Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 7.2.3 — the self-dual program (SD) and its strictly complementary optima

Proved
MatousekLP.InteriorPoint.self_dual_embedding

by mikedeng1 · Oct 3, 2026 · Mathlib 0df444a (Lean v4.33.1)

interior-point-methodslinear-programmingp2o-batch-b23bp2o-gran-per-chapterp2o-plan-bookp2o-v1self-dual-embedding

Let AAA be a real m×nm\times nm×n matrix, b∈Rmb\in\mathbb{R}^mb∈Rm, c∈Rnc\in\mathbb{R}^nc∈Rn, k=n+m+1k=n+m+1k=n+m+1, and let

M0=(0A−b−AT0cbT−cT0),r=1+M01,M=(M0−rrT0),q=(0,…,0,k+1)∈Rk+1.M_0=\begin{pmatrix}0 & A & -b\\ -A^{T} & 0 & c\\ b^{T} & -c^{T} & 0\end{pmatrix},\quad r=\mathbf 1+M_0\mathbf 1,\quad M=\begin{pmatrix}M_0 & -r\\ r^{T} & 0\end{pmatrix},\quad q=(0,\dots,0,k+1)\in\mathbb{R}^{k+1}.M0​=​0−ATbT​A0−cT​−bc0​​,r=1+M0​1,M=(M0​rT​−r0​),q=(0,…,0,k+1)∈Rk+1.

Consider the linear program

(SD)maximize −qTv  subject to Mv≤q, v≥0,\text{(SD)}\qquad \text{maximize } -q^{T}v \ \text{ subject to } Mv\le q,\ v\ge 0,(SD)maximize −qTv  subject to Mv≤q, v≥0,

in v=(u,ϑ)v=(u,\vartheta)v=(u,ϑ) with u=(y,x,τ)∈Rku=(y,x,\tau)\in\mathbb{R}^{k}u=(y,x,τ)∈Rk, and write z=q−Mvz=q-Mvz=q−Mv for its slacks. Then:

  1. (SD) has a feasible solution, and its objective is bounded from above on the feasible set;
  2. every optimal solution v=(u,ϑ)v=(u,\vartheta)v=(u,ϑ) of (SD) has ϑ=0\vartheta=0ϑ=0, and its uuu-part (y,x,τ)(y,x,\tau)(y,x,τ) is a solution of the Goldman–Tucker system (GTS) Ax−τb≤0Ax-\tau b\le 0Ax−τb≤0, −ATy+τc≤0-A^{T}y+\tau c\le 0−ATy+τc≤0, bTy−cTx≤0b^{T}y-c^{T}x\le 0bTy−cTx≤0, x,y≥0x,y\ge 0x,y≥0, τ≥0\tau\ge 0τ≥0;
  3. if moreover vvv is strictly complementary (for each coordinate jjj, vj>0v_j>0vj​>0 or zj>0z_j>0zj​>0), then the resulting solution of (GTS) has τ>0\tau>0τ>0 or ρ=cTx−bTy>0\rho=c^{T}x-b^{T}y>0ρ=cTx−bTy>0.

Together with Lemma 7.2.2, this reduces solving (7.7) to finding a strictly complementary optimal solution of (SD), a program with an explicit interior starting point.

Formalization Note uuu is indexed by Fin m ⊕ Fin n ⊕ Unit and vvv by (Fin m ⊕ Fin n ⊕ Unit) ⊕ Unit; uY, uX, uTau, vU, vTheta extract yyy, xxx, τ\tauτ, uuu, ϑ\varthetaϑ. Optimality and boundedness are stated against every feasible point.

Preamble
import Mathlib
import Definitions.Def_MatousekLP_InteriorPoint_SelfDual
Formal statement
namespace MatousekLP.InteriorPoint

open Matrix

/-- Lemma 7.2.3 (Matoušek & Gärtner, p. 128). The linear program
(SD) `maximize −qᵀv subject to Mv ≤ q, v ≥ 0` built from `A, b, c` is feasible and bounded,
every optimal solution `v = (u, ϑ)` has `ϑ = 0`, and hence its `u`-part `u = (y, x, τ)` is a
solution of the Goldman–Tucker system (GTS). Moreover, every strictly complementary optimal
solution yields a solution of (GTS) with `τ > 0` or `ρ > 0`. -/
theorem self_dual_embedding {m n : ℕ} (A : Matrix (Fin m) (Fin n) ℝ) (b : Fin m → ℝ)
    (c : Fin n → ℝ) :
    (∃ v : VIdx m n → ℝ, IsSDFeasible A b c v) ∧
    IsSDBoundedAbove A b c ∧
    (∀ v : VIdx m n → ℝ, IsSDOptimal A b c v →
        vTheta v = 0 ∧ IsGTSSolution A b c (uX (vU v)) (uY (vU v)) (uTau (vU v))) ∧
    (∀ v : VIdx m n → ℝ, IsSDOptimal A b c v → IsStrictlyComplementary A b c v →
        0 < uTau (vU v) ∨ 0 < gtsSlack b c (uX (vU v)) (uY (vU v))) := by sorry

end MatousekLP.InteriorPoint
Source
Matoušek & Gärtner, Understanding and Using Linear Programming, Springer 2007, §7.2, p. 128, Lemma 7.2.3 (with M_0, r, M p. 127 and q, (SD) p. 128)
Human review
  • Endorsed by Shuze Chen · Oct 5, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 5, 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