Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Fixed-contact localization boxes shrink linearly with the rate gap

Open
PGDRevision20261113.localization_linear_shrinkage

by Somehow · Nov 13, 2026 · Mathlib 0df444a (Lean v4.33.1)

bregman-geometryconvex-optimizationformalization

For a positive kernel kc(t)k_c(t)kc​(t), error E=(pkc−A)/(pkc+B)E=(pk_c-A)/(pk_c+B)E=(pkc​−A)/(pkc​+B), and 0<R<10<R<10<R<1, write

Lt(R)=A(t)−RB(t)1+R,Ut(R)=A(t)+RB(t)1−R.\mathcal L_t(R)=\frac{A(t)-RB(t)}{1+R},\quad \mathcal U_t(R)=\frac{A(t)+RB(t)}{1-R}.Lt​(R)=1+RA(t)−RB(t)​,Ut​(R)=1−RA(t)+RB(t)​.

For fixed ordered contacts t1,t2,t3t_1,t_2,t_3t1​,t2​,t3​ with signs si=σ(−1)i−1s_i=\sigma(-1)^{i-1}si​=σ(−1)i−1, let

βi={log⁡Uti(U)−log⁡Uti(L),si=1,log⁡Lti(L)−log⁡Lti(U),si=−1.\beta_i=\begin{cases}\log\mathcal U_{t_i}(U)-\log\mathcal U_{t_i}(L),&s_i=1,\\ \log\mathcal L_{t_i}(L)-\log\mathcal L_{t_i}(U),&s_i=-1.\end{cases}βi​={logUti​​(U)−logUti​​(L),logLti​​(L)−logLti​​(U),​si​=1,si​=−1.​

Fix the contact states, signs, compact parameter interval, and a kernel satisfying the score-ordering hypotheses. Fix 0<r0≤r1<10<r_0\le r_1<10<r0​≤r1​<1, and require A(ti)>r1B(ti)A(t_i)>r_1B(t_i)A(ti​)>r1​B(ti​) at every negative contact.

There is a constant K≥0K\ge0K≥0, independent of L,UL,UL,U, such that for every r0≤L≤U≤r1r_0\le L\le U\le r_1r0​≤L≤U≤r1​, each βi\beta_iβi​, both exponent radii d−,d+d_-,d_+d−​,d+​, and the logarithmic scale radius are nonnegative and at most

K(U−L).K(U-L).K(U−L).

This states the linear shrinking assertion with uniform positivity of all relevant band denominators. The contacts and interval remain fixed while the two rate levels vary.

Related results. Parameter localization (Theorem 7.1) · Error-band equivalence

Preamble
import Definitions.Def_PGDRevision20261113_NumericalCertificates
open PGDRevision20261113
Formal statement
theorem PGDRevision20261113.localization_linear_shrinkage : LocalizationLinearShrinkageClaim := by sorry
Source
Author manuscript, Exact Residual Bounds and Optimal Geometry Selection for Proximal Gradient Methods, revised reading edition, 13 November 2026; Section 7, paragraph following the rational scale enclosure; sections/07_quantitative_certificates.tex, all beta_i=O(U-L).

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