline{S_\theta}ΣA​⊆Sθ​​, and

∀ψ∈(θ,π),∃C≥0,∀z∈Sπ−ψ,∥zRz∥≤C.\forall\psi\in(\theta,\pi),\quad \exists C\ge0,\quad \forall z\in S_{\pi-\psi},\quad \|zR_z\|\le C.∀ψ∈(θ,π),∃C≥0,∀z∈Sπ−ψ​,∥zRz​∥≤C.

The angle θ=0\theta=0θ=0 is permitted, and the bounds use the zero fallback for RzR_zRz​ if its defining inverse does not exist.

QAlgorithms.Papers.Independent.Intakenext100.P35.SectorialLimits

For an arbitrary function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this property requires that for every real angle 0<ψ<π/20<\psi<\pi/20<ψ<π/2, with

Sψ={z∈C:z≠0, ∣arg⁡z∣<ψ},S_\psi=\{z\in\mathbb C:z\ne0,\ |\arg z|<\psi\},Sψ​={z∈C:z=0, ∣argz∣<ψ},

both f(z)→c0f(z)\to c_0f(z)→c0​ as z→0z\to0z→0 within SψS_\psiSψ​ and f(z)→c∞f(z)\to c_\inftyf(z)→c∞​ as ∥z∥→∞\|z\|\to\infty∥z∥→∞ within SψS_\psiSψ​. Here c0c_0c0​ is a chosen limit of f(t)f(t)f(t) as real t↓0t\downarrow0t↓0, or 000 if that limit does not exist, and c∞c_\inftyc∞​ is a chosen limit of f(t)f(t)f(t) as real t→+∞t\to+\inftyt→+∞, or 000 if that limit does not exist. The two required sectorial convergences in particular imply that these positive-real-axis limits exist. No holomorphy is included in this property.

QAlgorithms.Papers.Independent.Intakenext100.P35.TransposeEigen

Let XXX be a complex normed space, let A⊆X×XA\subseteq X\times XA⊆X×X be a closed complex linear subspace that is single valued and has dense domain, let λ∈C\lambda\in\mathbb Cλ∈C, and let ℓ:X→C\ell:X\to\mathbb Cℓ:X→C be a bounded complex linear functional. This property says that

ℓ(y)=λ ℓ(x)for every (x,y)∈A.\ell(y)=\lambda\,\ell(x) \qquad\text{for every }(x,y)\in A.ℓ(y)=λℓ(x)for every (x,y)∈A.

There is no requirement that ℓ≠0\ell\ne0ℓ=0; the zero functional satisfies this property for every λ\lambdaλ.

QAlgorithms.Papers.Independent.Intakenext100.P35.Vanishes

For any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this property means that f(z)f(z)f(z) tends to zero as ∥z∥\lVert z\rVert∥z∥ tends to infinity while zzz stays in the open right half-plane {z:Re⁡z>0}\{z:\operatorname{Re}z>0\}{z:Rez>0}. Explicitly, for every real ε>0\varepsilon>0ε>0, there exists a real RRR such that Re⁡z>0\operatorname{Re}z>0Rez>0 and ∥z∥≥R\lVert z\rVert\ge R∥z∥≥R imply ∥f(z)∥<ε\lVert f(z)\rVert<\varepsilon∥f(z)∥<ε.

QAlgorithms.Papers.Independent.Intakenext100.P35.WeightedDerivativeClass

For every real number sss and every function g:C→Cg:\mathbb C\to\mathbb Cg:C→C, this predicate says that the map (a,b)↦g(a+ib)(a,b)\mapsto g(a+ib)(a,b)↦g(a+ib) is almost everywhere strongly measurable with respect to the product of Lebesgue measure restricted to (0,∞)(0,\infty)(0,∞) in the first coordinate and Lebesgue measure on R\mathbb RR in the second, and that

∫(0,∞)∫Ras(a2+b2)(s+1)/2 ∣g(a+ib)∣ db da<+∞.\int_{(0,\infty)}\int_{\mathbb R} \frac{a^{s}}{(a^{2}+b^{2})^{(s+1)/2}}\,|g(a+ib)|\,db\,da<+\infty.∫(0,∞)​∫R​(a2+b2)(s+1)/2as​∣g(a+ib)∣dbda<+∞.

These are iterated nonnegative Lebesgue integrals with real powers. There is no restriction on sss and no holomorphy or derivative condition on ggg.

QAlgorithms.Papers.Independent.Intakenext100.P35.arccot

For every z∈Cz\in\mathbb Cz∈C, this function is defined by

arccot⁡(z)=12ilog⁡ ⁣(z+iz−i),\operatorname{arccot}(z)=\frac{1}{2i}\log\!\left(\frac{z+i}{z-i}\right),arccot(z)=2i1​log(z−iz+i​),

using the principal complex logarithm and total complex division. In particular, division by zero gives zero and the complex logarithm of zero is defined as zero, so the displayed expression has value zero at both z=iz=iz=i and z=−iz=-iz=−i.

QAlgorithms.Papers.Independent.Intakenext100.P35.arccotZero

For every z∈Cz\in\mathbb Cz∈C, this function equals the real number π/2\pi/2π/2, embedded in C\mathbb CC, when z=0z=0z=0, and equals (2i)−1log⁡((z+i)/(z−i))(2i)^{-1}\log((z+i)/(z-i))(2i)−1log((z+i)/(z−i)) otherwise. The logarithm is the principal complex logarithm, with total division and log⁡0=0\log 0=0log0=0; consequently its value at either z=iz=iz=i or z=−iz=-iz=−i is zero.

