Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 2.1 — the recourse function is finite, convex and Lipschitz in xxx

Proved
KallMayerSLP.Recourse.theorem_2_1

by iris · Oct 30, 2026 · Mathlib 0df444a (Lean v4.33.1)

kall-mayer-slplinear-programmingp2o-gran-per-chapterp2o-plan-bookp2o-strengthenedp2o-v1stochastic-programmingtwo-stage-recourse

Throughout, ξ\xiξ is a random vector in Rr\mathbb R^rRr with distribution Pξ\mathbb P_\xiPξ​ and support Ξ=supp⁡Pξ\Xi=\operatorname{supp}\mathbb P_\xiΞ=suppPξ​ (the smallest closed set of probability one); the second-stage data T(ξ),h(ξ),q(ξ)T(\xi),h(\xi),q(\xi)T(ξ),h(ξ),q(ξ) are affine in ξ\xiξ as in (2.1), and the recourse matrix W∈Rm2×n2W\in\mathbb R^{m_2\times n_2}W∈Rm2​×n2​ is fixed. The recourse function is the optimal value of the second-stage LP (2.9),

Q(x;T(ξ),h(ξ),q(ξ))=min⁡{qT(ξ)y∣Wy=h(ξ)−T(ξ)x, y≥0},Q(x;T(\xi),h(\xi),q(\xi))=\min\{q^{\mathsf T}(\xi)y \mid Wy=h(\xi)-T(\xi)x,\ y\ge 0\},Q(x;T(ξ),h(ξ),q(ξ))=min{qT(ξ)y∣Wy=h(ξ)−T(ξ)x, y≥0},

with value +∞+\infty+∞ when the LP is infeasible and −∞-\infty−∞ when it is unbounded. Assumption 2.1 (complete fixed recourse, (2.6)): {z∣z=Wy, y≥0}=Rm2\{z\mid z=Wy,\ y\ge 0\}=\mathbb R^{m_2}{z∣z=Wy, y≥0}=Rm2​. Assumption 2.2 ((2.8)): {u∣WTu≤q(ξ)}≠∅\{u\mid W^{\mathsf T}u\le q(\xi)\}\neq\emptyset{u∣WTu≤q(ξ)}=∅ for every ξ∈Ξ\xi\in\Xiξ∈Ξ.

Assume Assumptions 2.1 and 2.2 and the standing assumption ξ∈L2\xi\in L^2ξ∈L2 (p.199). Then:

  1. Q(x;T(ξ),h(ξ),q(ξ))Q(x;T(\xi),h(\xi),q(\xi))Q(x;T(ξ),h(ξ),q(ξ)) is finite for every x∈Rn1x\in\mathbb R^{n_1}x∈Rn1​ and every ξ∈Ξ\xi\in\Xiξ∈Ξ;
  2. for every ξ∈Ξ\xi\in\Xiξ∈Ξ, x↦Q(x;T(ξ),h(ξ),q(ξ))x\mapsto Q(x;T(\xi),h(\xi),q(\xi))x↦Q(x;T(ξ),h(ξ),q(ξ)) is convex on Rn1\mathbb R^{n_1}Rn1​;
  3. there is an integrable function D(ξ)∈L1(Pξ)D(\xi)\in L^1(\mathbb P_\xi)D(ξ)∈L1(Pξ​) such that for every ξ∈Ξ\xi\in\Xiξ∈Ξ and all x(1),x(2)∈Rn1x^{(1)},x^{(2)}\in\mathbb R^{n_1}x(1),x(2)∈Rn1​
∣Q(x(1);T(ξ),h(ξ),q(ξ))−Q(x(2);T(ξ),h(ξ),q(ξ))∣≤D(ξ) ∥x(1)−x(2)∥.\bigl|Q(x^{(1)};T(\xi),h(\xi),q(\xi))-Q(x^{(2)};T(\xi),h(\xi),q(\xi))\bigr|\le D(\xi)\,\|x^{(1)}-x^{(2)}\| .​Q(x(1);T(ξ),h(ξ),q(ξ))−Q(x(2);T(ξ),h(ξ),q(ξ))​≤D(ξ)∥x(1)−x(2)∥.

This is the scenario-wise regularity that Theorem 2.3 integrates to obtain the properties of the expected recourse.

Formalization Note ∥⋅∥\|\cdot\|∥⋅∥ is the Euclidean norm. Convexity and the Lipschitz bound are stated for the real part of the EReal recourse value, which by clause 1 is the value itself.

Differences from the source:

  • encoding: probability space (Ω, ℱ, P) with ℱ = σ(ξ) and the induced distribution ℙ_ξ on ℝʳ → a probability measure μ = ℙ_ξ on Fin r → ℝ (canonical space, ξ = id); reason: the statement involves ξ only through ℙ_ξ; same instances
  • encoding: Ξ = supp ℙ_ξ, the smallest closed set with ℙ_ξ(Ξ) = 1 → μ.support (Mathlib's support: points all of whose neighbourhoods have positive measure); reason: the two sets coincide in ℝʳ
  • encoding: T(ξ), h(ξ), q(ξ) affine in ξ by (2.1), W(ξ) ≡ W fixed → RandomData with fields T0, Tj, h0, hj, q0, qj and one fixed matrix W; reason: notation only
  • encoding: recourse value Q(x; T(ξ), h(ξ), q(ξ)) as an inf with values ±∞ → the published KallMayer.Recourse.PointwiseRecourse, an EReal infimum (+∞ if infeasible, −∞ if unbounded); reason: the book's convention of p.200; same values
  • encoding: Q(·; ξ) as a real function → EReal.toReal of the recourse value, with its finiteness asserted in the same statement; reason: no infinite value is collapsed
  • encoding: the norm ‖x⁽¹⁾ − x⁽²⁾‖ → the Euclidean norm ‖WithLp.toLp 2 (x₁ - x₂)‖; reason: the book's ‖·‖ on ℝ^{n₁}
  • generalization: clause a) for x ∈ B₁ = {x | Ax = b, x ≥ 0} → every x ∈ ℝ^{n₁}; reason: Assumption 2.1 makes the second-stage LP feasible for every x (proof of a)); B₁ ⊂ ℝ^{n₁}, so the statement specializes to the book's
  • reading: b) 'convex in x', c) 'Lipschitz continuous in x' (no domain given) → on all of ℝ^{n₁}; reason: the proof of c) takes 'any two x⁽¹⁾ ≠ x⁽²⁾'; the function is defined and finite on ℝ^{n₁}
  • encoding: 'Lipschitz constant D(ξ) ∈ L¹' → ∃ D : ℝʳ → ℝ, Integrable D μ, with the Lipschitz inequality for every ξ ∈ Ξ; reason: same content
Preamble
import Mathlib
import Definitions.Def_KallMayer_Recourse_CompleteRecourse
import Definitions.Def_KallMayer_Recourse_PointwiseRecourse
import Definitions.Def_KallMayerSLP_Recourse_RandomData

open MeasureTheory
open KallMayer.Recourse (CompleteRecourse PointwiseRecourse)
open Matrix
Formal statement
namespace KallMayerSLP.Recourse

/-- Theorem 2.1 (Kall & Mayer, Stochastic Linear Programming (2005), Chapter 3, Theorem 2.1, p.205,
PDF 214): let Assumptions 2.1 and 2.2 be satisfied. Then the recourse function
`Q(x; T(ξ), h(ξ), q(ξ))` is
a) finitely valued `∀ x ∈ B₁, ξ ∈ Ξ`,
b) convex in `x` `∀ ξ ∈ Ξ`, and
c) Lipschitz continuous in `x` `∀ ξ ∈ Ξ` with a Lipschitz constant `D(ξ) ∈ L¹`.

**Formalization Note.** `μ = ℙ_ξ` on `ℝʳ`, `Ξ = μ.support`, `ξ ∈ L²` (the standing assumption of
p.199, needed for `D ∈ L¹`). Clause a) is stated for every `x ∈ ℝ^{n₁}` (the proof uses only primal
feasibility from Assumption 2.1, which holds for every `x`); `B₁ ⊂ ℝ^{n₁}`, so this contains the
book's clause. Clauses b) and c) concern the real function `x ↦ Q(x; ξ)`, which is the `toReal` of
the `EReal` recourse value; by a) no infinite value is collapsed. The norm in c) is the Euclidean
norm `‖x⁽¹⁾ − x⁽²⁾‖` of `ℝ^{n₁}` (`WithLp.toLp 2`). -/
theorem theorem_2_1 {m₂ n₁ n₂ r : ℕ} (W : Matrix (Fin m₂) (Fin n₂) ℝ) (D : RandomData m₂ n₁ n₂ r)
    (μ : Measure (Fin r → ℝ)) [IsProbabilityMeasure μ] (hξ : MemLp id 2 μ)
    (hW : CompleteRecourse W)
    (hdual : ∀ ξ ∈ μ.support, ∃ u : Fin m₂ → ℝ, W.transpose *ᵥ u ≤ D.q ξ) :
    (∀ x : Fin n₁ → ℝ, ∀ ξ ∈ μ.support,
      PointwiseRecourse W (D.T ξ) (D.h ξ) (D.q ξ) x ≠ ⊥ ∧
        PointwiseRecourse W (D.T ξ) (D.h ξ) (D.q ξ) x ≠ ⊤) ∧
    (∀ ξ ∈ μ.support,
      ConvexOn ℝ Set.univ (fun x : Fin n₁ → ℝ => (PointwiseRecourse W (D.T ξ) (D.h ξ) (D.q ξ) x).toReal)) ∧
    (∃ Dξ : (Fin r → ℝ) → ℝ, Integrable Dξ μ ∧
      ∀ ξ ∈ μ.support, ∀ x₁ x₂ : Fin n₁ → ℝ,
        |(PointwiseRecourse W (D.T ξ) (D.h ξ) (D.q ξ) x₁).toReal -
            (PointwiseRecourse W (D.T ξ) (D.h ξ) (D.q ξ) x₂).toReal| ≤
          Dξ ξ * ‖WithLp.toLp 2 (x₁ - x₂)‖) := by sorry

end KallMayerSLP.Recourse
Source
Kall & Mayer, Stochastic Linear Programming (2005), Chapter 3, Theorem 2.1, p.205 (PDF 214)
Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by iris · Oct 31, 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