Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proof of Proposition 2, p. 212 (corrected: the dual optimal set of (14)) — ∂v(x, Y) = {optimal dual prices}, uniformly bounded by r̄

Proved
RLPBidPrice.Directional.superdiff_eq_dualOpt

by mikedeng1 · Oct 3, 2026 · Mathlib 0df444a (Lean v4.33.1)

convex-analysislinear-programmingp2o-batch-rm-cafterp2o-gran-per-chapterp2o-plan-paperp2o-v1revenue-management

Let AAA be a 0/1 incidence matrix with nonzero columns, r≥0r \ge 0r≥0, xxx a capacity vector with every component xi>0x^i > 0xi>0, and Y≥0Y \ge 0Y≥0 a demand vector. Then the superdifferential of the concave function z↦v(z,Y)z \mapsto v(z, Y)z↦v(z,Y) at xxx,

∂v(x,Y)={μ∈Rm:v(z,Y)≤v(x,Y)+μ⊤(z−x) for all z≥0},\partial v(x, Y) = \{\mu \in \mathbb{R}^m : v(z, Y) \le v(x, Y) + \mu^\top (z - x) \text{ for all } z \ge 0\},∂v(x,Y)={μ∈Rm:v(z,Y)≤v(x,Y)+μ⊤(z−x) for all z≥0},

is exactly the set of optimal vectors of dual prices,

∂v(x,Y)={μ≥0:L(x,μ)=min⁡μ′≥0L(x,μ′)},L(x,μ)=∑j=1nYj(rj−μ⊤Aj)++μ⊤x,\partial v(x, Y) = \Big\{\mu \ge 0 : L(x, \mu) = \min_{\mu' \ge 0} L(x, \mu')\Big\}, \qquad L(x, \mu) = \sum_{j=1}^n Y^j (r^j - \mu^\top A^j)^+ + \mu^\top x,∂v(x,Y)={μ≥0:L(x,μ)=μ′≥0min​L(x,μ′)},L(x,μ)=j=1∑n​Yj(rj−μ⊤Aj)++μ⊤x,

and every μ∈∂v(x,Y)\mu \in \partial v(x, Y)μ∈∂v(x,Y) satisfies 0≤μi≤rˉ=max⁡jrj0 \le \mu^i \le \bar r = \max_j r^j0≤μi≤rˉ=maxj​rj for every leg iii. In particular ∂v(x,Y)\partial v(x, Y)∂v(x,Y) is bounded uniformly in YYY.

Formalization Note. The superdifferential is taken over the feasible capacities z≥0z \ge 0z≥0. rˉ\bar rrˉ is written ⨆ j, r j. x>0x > 0x>0 is the paper's implicit setting for differentiating at a capacity vector (at a zero capacity the dual prices of that leg are unbounded). The time index kkk is dropped.

Correction. The paper states ∂vk(x,Y)={μ:μ⊤x=vk(x,Y), μ⊤A≤r, μ≥0}\partial v_k(x, Y) = \{\mu : \mu^\top x = v_k(x, Y),\ \mu^\top A \le r,\ \mu \ge 0\}∂vk​(x,Y)={μ:μ⊤x=vk​(x,Y), μ⊤A≤r, μ≥0}, which omits the dual variables of the constraints y≤Yy \le Yy≤Y. With one leg, one itinerary, r=1r = 1r=1, x=1x = 1x=1 and Y=1/2Y = 1/2Y=1/2, the printed set is {1/2}\{1/2\}{1/2} while v(z,Y)=min⁡(z,1/2)v(z, Y) = \min(z, 1/2)v(z,Y)=min(z,1/2) has superdifferential {0}\{0\}{0} at x=1x = 1x=1.

Preamble
import Mathlib
import Definitions.Def_RLPBidPrice_Directional_Model

open MeasureTheory Matrix
Formal statement
namespace RLPBidPrice.Directional

/-- Proof of Proposition 2, p. 212 (corrected): for `x > 0` the superdifferential of
`z ↦ v(z, y₀)` at `x` is the set of optimal dual price vectors of (14), and it lies in the box
`0 ≤ μ ≤ r̄`, `r̄ = max_j r^j`, whatever `y₀` is. -/
theorem superdiff_eq_dualOpt {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 : ∀ i, 0 < x i)
    (y₀ : Fin n → ℝ) (hy₀ : 0 ≤ y₀) :
    superdiff A r x y₀ = {μ | IsDualOpt A r x y₀ μ} ∧
      ∀ μ ∈ superdiff A r x y₀, ∀ i, 0 ≤ μ i ∧ μ i ≤ ⨆ j, r j := by sorry

end RLPBidPrice.Directional
Source
Talluri, van Ryzin, A Randomized Linear Programming Method for Computing Network Bid Prices, Transp. Sci. 33(2), 1999, p. 212, proof of Proposition 2, first two sentences
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