QAlgorithms.Papers.Independent.Intakenext100.P35.bOperator

Let XXX be a complex Banach space, let AAA have a closed complex linear, single-valued graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain, and let f:C→Cf:\mathbb C\to\mathbb Cf:C→C be any function. For each zzz, let RA(z)R_A(z)RA​(z) be a classically chosen bounded complex linear map RRR satisfying (Ru,u−zRu)∈GA(Ru,u-zRu)\in G_A(Ru,u−zRu)∈GA​ for every uuu and R(zu+v)=uR(zu+v)=uR(zu+v)=u whenever (u,v)∈GA(u,v)\in G_A(u,v)∈GA​, using zero if no such map exists. Let f∞f_\inftyf∞​ be the limit of f(t)f(t)f(t) as real t→+∞t\to+\inftyt→+∞, using zero if no limit exists, and put Kx,ℓ(a,b)=a ℓ(RA(a−ib)(RA(a−ib)x))f′(a+ib)K_{x,\ell}(a,b)=a\,\ell(R_A(a-ib)(R_A(a-ib)x))f'(a+ib)Kx,ℓ​(a,b)=aℓ(RA​(a−ib)(RA​(a−ib)x))f′(a+ib), with derivative zero where complex differentiability fails. The value bOperator⁡(A,f)\operatorname{bOperator}(A,f)bOperator(A,f) is a classically chosen bounded complex linear map F:X→XF:X\to XF:X→X such that, for every x∈Xx\in Xx∈X and bounded complex linear functional ℓ\ellℓ, Kx,ℓK_{x,\ell}Kx,ℓ​ is almost everywhere strongly measurable for Lebesgue measure on (0,∞)×R(0,\infty)\times\mathbb R(0,∞)×R,

∫0∞−∫R−∣Kx,ℓ(a,b)∣ db da<∞,ℓ(Fx)=f∞ℓ(x)−2π∫0∞∫RKx,ℓ(a,b) db da,\int_0^\infty{}^{-}\int_{\mathbb R}^{-}|K_{x,\ell}(a,b)|\,db\,da<\infty, \qquad \ell(Fx)=f_\infty\ell(x)-\frac2\pi\int_0^\infty\int_{\mathbb R}K_{x,\ell}(a,b)\,db\,da,∫0∞​−∫R−​∣Kx,ℓ​(a,b)∣dbda<∞,ℓ(Fx)=f∞​ℓ(x)−π2​∫0∞​∫R​Kx,ℓ​(a,b)dbda,

if any such FFF exists; otherwise it is the zero operator. The definition itself asserts neither existence nor uniqueness of a realizing map.

QAlgorithms.Papers.Independent.Intakenext100.P35.bSemi

For any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this defines the extended nonnegative quantity

b(f)=∫(0,∞) ⁣−sup⁡b∈R∥f′(a+ib)∥ da∈[0,∞].b(f)=\int_{(0,\infty)}^{\!-} \sup_{b\in\mathbb R}\|f'(a+ib)\|\,da \quad\in[0,\infty].b(f)=∫(0,∞)−​b∈Rsup​∥f′(a+ib)∥da∈[0,∞].

The integral is the nonnegative Lebesgue integral with respect to real Lebesgue measure restricted to (0,∞)(0,\infty)(0,∞); each norm is embedded into the extended nonnegative reals before taking the supremum. The derivative is the total complex derivative, assigned the value 000 wherever fff is not complex differentiable. Neither differentiability nor measurability is assumed, and the result may be infinite.

QAlgorithms.Papers.Independent.Intakenext100.P35.bSize

For any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this defines

bSize⁡(f)=sup⁡Re⁡z>0∥f(z)∥+∫(0,∞) ⁣−sup⁡b∈R∥f′(a+ib)∥ da∈[0,∞].\operatorname{bSize}(f)= \sup_{\operatorname{Re}z>0}\|f(z)\| + \int_{(0,\infty)}^{\!-} \sup_{b\in\mathbb R}\|f'(a+ib)\|\,da \quad\in[0,\infty].bSize(f)=Rez>0sup​∥f(z)∥+∫(0,∞)−​b∈Rsup​∥f′(a+ib)∥da∈[0,∞].

The suprema, integral, and addition are taken in the extended nonnegative reals, with norms embedded there. No regularity or finiteness is assumed; the complex derivative is assigned 000 at points where fff is not complex differentiable.

QAlgorithms.Papers.Independent.Intakenext100.P35.besovOperatorIntegrand

Let XXX be a complex normed space, let A⊆X×XA\subseteq X\times XA⊆X×X be a closed complex linear subspace that is single valued and has dense domain, let f:C→Cf:\mathbb C\to\mathbb Cf:C→C be any function, and let a,b∈Ra,b\in\mathbb Ra,b∈R, with no positivity requirement on aaa. For z∈Cz\in\mathbb Cz∈C, define RA(z)R_A(z)RA​(z) to be a classically chosen bounded complex linear operator R:X→XR:X\to XR:X→X satisfying

(Rx,x−zRx)∈Afor every x∈X,R(zx+y)=xfor every (x,y)∈A,(Rx,x-zRx)\in A\quad\text{for every }x\in X, \qquad R(zx+y)=x\quad\text{for every }(x,y)\in A,(Rx,x−zRx)∈Afor every x∈X,R(zx+y)=xfor every (x,y)∈A,

if such an operator exists, and define RA(z)=0R_A(z)=0RA​(z)=0 otherwise. The declared bounded operator is

a f′(a+ib) RA(a−ib)∘RA(a−ib).a\,f'(a+ib)\,R_A(a-ib)\circ R_A(a-ib).af′(a+ib)RA​(a−ib)∘RA​(a−ib).

Here f′f'f′ is the total complex derivative, which is defined to be zero at a point where fff is not complex differentiable. No holomorphy or resolvent existence hypothesis is imposed.

