Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

§A.7(b) — if demand D(p, r) is convex in r, the value function V(r) is convex in r

Proved
ReferencePricing.PriceMonotone.value_convex_of_demand_convex

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

convexitydynamic-pricingdynamic-programmingp2o-batch-rm-cafterp2o-gran-per-chapterp2o-plan-paperp2o-v1

Convexity of the value function (Popescu and Wu 2007, Appendix A.7(b), p. 426). In the dynamic pricing problem with reference effects (price set P=[0,pˉ]P = [0,\bar p]P=[0,pˉ​] with pˉ>0\bar p > 0pˉ​>0, memory α∈[0,1)\alpha \in [0,1)α∈[0,1), discount β∈(0,1)\beta \in (0,1)β∈(0,1), demand D(p,r)=d(p)+R(r−p,r)D(p,r) = d(p) + R(r-p,r)D(p,r)=d(p)+R(r−p,r), short-term profit Π(p,r)=p D(p,r)\Pi(p,r) = p\,D(p,r)Π(p,r)=pD(p,r)), assume Assumptions 1–3 and that ddd is differentiable. Suppose that demand is convex in the reference price: for every p∈Pp \in Pp∈P, r↦D(p,r)r \mapsto D(p, r)r↦D(p,r) is convex on PPP. Then the value function

V(r0)=sup⁡pt∈P∑t≥0βt Π(pt,rt),rt+1=αrt+(1−α)pt,V(r_0) = \sup_{p_t \in P} \sum_{t \ge 0} \beta^t\, \Pi(p_t, r_t), \qquad r_{t+1} = \alpha r_t + (1-\alpha) p_t,V(r0​)=pt​∈Psup​t≥0∑​βtΠ(pt​,rt​),rt+1​=αrt​+(1−α)pt​,

is convex on PPP.

This is the step of the proof of Lemma 3(b) that makes the Bellman objective W(p,r)=Π(p,r)+βV(αr+(1−α)p)W(p, r) = \Pi(p,r) + \beta V(\alpha r + (1-\alpha)p)W(p,r)=Π(p,r)+βV(αr+(1−α)p) supermodular. The paper attributes it to Smith and McCardle (2002, Proposition 5).

Formalization Note. Convexity is required on the price set PPP, the domain of the reference price. VVV is the supremum over price sequences (see the definition value). Differentiable ℝ d is added, as in every item of the mission.

Preamble
import Mathlib
import Definitions.Def_RobustBooking_Shared_Assumptions
import Definitions.Def_RobustBooking_Shared_optPrice
open RobustBooking.Shared
Formal statement
namespace ReferencePricing.PriceMonotone

theorem value_convex_of_demand_convex
    (pbar α β : ℝ) (d : ℝ → ℝ) (R : ℝ → ℝ → ℝ)
    (hpbar : 0 < pbar) (hα₀ : 0 ≤ α) (hα₁ : α < 1) (hβ₀ : 0 < β) (hβ₁ : β < 1)
    (hA1 : Assumption1 pbar d R) (hA2 : Assumption2 R) (hA3 : Assumption3 pbar d R)
    (hd : Differentiable ℝ d)
    (hconv : ∀ p ∈ Set.Icc (0 : ℝ) pbar,
      ConvexOn ℝ (Set.Icc (0 : ℝ) pbar) (fun r => demand d R p r)) :
    ConvexOn ℝ (Set.Icc (0 : ℝ) pbar) (value (profit d R) α β pbar) := by sorry

end ReferencePricing.PriceMonotone
Source
Popescu and Wu, Dynamic Pricing Strategies with Reference Effects, Operations Research 55(3), 2007, p. 426, Appendix A.7(b), second sentence
Read-back

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

Setting. Let pˉ,α,β∈R\bar p,\alpha,\beta\in\mathbb Rpˉ​,α,β∈R, let d:R→Rd:\mathbb R\to\mathbb Rd:R→R, and let R:R×R→RR:\mathbb R\times\mathbb R\to\mathbb RR:R×R→R.

Hypotheses.

  • pˉ>0\bar p>0pˉ​>0.
  • 0≤α<10\le\alpha<10≤α<1.
  • 0<β<10<\beta<10<β<1.
  • The imported predicates Assumption1 (pˉ,d,R)(\bar p,d,R)(pˉ​,d,R), Assumption2 (R)(R)(R) and Assumption3 (pˉ,d,R)(\bar p,d,R)(pˉ​,d,R) hold. Their definitions are not shown.
  • ddd is differentiable everywhere.
  • For every p∈[0,pˉ]p\in[0,\bar p]p∈[0,pˉ​], the function r↦D(p,r)r\mapsto D(p,r)r↦D(p,r) is convex on [0,pˉ][0,\bar p][0,pˉ​], where D(p,r)D(p,r)D(p,r) is the imported demand d R p r (definition not shown).

Conclusion. Let Π\PiΠ be the imported profit d R, and let V:R→RV:\mathbb R\to\mathbb RV:R→R be the imported value (Π,α,β,pˉ)(\Pi,\alpha,\beta,\bar p)(Π,α,β,pˉ​). Neither definition is shown. The theorem asserts that VVV is convex on [0,pˉ][0,\bar p][0,pˉ​]: for all r,s∈[0,pˉ]r,s\in[0,\bar p]r,s∈[0,pˉ​] and all λ∈[0,1]\lambda\in[0,1]λ∈[0,1],

V(λr+(1−λ)s) ≤ λV(r)+(1−λ)V(s).V(\lambda r+(1-\lambda)s)\ \le\ \lambda V(r)+(1-\lambda)V(s).V(λr+(1−λ)s) ≤ λV(r)+(1−λ)V(s).

Nothing is claimed outside [0,pˉ][0,\bar p][0,pˉ​].

Degenerate cases.

  • The interval [0,pˉ][0,\bar p][0,pˉ​] is non-degenerate, and α=0\alpha=0α=0 is allowed.
  • The convexity hypothesis on DDD applies only for ppp and rrr in [0,pˉ][0,\bar p][0,pˉ​].
  • Any default value that value returns, for example a supremum of an unbounded or empty set, is fixed by the hidden definition and cannot be seen here.
  • If the hypotheses cannot all hold together, the theorem is vacuous.
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