Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proof of Theorem 8 — (xᵏ − x*)ᵀ(sᵏ − s*) + (τᵏ − τ*)(κᵏ − κ*) = 0 and xᵏᵀs* + sᵏᵀx* + κᵏτ* = (n+1)μᵏ = (n+1)θᵏ

Proved
SelfDualLP.Output.orthogonality_identity

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

interior-point-methodslinear-programmingp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1self-dual-embedding

Work with (HLP) under the choice (7). Let (y,x,τ,θ,s,κ)∈Fh(y,x,\tau,\theta,s,\kappa)\in\mathcal F_h(y,x,τ,θ,s,κ)∈Fh​ be a feasible point, and let (y∗,x∗,τ∗,θ∗,s∗,κ∗)(y^*,x^*,\tau^*,\theta^*,s^*,\kappa^*)(y∗,x∗,τ∗,θ∗,s∗,κ∗) be a self-complementary (optimal) solution of (HLP) with κ∗=0\kappa^*=0κ∗=0. Then

(x−x∗)T(s−s∗)+(τ−τ∗)(κ−κ∗)=0,(x-x^*)^T(s-s^*)+(\tau-\tau^*)(\kappa-\kappa^*)=0,(x−x∗)T(s−s∗)+(τ−τ∗)(κ−κ∗)=0,

and this can be rewritten as

xTs∗+sTx∗+κτ∗=(n+1)μ=(n+1)θ,μ=xTs+τκn+1.x^Ts^*+s^Tx^*+\kappa\tau^*=(n+1)\mu=(n+1)\theta,\qquad \mu=\frac{x^Ts+\tau\kappa}{n+1}.xTs∗+sTx∗+κτ∗=(n+1)μ=(n+1)θ,μ=n+1xTs+τκ​.

In the proof of Theorem 8 it is applied with the iterate zkz^kzk in place of (y,x,τ,θ,s,κ)(y,x,\tau,\theta,s,\kappa)(y,x,τ,θ,s,κ); dropping the nonnegative terms xTs∗x^Ts^*xTs∗ and sTx∗s^Tx^*sTx∗ gives κkτ∗≤(n+1)μk\kappa^k\tau^*\le(n+1)\mu^kκkτ∗≤(n+1)μk.

Formalization Note The first equality holds for any two points of Fh\mathcal F_hFh​. The rewritten form uses that the second point is optimal (so θ∗=0\theta^*=0θ∗=0 and (x∗)Ts∗+τ∗κ∗=0(x^*)^Ts^*+\tau^*\kappa^*=0(x∗)Ts∗+τ∗κ∗=0 by Theorem 2 (iv)) and that κ∗=0\kappa^*=0κ∗=0; these are exactly the hypotheses of the statement.

Preamble
import Mathlib
import Definitions.Def_SelfDualLP_Complexity_LPData
import Definitions.Def_SelfDualLP_Output_HLP
import Definitions.Def_SelfDualLP_Output_Neighborhood
import Definitions.Def_SelfDualLP_Output_PCSequence

open Matrix
Formal statement
namespace SelfDualLP.Output

/-- Proof of Theorem 8 (p. 63), orthogonality identity, under (7). Let `z = (y, x, τ, θ, s, κ) ∈ 𝓕_h`
and let `w = (y*, x*, τ*, θ*, s*, κ*)` be a self-complementary (optimal) solution of (HLP) with
`κ* = 0`. Then `(x − x*)ᵀ(s − s*) + (τ − τ*)(κ − κ*) = 0`, which can be rewritten as
`xᵀs* + sᵀx* + κτ* = (n + 1)μ = (n + 1)θ`, where `μ = (xᵀs + τκ)/(n + 1)`. -/
theorem orthogonality_identity {m n : ℕ} (A : Matrix (Fin m) (Fin n) ℝ) (b : Fin m → ℝ)
    (c : Fin n → ℝ) (z w : HLPPoint m n) (hz : HLPFeasible7 A b c z)
    (hw : HLPOptimal7 A b c w) (hκ : w.κ = 0) :
    (z.x - w.x) ⬝ᵥ (z.s - w.s) + (z.τ - w.τ) * (z.κ - w.κ) = 0 ∧
    z.x ⬝ᵥ w.s + z.s ⬝ᵥ w.x + z.κ * w.τ = ((n : ℝ) + 1) * mu z ∧
    ((n : ℝ) + 1) * mu z = ((n : ℝ) + 1) * z.θ := by sorry
end SelfDualLP.Output
Source
Ye, Todd, Mizuno, An O(√nL)-Iteration Homogeneous and Self-Dual Linear Programming Algorithm, Math. Oper. Res. 19(1) (1994), p. 63, proof of Theorem 8; DOI 10.1287/moor.19.1.53
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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