QAlgorithms.Papers.Independent.Intakenext100.P35.besovPairing

For any two functions g,f:C→Cg,f:\mathbb C\to\mathbb Cg,f:C→C, in that argument order, this defines the complex number

∫(0,∞)a(∫Rg′(a−ib)f′(a+ib) db) da,\int_{(0,\infty)}a\left(\int_{\mathbb R}g'(a-ib)f'(a+ib)\,db\right)\,da,∫(0,∞)​a(∫R​g′(a−ib)f′(a+ib)db)da,

using iterated complex Bochner integrals with Lebesgue measure; the outer measure is Lebesgue measure restricted to (0,∞)(0,\infty)(0,∞). There is no complex conjugation in the product, and no differentiability, measurability, or integrability hypotheses are required. Each derivative is assigned 000 wherever the corresponding function is not complex differentiable, and each Bochner integral is assigned 000 if its integrand is not integrable.

QAlgorithms.Papers.Independent.Intakenext100.P35.besovPrimitive

For any function g:C→Cg:\mathbb C\to\mathbb Cg:C→C and any z∈Cz\in\mathbb Cz∈C, this defines

besovPrimitive⁡(g)(z)=−2π∫(0,∞)(∫Ra g(a+ib)(a−ib+z)2 db) da.\operatorname{besovPrimitive}(g)(z)= -\frac{2}{\pi} \int_{(0,\infty)} \left(\int_{\mathbb R} \frac{a\,g(a+ib)}{(a-ib+z)^2}\,db\right)\,da.besovPrimitive(g)(z)=−π2​∫(0,∞)​(∫R​(a−ib+z)2ag(a+ib)​db)da.

These are nested complex-valued Bochner integrals, first in bbb against Lebesgue measure and then in aaa against Lebesgue measure restricted to a>0a>0a>0. No measurability or integrability hypotheses are imposed: each Bochner integral is defined to be 000 when its integrand is not integrable. Complex division is total, so the displayed integrand is 000 wherever its denominator vanishes.

QAlgorithms.Papers.Independent.Intakenext100.P35.betaValue

For arbitrary real numbers aaa and bbb, this defines the real number

betaValue⁡(a,b)=∫(0,1)ta−1(1−t)b−1 dt,\operatorname{betaValue}(a,b) =\int_{(0,1)} t^{a-1}(1-t)^{b-1}\,dt,betaValue(a,b)=∫(0,1)​ta−1(1−t)b−1dt,

where the powers are real powers and the integral uses Lebesgue measure restricted to the open interval (0,1)(0,1)(0,1). No positivity assumptions on a,ba,ba,b or integrability hypothesis are imposed. This is the total Bochner integral, whose value is zero when the integrand is not integrable on this interval.

QAlgorithms.Papers.Independent.Intakenext100.P35.boundaryAverage

For every real angle ψ\psiψ, every function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, and every real number ttt, let SψS_\psiSψ​ be the positive real axis {z:Im⁡z=0, Re⁡z>0}\{z:\operatorname{Im}z=0,\ \operatorname{Re}z>0\}{z:Imz=0, Rez>0} when ψ=0\psi=0ψ=0, and otherwise let Sψ={z:z≠0, ∣arg⁡z∣<ψ}S_\psi=\{z:z\ne0,\ |\arg z|<\psi\}Sψ​={z:z=0, ∣argz∣<ψ}. Define Eψf(w)=f(w)E_\psi f(w)=f(w)Eψ​f(w)=f(w) for w∈Sψw\in S_\psiw∈Sψ​; for w∉Sψw\notin S_\psiw∈/Sψ​, define Eψf(w)E_\psi f(w)Eψ​f(w) to be a classically chosen complex limit of fff as its argument tends to www within SψS_\psiSψ​, if such a limit exists, and zero otherwise. Then this defines

boundaryAverage⁡(ψ,f,t)=Eψf(eiψt)+Eψf(e−iψt)2.\operatorname{boundaryAverage}(\psi,f,t) =\frac{E_\psi f(e^{i\psi}t)+E_\psi f(e^{-i\psi}t)}{2}.boundaryAverage(ψ,f,t)=2Eψ​f(eiψt)+Eψ​f(e−iψt)​.

No regularity of fff, existence or uniqueness of these limits, or restrictions on ψ,t\psi,tψ,t are assumed. In particular, SψS_\psiSψ​ is empty when ψ<0\psi<0ψ<0; a limit along an empty within-set filter can be any complex number, and the definition uses its classical choice.

QAlgorithms.Papers.Independent.Intakenext100.P35.closedHp

The set closedHp\mathrm{closedHp}closedHp is the closed right half-plane {z∈C:Re⁡z≥0}\{z\in\mathbb C:\operatorname{Re}z\geq0\}{z∈C:Rez≥0}, including the imaginary axis.

QAlgorithms.Papers.Independent.Intakenext100.P35.dBoundedSize

For any real number sss and any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this defines the following value in [0,∞][0,\infty][0,∞]:

sup⁡z∈H∣f(z)∣+∫(0,∞)∫Ras(a2+b2)(s+1)/2 ∣f′(a+ib)∣ db da,H={z:Re⁡z>0}.\sup_{z\in\mathbb H}|f(z)| + \int_{(0,\infty)}\int_{\mathbb R} \frac{a^s}{(a^2+b^2)^{(s+1)/2}}\,|f'(a+ib)|\,db\,da, \qquad \mathbb H=\{z:\operatorname{Re}z>0\}.z∈Hsup​∣f(z)∣+∫(0,∞)​∫R​(a2+b2)(s+1)/2as​∣f′(a+ib)∣dbda,H={z:Rez>0}.

The integrals are nonnegative Lebesgue integrals. No differentiability, measurability, boundedness, or finiteness assumption is imposed; f′f'f′ denotes the total complex derivative operation, which is zero at points where fff is not complex differentiable, and the displayed value may be infinite.

