Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 1 — for λ0/λ2>M\sqrt{\lambda_0/\lambda_2}>Mλ0​/λ2​​>M the perspective relaxation dominates the Big-M and PR(∞) relaxations

Proved
L0BnB.Strength.proposition_1

by mikedeng1 · Oct 8, 2026 · Mathlib 0df444a (Lean v4.33.1)

mixed-integer-programmingp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1perspective-relaxationrelaxation-strengthsparse-regression

Let X∈Rn×pX\in\mathbb R^{n\times p}X∈Rn×p, y∈Rny\in\mathbb R^ny∈Rn, λ0,λ2>0\lambda_0,\lambda_2>0λ0​,λ2​>0 and M>0M>0M>0. Let VPR(M)V_{PR(M)}VPR(M)​ be the optimal value of the reduced perspective relaxation (5),

VPR(M)=inf⁡∥β∥∞≤MF(β),F(β)=12∥y−Xβ∥22+∑i∈[p]ψ(βi;λ0,λ2,M),V_{PR(M)} = \inf_{\|\beta\|_\infty\le M} F(\beta),\qquad F(\beta) = \tfrac12\|y-X\beta\|_2^2 + \sum_{i\in[p]}\psi(\beta_i;\lambda_0,\lambda_2,M),VPR(M)​=∥β∥∞​≤Minf​F(β),F(β)=21​∥y−Xβ∥22​+i∈[p]∑​ψ(βi​;λ0​,λ2​,M),

let VB(M)V_{B(M)}VB(M)​ be the optimal value of the interval relaxation of the Big-M formulation (2), and let VPR(∞)=inf⁡β∈RpG(β)V_{PR(\infty)} = \inf_{\beta\in\mathbb R^p}G(\beta)VPR(∞)​=infβ∈Rp​G(β) with G(β)=12∥y−Xβ∥22+∑iψ1(βi;λ0,λ2)G(\beta) = \tfrac12\|y-X\beta\|_2^2 + \sum_i \psi_1(\beta_i;\lambda_0,\lambda_2)G(β)=21​∥y−Xβ∥22​+∑i​ψ1​(βi​;λ0​,λ2​) as in (6). Let β∗\beta^*β∗ be an optimal solution to (5), and h(λ0,λ2,M)=λ0/M+λ2M−2λ0λ2h(\lambda_0,\lambda_2,M) = \lambda_0/M + \lambda_2M - 2\sqrt{\lambda_0\lambda_2}h(λ0​,λ2​,M)=λ0​/M+λ2​M−2λ0​λ2​​. If λ0/λ2>M\sqrt{\lambda_0/\lambda_2} > Mλ0​/λ2​​>M, then

VPR(M)≥VB(M)+λ2(M∥β∗∥1−∥β∗∥22),(7)V_{PR(M)} \ge V_{B(M)} + \lambda_2\big(M\|\beta^*\|_1 - \|\beta^*\|_2^2\big), \tag{7}VPR(M)​≥VB(M)​+λ2​(M∥β∗∥1​−∥β∗∥22​),(7) VPR(M)≥VPR(∞)+h(λ0,λ2,M) ∥β∗∥1.(8)V_{PR(M)} \ge V_{PR(\infty)} + h(\lambda_0,\lambda_2,M)\,\|\beta^*\|_1. \tag{8}VPR(M)​≥VPR(∞)​+h(λ0​,λ2​,M)∥β∗∥1​.(8)

Both added terms are nonnegative in this regime, so the perspective relaxation is at least as tight as both the Big-M relaxation and PR(∞)\mathrm{PR}(\infty)PR(∞), with explicit gaps that depend on the relaxation's own optimal solution. This is the theoretical basis for using PR(M)\mathrm{PR}(M)PR(M) with a tight, valid MMM in branch-and-bound.

Formalization Note β∗\beta^*β∗ is a binder with the hypotheses ∥β∗∥∞≤M\|\beta^*\|_\infty \le M∥β∗∥∞​≤M and F(β∗)≤F(β)F(\beta^*) \le F(\beta)F(β∗)≤F(β) for all feasible β\betaβ. The regime is written M<λ0/λ2M < \sqrt{\lambda_0/\lambda_2}M<λ0​/λ2​​.

Preamble
import Mathlib
import Definitions.Def_L0BnB_Strength_Penalties
import Definitions.Def_L0BnB_Strength_Relaxations
Formal statement
namespace L0BnB.Strength

/-- Proposition 1, p. 7: for `√(λ₀/λ₂) > M` and `β*` an optimal solution of (5),
`V_{PR(M)} ≥ V_{B(M)} + λ₂ (M ‖β*‖₁ − ‖β*‖₂²)` (7) and
`V_{PR(M)} ≥ V_{PR(∞)} + h(λ₀, λ₂, M) ‖β*‖₁` (8). -/
theorem proposition_1 {n p : ℕ} (X : Matrix (Fin n) (Fin p) ℝ) (y : Fin n → ℝ) (lam0 lam2 M : ℝ)
    (hlam0 : 0 < lam0) (hlam2 : 0 < lam2) (hM : 0 < M)
    (hregime : M < Real.sqrt (lam0 / lam2))
    (βs : Fin p → ℝ) (hβs_feas : ∀ i, |βs i| ≤ M)
    (hβs_opt : ∀ β : Fin p → ℝ, (∀ i, |β i| ≤ M) → L0BnB.Reduced.F X y lam0 lam2 M βs ≤ L0BnB.Reduced.F X y lam0 lam2 M β) :
    L0BnB.Reduced.VPR X y lam0 lam2 M ≥
        VB X y lam0 lam2 M + lam2 * (M * ∑ i, |βs i| - ∑ i, βs i ^ 2) ∧
      L0BnB.Reduced.VPR X y lam0 lam2 M ≥ VPRinf X y lam0 lam2 + hGap lam0 lam2 M * ∑ i, |βs i| := by sorry

end L0BnB.Strength
Source
Hazimeh, Mazumder, Saab, Sparse Regression at Scale: Branch-and-Bound rooted in First-Order Optimization, arXiv:2004.06152v2, p. 7, Proposition 1; App. A, Proof of Proposition 1, pp. 28–29
Human review
  • Endorsed by Shuze Chen · Oct 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 8, 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