Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Existence and uniqueness of the [m/n][m/n][m/n] Padé approximant

Proved
Pade.pade_existence_uniqueness

by LG · Sep 21, 2026 · Mathlib 0df444a (Lean v4.33.1)

analysisapproximation-theorynumerical-analysispower-series

Fundamental theorem of Padé approximation (Frobenius). Let FFF be a field, let f∈F[[x]]f \in F[[x]]f∈F[[x]] be a formal power series and let m,n≥0m, n \ge 0m,n≥0 be integers. Call (P,Q)(P,Q)(P,Q) a Padé pair of type [m/n][m/n][m/n] for fff when Q≠0Q \ne 0Q=0, deg⁡P≤m\deg P \le mdegP≤m, deg⁡Q≤n\deg Q \le ndegQ≤n and Qf−P≡0(modxm+n+1)Qf - P \equiv 0 \pmod{x^{m+n+1}}Qf−P≡0(modxm+n+1). Then:

(E)a Padeˊ pair of type [m/n] for f exists;\textbf{(E)}\quad \text{a Padé pair of type } [m/n] \text{ for } f \text{ exists;}(E)a Padeˊ pair of type [m/n] for f exists; (U)if (P1,Q1) and (P2,Q2) are Padeˊ pairs of type [m/n] for f, then P1Q2=P2Q1.\textbf{(U)}\quad \text{if } (P_1,Q_1) \text{ and } (P_2,Q_2) \text{ are Padé pairs of type } [m/n] \text{ for } f, \text{ then } P_1Q_2 = P_2Q_1 .(U)if (P1​,Q1​) and (P2​,Q2​) are Padeˊ pairs of type [m/n] for f, then P1​Q2​=P2​Q1​.

Part (E) holds for every fff, mmm and nnn, with no condition on fff; part (U) is the precise sense in which "the Padé approximant is unique as a formal power series for the given mmm and nnn": two solutions need not be equal as pairs — they may differ by a common polynomial factor or a scalar — but they cross-multiply, hence define the same element of the field of rational functions F(x)F(x)F(x).

Together the two parts make the Padé table ([m/n]f)m,n≥0\big([m/n]_f\big)_{m,n\ge0}([m/n]f​)m,n≥0​ a well-defined array, which is the foundation for every identity relating its entries and for every algorithm (extended Euclid, Wynn's ε\varepsilonε-algorithm) that computes them.

Preamble
import Mathlib
import Definitions.Def_pade_approximant_def
open Polynomial
Formal statement
namespace Pade

theorem pade_existence_uniqueness {F : Type*} [Field F] (f : PowerSeries F) (m n : ℕ) :
    (∃ P Q : Polynomial F, IsPadeApproximant f m n P Q) ∧
      (∀ P₁ Q₁ P₂ Q₂ : Polynomial F, IsPadeApproximant f m n P₁ Q₁ →
        IsPadeApproximant f m n P₂ Q₂ → P₁ * Q₂ = P₂ * Q₁) := by sorry

end Pade
Source
Padé approximant, Wikipedia, revision oldid=1374746248, https://en.wikipedia.org/w/index.php?title=Pad%C3%A9_approximant&oldid=1374746248, section 'Definition' ('When it exists, the Padé approximant is unique as a formal power series for the given m and n.')
Read-back

What the Lean code literally says, in plain math · Aristotle (Harmonic)

Provenance note — this read-back is NOT blind and NOT independent. It was written by the same agent that drafted this proposal's Lean statements, at the explicit instruction of the proposal owner; no separate, blind auditor produced it. It is therefore not independent testimony and must not be mistaken for it: its author already knew what the code was intended to say, which is exactly the bias a read-back is normally meant to exclude. Please check it directly against the Lean code and against the source material.

Throughout, FFF is an arbitrary type equipped with a field structure (no assumption on its characteristic, cardinality, or ordering), fff is an arbitrary formal power series in F[[x]]F[[x]]F[[x]], and m,nm, nm,n are arbitrary natural numbers — so m=0m = 0m=0, n=0n = 0n=0 and f=0f = 0f=0 are all included.

The statement uses an auxiliary predicate on a pair of polynomials P,Q∈F[x]P, Q \in F[x]P,Q∈F[x], which unfolds to the conjunction of: Q≠0Q \neq 0Q=0; deg⁡P≤m\deg P \le mdegP≤m; deg⁡Q≤n\deg Q \le ndegQ≤n (degrees valued in N∪{−∞}\mathbb{N}\cup\{-\infty\}N∪{−∞}, so the zero polynomial meets every degree bound); and the vanishing of the coefficient of xkx^kxk in Qf−PQ f - PQf−P for every kkk with 0≤k≤m+n0 \le k \le m+n0≤k≤m+n, the products and differences being taken in F[[x]]F[[x]]F[[x]] after including PPP and QQQ into F[[x]]F[[x]]F[[x]]. Call such a pair admissible.

The theorem asserts a conjunction of two claims, for the given fff, mmm, nnn:

(1)∃ P,Q∈F[x] with (P,Q) admissible;\textbf{(1)} \quad \exists\, P, Q \in F[x] \ \text{with} \ (P,Q) \ \text{admissible};(1)∃P,Q∈F[x] with (P,Q) admissible; (2)∀ P1,Q1,P2,Q2∈F[x],(P1,Q1) admissible ∧ (P2,Q2) admissible ⟹ P1Q2=P2Q1,\textbf{(2)} \quad \forall\, P_1, Q_1, P_2, Q_2 \in F[x], \quad (P_1,Q_1) \ \text{admissible} \ \wedge\ (P_2,Q_2) \ \text{admissible} \ \Longrightarrow\ P_1 Q_2 = P_2 Q_1 ,(2)∀P1​,Q1​,P2​,Q2​∈F[x],(P1​,Q1​) admissible ∧ (P2​,Q2​) admissible ⟹ P1​Q2​=P2​Q1​,

the last equation being an identity of polynomials in F[x]F[x]F[x].

Three observations on the exact strength of the claim. Part (2) does not assert P1=P2P_1 = P_2P1​=P2​ or Q1=Q2Q_1 = Q_2Q1​=Q2​; it asserts only the cross-multiplication identity, which is the statement that the two formal quotients agree as elements of the field of rational functions whenever the denominators are nonzero. Part (2) is a statement about all admissible pairs, including pairs whose denominator vanishes at the origin, so no normalisation such as Q(0)=1Q(0) = 1Q(0)=1 is available to it. Finally, (1) and (2) are asserted together but independently: (1) guarantees the hypotheses of (2) are satisfiable, so (2) is not vacuous.

Human review
  • Endorsed by Shuze Chen · Sep 24, 2026

    Confirmed by the moderator at approval.

  • Endorsed by LG · Sep 24, 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