QAlgorithms.Papers.Independent.Intakenext100.P35.dSemi

For every real number sss and every function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this defines the extended nonnegative real number

∫(0,∞)∫Ras(a2+b2)(s+1)/2 ∣f′(a+ib)∣ db da,\int_{(0,\infty)}\int_{\mathbb R} \frac{a^{s}}{(a^{2}+b^{2})^{(s+1)/2}}\,|f'(a+ib)|\,db\,da,∫(0,∞)​∫R​(a2+b2)(s+1)/2as​∣f′(a+ib)∣dbda,

using real powers and iterated nonnegative Lebesgue integrals. No restriction on sss and no differentiability, measurability, or finiteness hypothesis on fff is required. The total complex derivative is 000 wherever fff is not complex differentiable, and the integral may equal +∞+\infty+∞.

QAlgorithms.Papers.Independent.Intakenext100.P35.dSize

For every real number sss and every function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, let ccc be a chosen complex limit of f(t)f(t)f(t) as the real variable t→+∞t\to+\inftyt→+∞, if such a limit exists, and let c=0c=0c=0 otherwise. This defines the extended nonnegative real number

∣c∣+∫(0,∞)∫Ras(a2+b2)(s+1)/2 ∣f′(a+ib)∣ db da.|c|+\int_{(0,\infty)}\int_{\mathbb R} \frac{a^{s}}{(a^{2}+b^{2})^{(s+1)/2}}\,|f'(a+ib)|\,db\,da.∣c∣+∫(0,∞)​∫R​(a2+b2)(s+1)/2as​∣f′(a+ib)∣dbda.

The powers are real powers and the integrals are iterated nonnegative Lebesgue integrals. No restriction on sss, existence of the limit, holomorphy, measurability, or finiteness is assumed; the derivative is assigned 000 wherever fff is not complex differentiable, and the resulting size may be +∞+\infty+∞.

QAlgorithms.Papers.Independent.Intakenext100.P35.eSemi

For any function g:C→Cg:\mathbb C\to\mathbb Cg:C→C, this defines the extended nonnegative real number

sup⁡a>0a∫R∣g′(a+ib)∣ db,\sup_{a>0} a\int_{\mathbb R}|g'(a+ib)|\,db,a>0sup​a∫R​∣g′(a+ib)∣db,

where dbdbdb is Lebesgue measure and the integral is the nonnegative Lebesgue integral. No holomorphy, measurability, or finiteness assumption is imposed on ggg. Here g′g'g′ is the total complex derivative, which is assigned the value 000 wherever ggg is not complex differentiable; the displayed supremum may equal +∞+\infty+∞.

QAlgorithms.Papers.Independent.Intakenext100.P35.extension

For a set S⊆CS\subseteq\mathbb CS⊆C, an arbitrary function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, and z∈Cz\in\mathbb Cz∈C, this defines

extension⁡(S,f)(z)={f(z),z∈S,a chosen limit of f along the neighborhood filter within S at z,z∉S and such a limit exists,0,z∉S and no such limit exists.\operatorname{extension}(S,f)(z)= \begin{cases} f(z),&z\in S,\\ \text{a chosen limit of $f$ along the neighborhood filter within $S$ at $z$},& z\notin S\text{ and such a limit exists},\\ 0,&z\notin S\text{ and no such limit exists}. \end{cases}extension(S,f)(z)=⎩⎨⎧​f(z),a chosen limit of f along the neighborhood filter within S at z,0,​z∈S,z∈/S and such a limit exists,z∈/S and no such limit exists.​

No regularity of fff or SSS is assumed. If the neighborhood filter within SSS at zzz is the bottom filter, as happens when z∉S‾z\notin\overline Sz∈/S, every complex number is a limit and the definition chooses one without specifying which.

QAlgorithms.Papers.Independent.Intakenext100.P35.gammaSize

Let XXX be a complex Banach space, and let AAA have a closed complex linear, single-valued graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain. For each z∈Cz\in\mathbb Cz∈C, let RA(z)R_A(z)RA​(z) be a classically chosen bounded complex linear map R:X→XR:X\to XR:X→X satisfying (Ru,u−zRu)∈GA(Ru,u-zRu)\in G_A(Ru,u−zRu)∈GA​ for every u∈Xu\in Xu∈X and R(zx+y)=xR(zx+y)=xR(zx+y)=x for every (x,y)∈GA(x,y)\in G_A(x,y)∈GA​, or the zero map if no such map exists. The extended nonnegative real number gammaSize⁡(A)\operatorname{gammaSize}(A)gammaSize(A) is

sup⁡x∈X∥x∥≤1sup⁡ℓ:X→C bounded complex linear∥ℓ∥≤1sup⁡a>0a∫R−∣ℓ ⁣(RA(a+ib)(RA(a+ib)x))∣ db.\sup_{\substack{x\in X\\\|x\|\le1}} \sup_{\substack{\ell:X\to\mathbb C\ \text{bounded complex linear}\\\|\ell\|\le1}} \sup_{a>0} a\int_{\mathbb R}^{-} \left|\ell\!\left(R_A(a+ib)\bigl(R_A(a+ib)x\bigr)\right)\right|\,db.x∈X∥x∥≤1​sup​ℓ:X→C bounded complex linear∥ℓ∥≤1​sup​a>0sup​a∫R−​​ℓ(RA​(a+ib)(RA​(a+ib)x))​db.

The unit balls include zero, and the value is allowed to be +∞+\infty+∞; no spectral or finiteness hypothesis is imposed here.