Setting. Let m,n∈Nm, n \in \mathbb{N}m,n∈N, let AAA be a real m×nm \times nm×n matrix, and let r∈Rnr \in \mathbb{R}^nr∈Rn. All vector inequalities are componentwise.

Hypotheses.

  • AAA is an incidence matrix: every entry is 000 or 111, and every column has at least one nonzero entry.
  • r≥0r \ge 0r≥0.
  • x∈Rmx \in \mathbb{R}^mx∈Rm with xi>0x_i > 0xi​>0 for every iii.
  • y0∈Rny_0 \in \mathbb{R}^ny0​∈Rn with y0≥0y_0 \ge 0y0​≥0.

Definitions used.

  • v(z,y0)=sup⁡{ r⋅y:Ay≤z, 0≤y≤y0 }v(z,y_0) = \sup\{\, r\cdot y : Ay \le z,\ 0 \le y \le y_0 \,\}v(z,y0​)=sup{r⋅y:Ay≤z, 0≤y≤y0​}, which is 000 if the set is empty.
  • ∂v(x,y0)\partial v(x,y_0)∂v(x,y0​) is the set of all μ∈Rm\mu \in \mathbb{R}^mμ∈Rm (no sign restriction) with
v(z,y0)≤v(x,y0)+μ⋅(z−x)for all z≥0.v(z,y_0) \le v(x,y_0) + \mu\cdot(z-x) \quad \text{for all } z \ge 0 .v(z,y0​)≤v(x,y0​)+μ⋅(z−x)for all z≥0.
  • D(x,y0)D(x,y_0)D(x,y0​) is the set of all μ≥0\mu \ge 0μ≥0 that minimise
L(x,y0,μ)=∑j(y0)jmax⁡(0, rj−∑iμiAij)+μ⋅xL(x,y_0,\mu) = \sum_j (y_0)_j \max\Big(0,\ r_j - \sum_i \mu_i A_{ij}\Big) + \mu\cdot xL(x,y0​,μ)=j∑​(y0​)j​max(0, rj​−i∑​μi​Aij​)+μ⋅x

over all μ≥0\mu \ge 0μ≥0.

Conclusion. Both of the following hold:

  1. ∂v(x,y0)=D(x,y0)\partial v(x, y_0) = D(x, y_0)∂v(x,y0​)=D(x,y0​), as sets.
  2. Every μ∈∂v(x,y0)\mu \in \partial v(x,y_0)μ∈∂v(x,y0​) satisfies
0≤μi≤sup⁡jrjfor every i,0 \le \mu_i \le \sup_j r_j \quad \text{for every } i,0≤μi​≤jsup​rj​for every i,

where sup⁡jrj\sup_j r_jsupj​rj​ is the real supremum over j∈{1,…,n}j \in \{1,\dots,n\}j∈{1,…,n}. For n≥1n \ge 1n≥1 this is max⁡jrj\max_j r_jmaxj​rj​, and for n=0n = 0n=0 it is 000.

Neither part asserts that the set is nonempty.

Degenerate cases.

  • If m=0m = 0m=0: the incidence hypothesis forces n=0n = 0n=0. Both sets are then the one-point set R0\mathbb{R}^0R0, and the box condition is vacuous.
  • If n=0n = 0n=0 and m≥1m \ge 1m≥1: v≡0v \equiv 0v≡0 on z≥0z \ge 0z≥0 and L=μ⋅xL = \mu\cdot xL=μ⋅x, so both sets must equal {0}\{0\}{0}. The box becomes 0≤μi≤00 \le \mu_i \le 00≤μi​≤0.
  • If y0=0y_0 = 0y0​=0: v(⋅,0)≡0v(\cdot, 0) \equiv 0v(⋅,0)≡0 on z≥0z \ge 0z≥0, and the claim covers this case too.
Human review
  • Endorsed by Shuze Chen · Oct 21, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 21, 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