Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Appendix, (II), p. 15 — conic quadratic duality: strictly feasible and bounded below ⇒ dual solvable, equal values

Proved
RobustUncLP.Ellipsoidal.cqp_duality

by mikedeng1 · Oct 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

conic-quadraticdualityp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1strong-duality

Consider the conic quadratic problem (CQPp_pp​) and its dual (CQPd_dd​). Suppose that:

  1. (CQPp_pp​) is strictly feasible: some z^\hat zz^ satisfies Rz^=rR\hat z = rRz^=r and ∥Aℓz^−bℓ∥<cℓTz^−dℓ\|A_\ell\hat z - b_\ell\| < c_\ell^T\hat z - d_\ell∥Aℓ​z^−bℓ​∥<cℓT​z^−dℓ​ for every ℓ=0,…,k\ell = 0, \dots, kℓ=0,…,k;
  2. the objective eTz+φe^Tz + \varphieTz+φ is bounded below on the feasible set of (CQPp_pp​).

Then there is a real number vvv which is the infimum of the primal objective over the primal feasible set, and which is attained as the maximum of the dual objective over the dual feasible set:

v=inf⁡z feasible(eTz+φ)=max⁡(λ,μ,ν) feasible(rTλ+∑ℓ=0k[dℓνℓ+bℓTμℓ]+φ).v = \inf_{z \text{ feasible}} (e^Tz + \varphi) = \max_{(\lambda,\mu,\nu) \text{ feasible}} \Big( r^T\lambda + \sum_{\ell=0}^k [d_\ell\nu_\ell + b_\ell^T\mu_\ell] + \varphi \Big).v=z feasibleinf​(eTz+φ)=(λ,μ,ν) feasiblemax​(rTλ+ℓ=0∑k​[dℓ​νℓ​+bℓT​μℓ​]+φ).

The paper quotes this from Nesterov and Nemirovski (1994), Theorem 4.2.1, and applies it to each problem (Pi[x])(P_i[x])(Pi​[x]).

Formalization Note "The dual is solvable and the optimal values are equal" is encoded as IsGLB of the primal objective values together with IsGreatest of the dual objective values at the same vvv. The dual uses the sum over ℓ=0,…,k\ell = 0, \dots, kℓ=0,…,k in its equality constraint (see the CQP definition). The rows of RRR need not be linearly independent.

Preamble
import Mathlib
import Definitions.Def_RobustUncLP_Ellipsoidal_Setting
import Definitions.Def_RobustUncLP_Ellipsoidal_CQP
Formal statement
namespace RobustUncLP.Ellipsoidal

open Matrix CQPData

/-- Appendix, claim (II), p. 15 (conic quadratic duality, cited from Nesterov–Nemirovski 1994,
Theorem 4.2.1): if `(CQP_p)` is strictly feasible and its objective is bounded below on its
feasible set, then `(CQP_d)` is solvable and the two optimal values coincide. -/
theorem cqp_duality {N p k : ℕ} (P : CQPData N p k)
    (hstrict : ∃ z, P.R *ᵥ z = P.r ∧ ∀ ℓ, euclidNorm (P.A ℓ *ᵥ z - P.b ℓ) < P.c ℓ ⬝ᵥ z - P.d ℓ)
    (hbdd : BddBelow (P.primalObj '' P.primalFeas)) :
    ∃ v : ℝ, IsGLB (P.primalObj '' P.primalFeas) v ∧ IsGreatest (P.dualObj '' P.dualFeas) v := by sorry

end RobustUncLP.Ellipsoidal
Source
Ben-Tal & Nemirovski, Robust solutions of uncertain linear programs, Oper. Res. Lett. 25 (1999); authors' manuscript, Appendix, p. 15, claim (II) (citing Nesterov–Nemirovski 1994, Thm 4.2.1)
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