Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Kuhn–Tucker conditions (10) for μ*: expected leg demand above the bid prices ≤ capacity, with complementary slackness

Proved
BidPriceControl.Asymptotic.kuhn_tucker_conditions

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

optimizationp2o-batch-rm-cafterp2o-gran-per-chapterp2o-plan-paperp2o-v1revenue-management

Assume the standing assumptions of §1; in particular the marginals of RtjR_t^jRtj​ are continuous on (0,+∞)(0, +\infty)(0,+∞). Let μ∗\mu^*μ∗ be an optimal solution of (9). Then for every leg iii,

∑t=1k∑j=1nP(Rtj>μ∗Aj) aij≤xiandμ∗i(xi−∑t=1k∑j=1nP(Rtj>μ∗Aj) aij)=0.\sum_{t=1}^k\sum_{j=1}^n P(R_t^j > \mu^* A^j)\, a_{ij} \le x^i \qquad\text{and}\qquad \mu^{*i}\Big(x^i - \sum_{t=1}^k\sum_{j=1}^n P(R_t^j > \mu^* A^j)\, a_{ij}\Big) = 0.t=1∑k​j=1∑n​P(Rtj​>μ∗Aj)aij​≤xiandμ∗i(xi−t=1∑k​j=1∑n​P(Rtj​>μ∗Aj)aij​)=0.

In words, the expected number of seats of leg iii requested at revenues above the bid prices does not exceed its capacity, and equals it when the leg's bid price is positive.

Formalization Note. The paper prints (10) as ∑t∑jP(Rtj>μ∗Aj)Aj−λ=x\sum_t\sum_j P(R_t^j > \mu^*A^j)A^j - \lambda = x∑t​∑j​P(Rtj​>μ∗Aj)Aj−λ=x, λμ∗=0\lambda\mu^* = 0λμ∗=0, λ≥0\lambda \ge 0λ≥0. With that sign and λ≥0\lambda \ge 0λ≥0, expected demand would exceed capacity, which contradicts the paper's own reading of (10) on p. 1584 and its use on p. 1585. The statement uses the corrected sign, slack λ=x−∑t∑jP(⋅)Aj≥0\lambda = x - \sum_t\sum_j P(\cdot)A^j \ge 0λ=x−∑t​∑j​P(⋅)Aj≥0, with λ\lambdaλ eliminated. Only the necessity direction is stated.

Preamble
import Mathlib
import Definitions.Def_BidPriceControl_Asymptotic_Model
Formal statement
namespace BidPriceControl.Asymptotic

open MeasureTheory ProbabilityTheory

/-- The Kuhn–Tucker conditions (10) (pp. 1583–1584), with the sign of the slack corrected: if the
marginals of the `R_t^j` are continuous on `(0, +∞)` (standing assumptions of §1), an optimal
solution `μ*` of (9) satisfies, for every leg `i`,
`∑_{t=1}^k ∑_j P(R_t^j > μ*A^j) a_ij ≤ x^i` and `μ*^i (x^i − ∑_{t=1}^k ∑_j P(R_t^j > μ*A^j) a_ij) = 0`
(the slack `λ^i = x^i − ∑_t ∑_j P(R_t^j > μ*A^j) a_ij ≥ 0` and `λ^i μ*^i = 0`). -/
theorem kuhn_tucker_conditions {m n : ℕ} (A : Fin m → Fin n → ℕ) (F : ℕ → Measure (Fin n → ℝ))
    (k : ℕ) (x : Fin m → ℕ) (hF : ∀ t ∈ Finset.Icc 1 k, IsRequestLaw (F t))
    (μstar : Fin m → ℝ) (hmin : IsRelaxedMinimizer A F k x μstar) (i : Fin m) :
    ∑ t ∈ Finset.Icc 1 k, ∑ j, exceedProb A F μstar t j * (A i j : ℝ) ≤ (x i : ℝ) ∧
      μstar i * ((x i : ℝ) - ∑ t ∈ Finset.Icc 1 k, ∑ j, exceedProb A F μstar t j * (A i j : ℝ))
        = 0 := by sorry

end BidPriceControl.Asymptotic
Source
Talluri, van Ryzin, An Analysis of Bid-Price Controls for Network Revenue Management, Management Sci. 44(11), 1998, pp. 1583–1584, eq. (10) (sign of λ corrected)
Read-back

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

Data. The theorem takes A=(aij)∈Nm×nA=(a_{ij})\in\mathbb N^{m\times n}A=(aij​)∈Nm×n, measures (Ft)t(F_t)_t(Ft​)t​ on Rn\mathbb R^nRn, k∈Nk\in\mathbb Nk∈N, x∈Nmx\in\mathbb N^mx∈Nm, a vector μ∗∈Rm\mu^*\in\mathbb R^mμ∗∈Rm and a leg iii.

Hypotheses.

  • For each t∈{1,…,k}t\in\{1,\dots,k\}t∈{1,…,k}, FtF_tFt​ is a request law: a probability measure, a.s. nonnegative with at most one positive coordinate, no atom of any rjr_jrj​ at a positive value, and integrable coordinates.
  • μ∗\mu^*μ∗ minimizes the relaxed value: μ∗≥0\mu^*\ge0μ∗≥0, and Jˉk(x,μ∗)≤Jˉk(x,μ)\bar J_k(x,\mu^*)\le\bar J_k(x,\mu)Jˉk​(x,μ∗)≤Jˉk​(x,μ) for all μ≥0\mu\ge0μ≥0, where
Jˉk(x,μ)=∑t=1k∑j∫(rj−μAj)+dFt+∑iμixi,μAj=∑i′μi′ai′j.\bar J_k(x,\mu)=\sum_{t=1}^k\sum_j\int(r_j-\mu A^j)^+dF_t+\sum_i\mu_ix_i,\qquad \mu A^j=\sum_{i'}\mu_{i'}a_{i'j}.Jˉk​(x,μ)=t=1∑k​j∑​∫(rj​−μAj)+dFt​+i∑​μi​xi​,μAj=i′∑​μi′​ai′j​.

Conclusion. Write ptj=Ft({r:rj>μ∗Aj})p_{tj}=F_t(\{r:r_j>\mu^*A^j\})ptj​=Ft​({r:rj​>μ∗Aj}), with strict inequality. Then both

∑t=1k∑jptj aij≤xiandμi∗(xi−∑t=1k∑jptj aij)=0\sum_{t=1}^k\sum_j p_{tj}\,a_{ij}\le x_i\qquad\text{and}\qquad \mu^*_i\Big(x_i-\sum_{t=1}^k\sum_j p_{tj}\,a_{ij}\Big)=0t=1∑k​j∑​ptj​aij​≤xi​andμi∗​(xi​−t=1∑k​j∑​ptj​aij​)=0

hold.

Degenerate cases.

  • No minimizer: if no minimizer exists for the given data, the statement asserts nothing.
  • k=0k=0k=0: it asserts 0≤xi0\le x_i0≤xi​ and μi∗xi=0\mu^*_ix_i=0μi∗​xi​=0.
  • n=0n=0n=0: the sums are 000, with the same conclusions as k=0k=0k=0.
  • Zero capacity: if xi=0x_i=0xi​=0, it asserts ptj=0p_{tj}=0ptj​=0 for every t≤kt\le kt≤k and every jjj with aij>0a_{ij}>0aij​>0.
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