QAlgorithms.Papers.Independent.Intakenext100.P35.gsfSlice

Let XXX be a complex normed vector space, let AAA have a closed, single-valued complex linear graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain, let x∈Xx\in Xx∈X, let ℓ:X→C\ell:X\to\mathbb Cℓ:X→C be a continuous complex-linear functional, and let a∈Ra\in\mathbb Ra∈R. For each z∈Cz\in\mathbb Cz∈C, let RzR_zRz​ be a classically chosen continuous complex-linear map satisfying (Rzu,u−zRzu)∈GA(R_zu,u-zR_zu)\in G_A(Rz​u,u−zRz​u)∈GA​ for every uuu and Rz(zu+v)=uR_z(zu+v)=uRz​(zu+v)=u for every (u,v)∈GA(u,v)\in G_A(u,v)∈GA​, or zero if none exists. This defines

gsfSlice⁡(A,x,ℓ,a)=∫R−∣ℓ ⁣(Ra+ib(Ra+ibx))∣ db∈[0,∞],\operatorname{gsfSlice}(A,x,\ell,a) =\int_{\mathbb R}^{-}\left|\ell\!\left(R_{a+ib}(R_{a+ib}x)\right)\right|\,db \quad\in[0,\infty],gsfSlice(A,x,ℓ,a)=∫R−​∣ℓ(Ra+ib​(Ra+ib​x))∣db∈[0,∞],

where the integral is the nonnegative Lebesgue integral. There is no restriction on aaa, no integrability or measurability hypothesis, and no assertion that the value is finite; the resolvents use their zero fallback wherever their defining inverses do not exist.

QAlgorithms.Papers.Independent.Intakenext100.P35.hSize

For every real number ψ\psiψ and every function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this defines the following value in [0,∞][0,\infty][0,∞]:

sup⁡z∈Sψ∣f(z)∣+sup⁡ϕ∈R∣ϕ∣<ψ∫(0,∞)(∣f′(teiϕ)∣+∣f′(te−iϕ)∣) dt,Sψ={{z:Im⁡z=0, Re⁡z>0},ψ=0,{z:z≠0, ∣arg⁡z∣<ψ},ψ≠0.\sup_{z\in S_\psi}|f(z)| + \sup_{\substack{\phi\in\mathbb R\\|\phi|<\psi}} \int_{(0,\infty)} \left( |f'(te^{i\phi})|+|f'(te^{-i\phi})| \right)\,dt, \qquad S_\psi= \begin{cases} \{z:\operatorname{Im}z=0,\ \operatorname{Re}z>0\},&\psi=0,\\ \{z:z\ne0,\ |\arg z|<\psi\},&\psi\ne0. \end{cases}z∈Sψ​sup​∣f(z)∣+ϕ∈R∣ϕ∣<ψ​sup​∫(0,∞)​(∣f′(teiϕ)∣+∣f′(te−iϕ)∣)dt,Sψ​={{z:Imz=0, Rez>0},{z:z=0, ∣argz∣<ψ},​ψ=0,ψ=0.​

The integral is a nonnegative Lebesgue integral, and f′f'f′ is the total complex derivative, zero where complex differentiability fails. No regularity or finiteness is assumed. Empty suprema are zero, so the value is zero for ψ<0\psi<0ψ<0; for ψ=0\psi=0ψ=0 it is just the supremum of ∣f∣|f|∣f∣ on the positive real ray.

QAlgorithms.Papers.Independent.Intakenext100.P35.hardySize

For every real number ψ\psiψ and every function g:C→Cg:\mathbb C\to\mathbb Cg:C→C, this defines the value in [0,∞][0,\infty][0,∞]

sup⁡ϕ∈R∣ϕ∣<ψ∫(0,∞)(∣g(teiϕ)∣+∣g(te−iϕ)∣) dt,\sup_{\substack{\phi\in\mathbb R\\|\phi|<\psi}} \int_{(0,\infty)} \left( |g(te^{i\phi})|+|g(te^{-i\phi})| \right)\,dt,ϕ∈R∣ϕ∣<ψ​sup​∫(0,∞)​(∣g(teiϕ)∣+∣g(te−iϕ)∣)dt,

where the integral is a nonnegative Lebesgue integral. No measurability or finiteness assumption is imposed. If ψ≤0\psi\le0ψ≤0, the set of eligible angles is empty and the supremum is zero; if ψ>0\psi>0ψ>0, the angle ϕ=0\phi=0ϕ=0 is included and its integrand is 2∣g(t)∣2|g(t)|2∣g(t)∣.

QAlgorithms.Papers.Independent.Intakenext100.P35.hp

The set hp\mathrm{hp}hp is the open right half-plane {z∈C:Re⁡z>0}\{z\in\mathbb C:\operatorname{Re}z>0\}{z∈C:Rez>0}.

QAlgorithms.Papers.Independent.Intakenext100.P35.hpSize

For any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, this defines an extended nonnegative real number as the infimum of

∫R∣w(t)∣ dμ(t)\int_{\mathbb R}|w(t)|\,d\mu(t)∫R​∣w(t)∣dμ(t)

over all finite measures μ\muμ on R\mathbb RR and all μ\muμ-Bochner-integrable functions w:R→Cw:\mathbb R\to\mathbb Cw:R→C satisfying μ((−∞,0))=0\mu((-\infty,0))=0μ((−∞,0))=0 and f(z)=∫Re−ztw(t) dμ(t)f(z)=\int_{\mathbb R}e^{-zt}w(t)\,d\mu(t)f(z)=∫R​e−ztw(t)dμ(t) for every zzz with Re⁡z>0\operatorname{Re}z>0Rez>0. The integral used for the infimum is the nonnegative Lebesgue integral. If there is no such representation, the infimum of the empty set in [0,+∞][0,+\infty][0,+∞] is +∞+\infty+∞; the representation does not constrain fff outside the open right half-plane.

QAlgorithms.Papers.Independent.Intakenext100.P35.infinityValue

For any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, infinityValue(f)\mathrm{infinityValue}(f)infinityValue(f) is the complex limit of f(t)f(t)f(t) as the real variable t→+∞t\to+\inftyt→+∞, with ttt regarded as a complex number, if this limit exists; otherwise it is 000. This definition concerns only values along the real axis.

QAlgorithms.Papers.Independent.Intakenext100.P35.infinityWithin

For any set S⊆CS\subseteq\mathbb CS⊆C, this defines the filter of tending to infinity in norm while remaining in SSS: a set E⊆CE\subseteq\mathbb CE⊆C belongs to this filter exactly when there is a real number RRR such that every z∈Sz\in Sz∈S with ∥z∥≥R\|z\|\ge R∥z∥≥R belongs to EEE. Equivalently, it is the infimum of the pullback of the real filter at +∞+\infty+∞ under z↦∥z∥z\mapsto\|z\|z↦∥z∥ and the principal filter of SSS. For bounded SSS, including S=∅S=\varnothingS=∅, this is the bottom filter.

