Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-generated
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

An inexact Newton residual gives a local error recurrence

Statement

For n=0, call a linear map invertible when it has a two-sided linear inverse, extending the positive-dimensional convention of Invertible Euclidean linear maps. Let n≥0 be an integer, let U⊆Rn be open, F:U→Rn be C1, and x∗∈U satisfy F(x∗)=0 with J∗:=DF(x∗) invertible. Suppose a closed Euclidean ball K=B‾2(x∗,R)⊆U with R>0 is convex and

∥J(u)−J(v)∥op≤L∥u−v∥2(u,v∈K),

where J(x):=DF(x) and ∥A∥op:=sup⁡∥v∥2=1∥Av∥2 for n≥1; for n=0 set the norm of the unique linear map to 0. Let ∥v∥∗:=∥J∗v∥2 and fix ηˉ∈[0,1). There are constants r0>0, CF>0, and C≥0 such that, whenever ∥x−x∗∥2≤r0, 0≤η≤ηˉ, and p,r∈Rn satisfy

J(x)p+F(x)=r,∥r∥2≤η∥F(x)∥2,

then J(x) is invertible, ∥F(x)∥2≤CF∥x−x∗∥∗, and

∥x+p−x∗∥∗≤(η+C∥x−x∗∥∗)∥x−x∗∥∗.

Facts & Assumptions

Given: U,F,x∗,J,K,L,ηˉ,x,η,p,r as in the Statement.

[F1]

Continuously differentiable maps, local inverses, and local diffeomorphisms: C1 means total differentiability at every point and continuity of the derivative-matrix entries.

[F2]

The Jacobian matrix of partial derivatives and the gradient in the scalar-valued case: the Jacobian matrix records the first partial derivatives in the standard coordinates.

[F3]

Newton maps are uniform contractions near a point with invertible derivative: if DF(x∗) is invertible for a C1 map, then on a sufficiently small closed ball every DF(x) is invertible and its inverse has a uniform operator-norm bound.

[F4]

Vector Newton remainder under a Lipschitz Jacobian: for x,x∗∈K, F(x∗)=0, and J L-Lipschitz on K, ∥F(x)−J(x)(x−x∗)∥2≤(L/2)∥x−x∗∥22.

[F5]

Invertible Euclidean linear maps: for n≥1, an invertible square linear map has a two-sided linear inverse; the Statement explicitly extends this convention to n=0.

[F6]

The Euclidean inner product ⟨x,y⟩=∑k<nxkyk on Rn: ∥v∥2=⟨v,v⟩ is the Euclidean norm.

Proof

Given: The hypotheses and notation in the Statement.

1.1givenF6algebra

If n=0, the space R0 is a singleton by [F6]. Its unique map A is linear and satisfies A(A(0))=0, so taking B=A verifies both inverse identities in the Statement's convention. Thus J(x)=J∗=A is invertible, and x=x∗=p=r=F(x)=0 in this space. Take r0=R, CF=1, and C=0. Every claimed estimate is 0≤0, so the assertion holds. For the remainder assume n≥1, so the imported positive-dimensional Newton lemmas and operator-norm comparisons apply.

1.2givenF1F2F3F5F6choose

Put K∗:=∥J∗−1∥op. The inverse identity gives ∥e∥2=∥J∗−1J∗e∥2≤K∗∥e∥∗, and the operator-norm definition gives ∥e∥∗≤∥J∗∥op∥e∥2. By [F3], choose a smaller radius r1≤R and M>0 such that every J(x) on B‾2(x∗,r1) is invertible and ∥J(x)−1∥op≤M. Decrease r1 if necessary so r1≤1, and set r0:=r1.

2.1givenF4F6step 1.2algebra

Let e:=x−x∗. By [F4] and ∥J(x)−J∗∥op≤L∥e∥2, ∥F(x)−J∗e∥2≤∥F(x)−J(x)e∥2+∥(J(x)−J∗)e∥2≤3L2∥e∥22. Since ∥e∥2≤K∗∥e∥∗ and ∥e∥∗≤∥J∗∥opr0, this gives ∥F(x)∥2≤(1+3L2K∗2∥J∗∥opr0)∥e∥∗=:CF∥e∥∗.

3.1step 1.2step 2.1givenalgebra

The linear residual equation and steps 1.2–2.1 imply ∥p∥2≤M∥J(x)p∥2≤M(1+η)∥F(x)∥2≤M(1+ηˉ)CF∥e∥∗.

4.1givenF4step 1.2step 3.1algebra∎

Put e+:=x+p−x∗. From J(x)p=−F(x)+r, J∗e+=J∗e−F(x)+r+(J∗−J(x))p. Step 1.2 and [F4] give ∥F(x)−J∗e∥2≤(3L/2)∥e∥22≤(3L/2)K∗2∥e∥∗2, so ∥r∥2≤η∥F(x)∥2≤η∥e∥∗+(3ηLK∗2/2)∥e∥∗2. Also ∥J∗−J(x)∥op≤LK∗∥e∥∗ and step 3.1 bounds ∥p∥2≤M(1+ηˉ)CF∥e∥∗. The triangle inequality and η≤ηˉ yield ∥e+∥∗≤η∥e∥∗+(3(1+ηˉ)LK∗22+LK∗M(1+ηˉ)CF)∥e∥∗2. Set C equal to this nonnegative coefficient; this is the claimed recurrence, and step 1.2 supplies invertibility.

Depends on

Used by

Dependency tree · two levels

40 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources