Lemma 4 (corrected: C1 on [A I], independent demands) — ∇ₓv(x, Y) exists with probability one
ProvedRLPBidPrice.Unbiased.lemma4_ae_gradient_existsLet be a 0/1 incidence matrix with nonzero columns, , and satisfying Condition C1′. Let be a random demand vector on a probability space : measurable, almost surely, with mutually independent components satisfying Condition 2 ( is continuous for every ). Then, with probability one, the perfect-information LP value is differentiable at :
This is the almost-sure differentiability that feeds Lemma 1 in the proof of Theorem 1.
Formalization Note. Differentiability is Fréchet differentiability of at . C1′ forces every , so with the point is interior. Condition 2 is kept in the paper's marginal form (continuity of each distribution function); the alternative of a joint density is not assumed. The time index is dropped.
Correction. Two hypotheses are strengthened. (i) Condition 1 is applied to the columns of (C1′): with having rows , , and i.i.d. uniform on , the printed Condition 1 and Condition 2 hold, but on (probability ) the optimal dual prices form the segment , so is not differentiable there. (ii) The components of are assumed mutually independent: with one leg, , fares and , uniform on , Conditions C1′ and 2 hold but surely and has a kink at almost surely.
import Mathlib import Definitions.Def_RLPBidPrice_Unbiased_Model open MeasureTheory Matrix
namespace RLPBidPrice.Unbiased
theorem lemma4_ae_gradient_exists {m n : ℕ} (A : Matrix (Fin m) (Fin n) ℝ) (r : Fin n → ℝ)
(hA : IsIncidence A) (hr : 0 ≤ r) (x : Fin m → ℝ) (hx : 0 ≤ x)
(hC1 : ConditionOnePrime A x)
{Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(Y : Ω → Fin n → ℝ) (hYm : Measurable Y) (hY0 : ∀ᵐ ω ∂P, 0 ≤ Y ω)
(hC2 : ConditionTwo P Y) (hind : ProbabilityTheory.iIndepFun (fun j ω => Y ω j) P) :
∀ᵐ ω ∂P, DifferentiableAt ℝ (fun x' => piValue A r x' (Y ω)) x := by sorry
end RLPBidPrice.Unbiased
Read-back
What the Lean code literally says, in plain math · claude-opus-5-5
Let be a real matrix whose entries are all or and whose every column has a nonzero entry. Assume with and with . Assume satisfies Condition C1′:
for all finite and , if lies in the real linear span of , then that span is all of .
Let be a probability space and a measurable map. Assume:
- coordinatewise for -almost every ;
- Condition 2: for each , is continuous on . That is, each coordinate has no atoms;
- the coordinates are mutually independent under .
For and , write
This is a real supremum, when the set is empty (for instance when has a negative coordinate).
Conclusion: for -almost every , the function on all of is Fréchet differentiable at .
Degenerate cases.
- C1′ restricts . Take and . Then C1′ forces every , so every . For and the hypotheses fail and the statement is vacuous.
- If , the incidence condition forces and the conclusion is trivial on .
- If , Condition 2 and independence hold vacuously and near .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.