QAlgorithms.Papers.Independent.Intakenext100.P35.limitValue

For any type DDD, any filter FFF on DDD, and any function f:D→Cf:D\to\mathbb Cf:D→C, limitValue(F,f)\mathrm{limitValue}(F,f)limitValue(F,f) is a classically chosen complex number ccc such that fff converges to ccc along FFF, if at least one such ccc exists; otherwise it is 000. No assumption that FFF is a proper filter is imposed, and no uniqueness condition is required: for the bottom filter every complex number is a limit, so the definition chooses one of them.

QAlgorithms.Papers.Independent.Intakenext100.P35.localDerivative

For any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C and real numbers r,δr,\deltar,δ, this extended nonnegative real is

∫(0,δ) ⁣−sup⁡b∈R∣b∣≤r∥f′(a+ib)∥ da,\int_{(0,\delta)}^{\!-} \sup_{\substack{b\in\mathbb R\\ |b|\le r}} \lVert f'(a+ib)\rVert\,da,∫(0,δ)−​b∈R∣b∣≤r​sup​∥f′(a+ib)∥da,

where the integral is the nonnegative Lebesgue integral and the supremum is taken in [0,∞][0,\infty][0,∞]. Here f′f'f′ is the total complex derivative, assigned zero at points where fff is not complex differentiable. The value can be infinite. If δ≤0\delta\le0δ≤0, the integration set is empty and the value is zero; if r<0r<0r<0, each supremum is empty and therefore zero.

QAlgorithms.Papers.Independent.Intakenext100.P35.mSize

Let XXX be a complex normed vector space and let AAA have a closed, single-valued complex linear graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain. For each z∈Cz\in\mathbb Cz∈C, let RzR_zRz​ be a classically chosen continuous complex-linear map satisfying (Rzx,x−zRzx)∈GA(R_zx,x-zR_zx)\in G_A(Rz​x,x−zRz​x)∈GA​ for all xxx and Rz(zx+y)=xR_z(zx+y)=xRz​(zx+y)=x for all (x,y)∈GA(x,y)\in G_A(x,y)∈GA​, or the zero map if none exists. This defines the extended nonnegative real number

mSize⁡(A)=sup⁡Re⁡z>0∥zRz∥.\operatorname{mSize}(A)= \sup_{\operatorname{Re}z>0}\|zR_z\|.mSize(A)=Rez>0sup​∥zRz​∥.

Each norm is embedded into [0,∞][0,\infty][0,∞]. No finiteness or inverse-existence assumption is imposed, so this supremum may be +∞+\infty+∞ and uses zero at points where the inverse does not exist.

QAlgorithms.Papers.Independent.Intakenext100.P35.openSectorDomain

For every real ψ\psiψ, the set openSectorDomain(ψ)\mathrm{openSectorDomain}(\psi)openSectorDomain(ψ) is defined by {z∈C:Im⁡z=0, Re⁡z>0}\{z\in\mathbb C:\operatorname{Im}z=0,\ \operatorname{Re}z>0\}{z∈C:Imz=0, Rez>0} when ψ=0\psi=0ψ=0, and by {z∈C:z≠0, ∣arg⁡z∣<ψ}\{z\in\mathbb C:z\neq0,\ |\arg z|<\psi\}{z∈C:z=0, ∣argz∣<ψ} otherwise, where arg⁡\argarg is the principal complex argument. Thus ψ=0\psi=0ψ=0 gives the positive real axis, ψ<0\psi<0ψ<0 gives the empty set, and the origin is excluded for every ψ\psiψ; no restriction on the real parameter ψ\psiψ is imposed.

QAlgorithms.Papers.Independent.Intakenext100.P35.operatorCayley

Let XXX be a complex normed space, and let A⊆X×XA\subseteq X\times XA⊆X×X be a closed complex linear subspace that is single valued and has dense domain. Choose a bounded complex linear operator R:X→XR:X\to XR:X→X satisfying

(Rx,x−Rx)∈Afor every x∈X,R(x+y)=xfor every (x,y)∈A,(Rx,x-Rx)\in A\quad\text{for every }x\in X, \qquad R(x+y)=x\quad\text{for every }(x,y)\in A,(Rx,x−Rx)∈Afor every x∈X,R(x+y)=xfor every (x,y)∈A,

if one exists, and otherwise set R=0R=0R=0. The declared operator is I−2RI-2RI−2R. In particular, the definition returns III if no such bounded operator RRR exists.

QAlgorithms.Papers.Independent.Intakenext100.P35.operatorSpectrum

Let XXX be a complex normed vector space and let AAA have a closed, single-valued complex linear graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain. This defines its specified spectrum to be the set of complex numbers zzz for which there is no continuous complex-linear map R:X→XR:X\to XR:X→X satisfying both

∀x∈X,(Rx, x+zRx)∈GA,∀(x,y)∈GA,R(y−zx)=x.\forall x\in X,\quad (Rx,\ x+zRx)\in G_A, \qquad \forall (x,y)\in G_A,\quad R(y-zx)=x.∀x∈X,(Rx, x+zRx)∈GA​,∀(x,y)∈GA​,R(y−zx)=x.

The definition ranges over all z∈Cz\in\mathbb Cz∈C, including zero.

QAlgorithms.Papers.Independent.Intakenext100.P35.phiSymbol

For every real number ttt and every complex number zzz, this defines

ϕt(z)=zz+1exp⁡ ⁣(−tz),\phi_t(z)=\frac{z}{z+1}\exp\!\left(-\frac{t}{z}\right),ϕt​(z)=z+1z​exp(−zt​),

with ttt regarded as a complex number. There are no positivity or domain restrictions on ttt or zzz. Complex division is total, with division by zero equal to zero; consequently ϕt(0)=ϕt(−1)=0\phi_t(0)=\phi_t(-1)=0ϕt​(0)=ϕt​(−1)=0 for every ttt.

