Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 2.1, p. 6 — solutions of the homogeneous model (7) are complementary, and give an optimal pair (τ* > 0) or an infeasibility certificate (κ* > 0)

Proved
ConicQuadIPM.Homogeneous.lemma_2_1

by mikedeng1 · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

conic-optimizationhomogeneous-modelinfeasibility-certificateinterior-pointp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Let A∈Rm×nA \in \mathbb R^{m\times n}A∈Rm×n, b∈Rmb \in \mathbb R^mb∈Rm, c∈Rnc \in \mathbb R^nc∈Rn, let K⊆RnK \subseteq \mathbb R^nK⊆Rn be a pointed closed convex cone with dual cone K∗={s:sTx≥0 ∀x∈K}K_* = \{s : s^Tx \ge 0\ \forall x \in K\}K∗​={s:sTx≥0 ∀x∈K}, and consider the conic pair (P)(P)(P): min⁡{cTx:Ax=b, x∈K}\min\{c^Tx : Ax = b,\ x \in K\}min{cTx:Ax=b, x∈K} and (D)(D)(D): max⁡{bTy:ATy+s=c, s∈K∗}\max\{b^Ty : A^Ty + s = c,\ s \in K_*\}max{bTy:ATy+s=c, s∈K∗​}. Let (x∗,τ∗,y∗,s∗,κ∗)(x^*, \tau^*, y^*, s^*, \kappa^*)(x∗,τ∗,y∗,s∗,κ∗) be any solution of the homogeneous model (7):

Ax∗−bτ∗=0,ATy∗+s∗−cτ∗=0,−cTx∗+bTy∗−κ∗=0,x∗∈K, τ∗≥0, s∗∈K∗, κ∗≥0.Ax^* - b\tau^* = 0,\quad A^Ty^* + s^* - c\tau^* = 0,\quad -c^Tx^* + b^Ty^* - \kappa^* = 0,\quad x^* \in K,\ \tau^* \ge 0,\ s^* \in K_*,\ \kappa^* \ge 0.Ax∗−bτ∗=0,ATy∗+s∗−cτ∗=0,−cTx∗+bTy∗−κ∗=0,x∗∈K, τ∗≥0, s∗∈K∗​, κ∗≥0.

Then:

  1. (i) (x∗)Ts∗+τ∗κ∗=0(x^*)^Ts^* + \tau^*\kappa^* = 0(x∗)Ts∗+τ∗κ∗=0.
  2. (ii) If τ∗>0\tau^* > 0τ∗>0, then x∗/τ∗x^*/\tau^*x∗/τ∗ is an optimal solution of (P)(P)(P) and (y∗,s∗)/τ∗(y^*, s^*)/\tau^*(y∗,s∗)/τ∗ is an optimal solution of (D)(D)(D).
  3. (iii) If κ∗>0\kappa^* > 0κ∗>0, then at least one of the strict inequalities bTy∗>0b^Ty^* > 0bTy∗>0 (8) and cTx∗<0c^Tx^* < 0cTx∗<0 (9) holds; if bTy∗>0b^Ty^* > 0bTy∗>0 then (P)(P)(P) is infeasible, and if cTx∗<0c^Tx^* < 0cTx∗<0 then (D)(D)(D) is infeasible.

The homogeneous model always has the trivial solution, so the lemma is about what a solution with τ∗+κ∗>0\tau^* + \kappa^* > 0τ∗+κ∗>0 delivers: either a scaled optimal primal–dual pair or a certificate of infeasibility. This is what makes an interior-point method applied to (7) a complete solution method for (P)(P)(P) and (D)(D)(D).

Formalization Note Rn\mathbb R^nRn is Fin n → ℝ, uTvu^TvuTv is the dot product u ⬝ᵥ v, AxAxAx is A *ᵥ x and ATyA^TyATy is Aᵀ *ᵥ y; the paper's first coordinate is index 0. The standing assumption that KKK is a pointed closed convex cone is carried as the hypothesis IsPointedClosedConvexCone K. The paper's additional assumption that AAA has full row rank ("for convenience and without loss of generality", p. 4) is not imposed, so the statement is a (harmless) generalization. Division by τ∗\tau^*τ∗ is written as scalar multiplication by (τ∗)−1(\tau^*)^{-1}(τ∗)−1 and is only used under τ∗>0\tau^* > 0τ∗>0. Optimality and infeasibility are order and non-existence predicates over all feasible points.

Preamble
import Mathlib
import Definitions.Def_ConicQuadIPM_Homogeneous_Setting
Formal statement
namespace ConicQuadIPM.Homogeneous

open Matrix

/-- Lemma 2.1 (p. 6). Let `(x*, τ*, y*, s*, κ*)` be any solution of the homogeneous model (7).
Then i) `(x*)ᵀs* + τ*κ* = 0`; ii) if `τ* > 0`, `x*/τ*` is optimal for (P) and `(y*, s*)/τ*` is
optimal for (D); iii) if `κ* > 0`, then `bᵀy* > 0` or `cᵀx* < 0`; if `bᵀy* > 0` then (P) is
infeasible, and if `cᵀx* < 0` then (D) is infeasible. -/
theorem lemma_2_1 {m n : ℕ} (A : Matrix (Fin m) (Fin n) ℝ) (b : Fin m → ℝ) (c : Fin n → ℝ)
    (K : Set (Fin n → ℝ)) (hK : IsPointedClosedConvexCone K)
    (x : Fin n → ℝ) (τ : ℝ) (y : Fin m → ℝ) (s : Fin n → ℝ) (κ : ℝ)
    (h : HomFeasible A b c K x τ y s κ) :
    x ⬝ᵥ s + τ * κ = 0 ∧
    (0 < τ → PrimalOptimal A b c K (τ⁻¹ • x) ∧ DualOptimal A b c K (τ⁻¹ • y) (τ⁻¹ • s)) ∧
    (0 < κ → (0 < b ⬝ᵥ y ∨ c ⬝ᵥ x < 0) ∧ (0 < b ⬝ᵥ y → PrimalInfeasible A b K) ∧
      (c ⬝ᵥ x < 0 → DualInfeasible A c K)) := by sorry

end ConicQuadIPM.Homogeneous
Source
Andersen, Roos & Terlaky, On implementing a primal-dual interior-point method for conic quadratic optimization, Math. Program. (2003); authors' preprint of 18 Dec 2000, p. 6, Lemma 2.1 i)–iii) with (8), (9); proof pp. 6–7
Human review
  • Endorsed by Shuze Chen · Oct 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 7, 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