Alphabeta Math
ExampleConstruction: AI-generatedVerification: AI-generatedPipeline-generatedprecheck pass
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.

Finite-time linearization bound for the planar pendulum with explicit constants

Example

Let l=g=1 and consider the planar pendulum field

F:R2→R2,F(θ,ω)=(ω, −sin⁡θ),

Let A:R2→R2 be the linear map A(θ,ω):=(ω,−θ), and let A=(01−10)∈M2(R) be its standard-coordinate matrix; use the matrix action of A from Finite-time linearization error bound: quadratic remainder, exponential growth bound and the variation-of-constants estimate. On the closed ball B‾(0,12) one has

∥F(z)−Az∥2≤112∥z∥22,∥F(z)∥2≤∥z∥2,

so the linearization theorem applies with r=12, C=112, L=1. For the datum z0=(1100,0) and T=1 one has δeLT=e100<3100<120<12, and the theorem bounds the linearization error by

∥z(t)−y(t)∥2≤δ212∫01e(1−s)e2s ds<10−4(0≤t≤1),

where y solves y′=Ay, y(0)=z0; in particular the linear solution stays within 10−4 of the nonlinear one on [0,1].

Facts & Assumptions

Given: The field F, the matrix A, the radius r=1/2, the datum z0=(1/100,0) and the horizon T=1.

[L1]

The linearization theorem (Finite-time linearization error bound: quadratic remainder, exponential growth bound and the variation-of-constants estimate): if U⊆Rn is open, 0∈U, F:U→Rn is continuous and locally Lipschitz in the state variable, r,T>0, C,L≥0, B‾(0,r)⊆U, δ=∥z0∥2>0, and ∥F(z)−Az∥2≤C∥z∥22 and ∥F(z)∥2≤L∥z∥2 for ∥z∥2≤r and δeLT<r, then the nonlinear solution exists on [0,T] and ∥z(t)−y(t)∥2≤Cδ2∫0te∥A∥(t−s)e2Lsds for t∈[0,T]. [given]

[L2]

Taylor remainder bound: if f has derivatives through order n+1 on the closed interval between a and x, with endpoint continuity, and ∣f(n+1)∣≤M there, then ∣Rn,af(x)∣≤M∣x−a∣n+1/(n+1)! (A uniform derivative bound gives a uniform Taylor remainder bound); with f=sin⁡, n=2, a=0 and M=1 this gives ∣sin⁡θ−θ∣≤∣θ∣3/6, since ∣sin⁡′′′∣=∣cos⁡∣≤1 (The derivatives of sine and cosine are cosine and minus sine, Signs, monotonicity intervals, and ranges of sine and cosine). [given]

[L3]

Operator norm: ∥A∥=sup⁡∥u∥2=1∥Au∥2 (The operator norm as the least bound and as the unit-sphere or unit-ball supremum); for A=(01−10) and u=(u1,u2) one has Au=(u2,−u1), so ∥Au∥2=∥u∥2 and ∥A∥=1. [given]

Verification

technique · direct; estimate the two remainder terms of the pendulum field on the ball and substitute the explicit constants into the theorem
1.1givenL2algebra

Quadratic remainder: for z=(θ,ω) with ∥z∥2≤1/2, F(z)−Az=(0, −sin⁡θ)−(0,−θ)=(0, θ−sin⁡θ), so ∥F(z)−Az∥2=∣sin⁡θ−θ∣≤∣θ∣3/6≤12∣θ∣2/6≤112∥z∥22 by [L2] and ∣θ∣≤∥z∥2≤1/2; hence C=1/12 is admissible.

1.2givenL1L3L4L6algebra

Regularity and growth: U=R2 is open and contains B‾(0,1/2). The differentiable sine and coordinate functions make F continuous. For z=(θ,ω) and w=(η,ν), [L6] and [L4] give ∣sin⁡θ−sin⁡η∣≤∣θ−η∣, hence ∥F(z)−F(w)∥22=(ω−ν)2+(sin⁡θ−sin⁡η)2≤∥z−w∥22. Thus F is globally 1-Lipschitz and in particular locally Lipschitz in the state variable. Growth bound: if θ≠0, the mean value theorem (The mean value theorem, as the case g(x)=x of Cauchy's: for f continuous on [a,b] with a<b and differentiable on (a,b) there is c∈(a,b) with f(b)−f(a)=f′(c)(b−a)) applied to sin⁡ between 0 and θ, with ∣cos⁡∣≤1 from [L4], gives ∣sin⁡θ∣=∣sin⁡θ−sin⁡0∣≤∣θ∣; this inequality is immediate at θ=0. Thus ∥F(z)∥22=ω2+sin⁡2θ≤ω2+θ2=∥z∥22, so L=1 is admissible. With δ=1/100 and T=1, [L4] gives δeLT=e/100<3/100<1/20<1/2=r, so the hypotheses of [L1] hold.

2.1step 1.1step 1.2L1L3L4L5algebra∎

Substitute: by [L1] and [L3], for 0≤t≤1 the error is at most 112δ2∫0te(1−s)e2sds≤112δ2∫01e(1−s)e2sds, since et−s≤e1−s for s≤t≤1 and the resulting integrand is positive by [L4]. The addition formula gives e(1−s)e2s=e1+s; since 0≤s≤1, monotonicity gives e1+s≤e2<9 by [L4]. Monotonicity of the integral [L5] therefore yields ∫01e1+sds≤9, and the error is at most 112⋅10−4⋅9=34⋅10−4<10−4.

Depends on

Dependency tree · 2 levels

Sources