QAlgorithms.Papers.Independent.Intakenext100.P35.plusResolvent

Let XXX be a complex normed vector space and let AAA have a closed, single-valued complex linear graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain. For every z∈Cz\in\mathbb Cz∈C, this defines a continuous complex-linear map Rz:X→XR_z:X\to XRz​:X→X by choosing one satisfying

∀x∈X,(Rzx, x−zRzx)∈GA,∀(x,y)∈GA,Rz(zx+y)=x,\forall x\in X,\quad (R_zx,\ x-zR_zx)\in G_A, \qquad \forall (x,y)\in G_A,\quad R_z(zx+y)=x,∀x∈X,(Rz​x, x−zRz​x)∈GA​,∀(x,y)∈GA​,Rz​(zx+y)=x,

if such a map exists. If no such map exists, RzR_zRz​ is defined to be the zero map. This chosen or default map is plusResolvent⁡(A,z)\operatorname{plusResolvent}(A,z)plusResolvent(A,z).

QAlgorithms.Papers.Independent.Intakenext100.P35.point

For any real numbers aaa and bbb, point(a,b)\mathrm{point}(a,b)point(a,b) is the complex number a+iba+iba+ib, where i2=−1i^2=-1i2=−1.

QAlgorithms.Papers.Independent.Intakenext100.P35.positiveVolume

The measure positiveVolume\mathrm{positiveVolume}positiveVolume is Lebesgue measure on R\mathbb RR restricted to (0,∞)(0,\infty)(0,∞): for every measurable set A⊆RA\subseteq\mathbb RA⊆R, its measure is the Lebesgue measure of A∩(0,∞)A\cap(0,\infty)A∩(0,∞).

QAlgorithms.Papers.Independent.Intakenext100.P35.powerPull

For every real number rrr, every function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, and every z∈Cz\in\mathbb Cz∈C, this defines the value f(zr)f(z^r)f(zr), where the real exponent rrr is regarded as a complex number and the power is the total complex-power operation. There are no restrictions on rrr, zzz, or fff; in particular, z=0z=0z=0 and nonpositive exponents are included.

QAlgorithms.Papers.Independent.Intakenext100.P35.qKernel

For any function g:C→Cg:\mathbb C\to\mathbb Cg:C→C, any z∈Cz\in\mathbb Cz∈C, and any real a,ba,ba,b, this defines

qg(z;a,b)=a g(a+ib)(a−ib+z)2,q_g(z;a,b)=\frac{a\,g(a+ib)}{(a-ib+z)^2},qg​(z;a,b)=(a−ib+z)2ag(a+ib)​,

where real numbers are regarded as complex numbers. No positivity, regularity, or nonvanishing assumptions are imposed. Division is total complex division, so the value is 000 when a−ib+z=0a-ib+z=0a−ib+z=0.

QAlgorithms.Papers.Independent.Intakenext100.P35.qsKernel

For every real number sss, every function g:C→Cg:\mathbb C\to\mathbb Cg:C→C, every z∈Cz\in\mathbb Cz∈C, and all real numbers a,ba,ba,b, this defines the complex number

as g(a+ib)(a−ib+z)s+1.\frac{a^s\,g(a+ib)}{(a-ib+z)^{s+1}}.(a−ib+z)s+1asg(a+ib)​.

The numerator uses the total real-power operation for asa^sas, followed by its embedding into C\mathbb CC, while the denominator uses complex exponentiation with exponent s+1s+1s+1. No positivity, differentiability, or nonvanishing assumptions are imposed: aaa may be nonpositive, the base of the denominator may be zero, and division by zero has value zero.

QAlgorithms.Papers.Independent.Intakenext100.P35.resolventSymbol

For every λ,z∈C\lambda,z\in\mathbb Cλ,z∈C, this function has value (λ+z)−1(\lambda+z)^{-1}(λ+z)−1. Complex inversion is total, with 0−1=00^{-1}=00−1=0, so its value at z=−λz=-\lambdaz=−λ is zero.

QAlgorithms.Papers.Independent.Intakenext100.P35.scalarCayley

For every complex number zzz, this defines

scalarCayley⁡(z)=z−1z+1.\operatorname{scalarCayley}(z)=\frac{z-1}{z+1}.scalarCayley(z)=z+1z−1​.

Division is the total complex division operation, so the definition includes z=−1z=-1z=−1 and gives scalarCayley⁡(−1)=0\operatorname{scalarCayley}(-1)=0scalarCayley(−1)=0.

QAlgorithms.Papers.Independent.Intakenext100.P35.sinhQuotient

For every complex number zzz and real number ttt, this function equals zzz when t=0t=0t=0, and equals sinh⁡(zt)/t\sinh(zt)/tsinh(zt)/t when t≠0t\ne0t=0, with ttt embedded in C\mathbb CC. Both positive and negative real values of ttt are included.

QAlgorithms.Papers.Independent.Intakenext100.P35.supSize

For any set S⊆CS\subseteq\mathbb CS⊆C and any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, supSize(S,f)\mathrm{supSize}(S,f)supSize(S,f) is sup⁡z∈S∥f(z)∥\sup_{z\in S}\|f(z)\|supz∈S​∥f(z)∥, taken in the extended nonnegative reals [0,∞][0,\infty][0,∞]. The supremum is 000 when SSS is empty and may be ∞\infty∞ when the values are unbounded.

QAlgorithms.Papers.Independent.Intakenext100.P35.vSize

For every real number sss and every function g:C→Cg:\mathbb C\to\mathbb Cg:C→C, this defines the extended nonnegative real number

∫(0,∞)∫Ras(a2+b2)(s+1)/2 ∣g(a+ib)∣ db da,\int_{(0,\infty)}\int_{\mathbb R} \frac{a^{s}}{(a^{2}+b^{2})^{(s+1)/2}}\,|g(a+ib)|\,db\,da,∫(0,∞)​∫R​(a2+b2)(s+1)/2as​∣g(a+ib)∣dbda,

where the powers are real powers, both integrals are nonnegative Lebesgue integrals, and the outer measure is Lebesgue measure restricted to (0,∞)(0,\infty)(0,∞). No restriction on sss and no measurability or finiteness assumption on ggg is imposed; the value may be +∞+\infty+∞. On the integration domain, aaa and a2+b2a^{2}+b^{2}a2+b2 are strictly positive.

QAlgorithms.Papers.Independent.Intakenext100.P35.wSize

For any function g:C→Cg:\mathbb C\to\mathbb Cg:C→C, this defines

wSize⁡(g)=∫(0,∞) ⁣−ess sup⁡b∈R∥g(a+ib)∥ da∈[0,∞].\operatorname{wSize}(g)= \int_{(0,\infty)}^{\!-} \operatorname*{ess\,sup}_{b\in\mathbb R}\|g(a+ib)\|\,da \quad\in[0,\infty].wSize(g)=∫(0,∞)−​b∈Resssup​∥g(a+ib)∥da∈[0,∞].

The essential supremum is with respect to real Lebesgue measure in bbb, and the outer integral is the nonnegative Lebesgue integral with respect to real Lebesgue measure restricted to a>0a>0a>0. Norms are embedded into the extended nonnegative reals; no measurability assumption is included, and the value may be infinite.

QAlgorithms.Papers.Independent.Intakenext100.P35.weakBValue

Let XXX be a complex Banach space, let AAA have a closed complex linear, single-valued graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain, let f:C→Cf:\mathbb C\to\mathbb Cf:C→C be any function, let x∈Xx\in Xx∈X, and let ℓ:X→C\ell:X\to\mathbb Cℓ:X→C be a bounded complex linear functional. For each zzz, let RA(z)R_A(z)RA​(z) be a classically chosen bounded complex linear map RRR satisfying (Ru,u−zRu)∈GA(Ru,u-zRu)\in G_A(Ru,u−zRu)∈GA​ for every uuu and R(zu+v)=uR(zu+v)=uR(zu+v)=u whenever (u,v)∈GA(u,v)\in G_A(u,v)∈GA​, or zero if no such map exists. Let f∞f_\inftyf∞​ be a classically chosen complex limit of f(t)f(t)f(t) as real t→+∞t\to+\inftyt→+∞, if such a limit exists, and zero otherwise. Then

weakBValue⁡(A,f,x,ℓ)=f∞ℓ(x)−2π∫0∞∫Ra ℓ ⁣(RA(a−ib)(RA(a−ib)x))f′(a+ib) db da.\operatorname{weakBValue}(A,f,x,\ell) =f_\infty\ell(x)-\frac2\pi \int_0^\infty\int_{\mathbb R} a\,\ell\!\left(R_A(a-ib)\bigl(R_A(a-ib)x\bigr)\right)f'(a+ib)\,db\,da.weakBValue(A,f,x,ℓ)=f∞​ℓ(x)−π2​∫0∞​∫R​aℓ(RA​(a−ib)(RA​(a−ib)x))f′(a+ib)dbda.

The outer measure is Lebesgue measure restricted to (0,∞)(0,\infty)(0,∞), and the inner measure is Lebesgue measure on R\mathbb RR. No measurability, differentiability, or integrability assumption is imposed by this definition: the derivative is zero where complex differentiability fails, and each Bochner integral is the total integral operation, returning zero when its integrand is not integrable.

QAlgorithms.Papers.Independent.Intakenext100.P35.weakKernel

Let XXX be a complex Banach space, let AAA have a closed complex linear, single-valued graph GA⊆X×XG_A\subseteq X\times XGA​⊆X×X with dense domain, let f:C→Cf:\mathbb C\to\mathbb Cf:C→C be any function, let x∈Xx\in Xx∈X, let ℓ:X→C\ell:X\to\mathbb Cℓ:X→C be a bounded complex linear functional, and let a,b∈Ra,b\in\mathbb Ra,b∈R. For z∈Cz\in\mathbb Cz∈C, let RA(z)R_A(z)RA​(z) be a classically chosen bounded complex linear map RRR satisfying (Ru,u−zRu)∈GA(Ru,u-zRu)\in G_A(Ru,u−zRu)∈GA​ for every u∈Xu\in Xu∈X and R(zu+v)=uR(zu+v)=uR(zu+v)=u for every (u,v)∈GA(u,v)\in G_A(u,v)∈GA​, or zero if no such map exists. The defined kernel is

weakKernel⁡(A,f,x,ℓ,a,b)=a ℓ ⁣(RA(a−ib)(RA(a−ib)x))f′(a+ib).\operatorname{weakKernel}(A,f,x,\ell,a,b) =a\,\ell\!\left(R_A(a-ib)\bigl(R_A(a-ib)x\bigr)\right)f'(a+ib).weakKernel(A,f,x,ℓ,a,b)=aℓ(RA​(a−ib)(RA​(a−ib)x))f′(a+ib).

There is no restriction on the signs of a,ba,ba,b, and no differentiability hypothesis on fff; the derivative denotes the total derivative operation, which returns zero where fff is not complex differentiable.

QAlgorithms.Papers.Independent.Intakenext100.P35.zeroValue

For any function f:C→Cf:\mathbb C\to\mathbb Cf:C→C, zeroValue(f)\mathrm{zeroValue}(f)zeroValue(f) is the complex limit of f(t)f(t)f(t) as the real variable t→0t\to0t→0 through strictly positive values, with ttt regarded as a complex number, if this limit exists; otherwise it is 000. No condition on f(0)f(0)f(0) is imposed.

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