Riemann-sum bound:
ProvedPersonalizedAssortment.IB.riemann_sum_boundp2o-batch-rm-cafterp2o-gran-per-chapterp2o-plan-paperp2o-v1penalty-functionreal-analysisriemann-sums
Let be a penalty function ( increasing, , ). Let be a positive integer and an integer with . Then
In the proof of Theorem 1, is the final inventory of a product with initial inventory . The bound turns the discrete dual value of that product into the continuous expression that appears in .
Preamble
import Mathlib import Definitions.Def_PersonalizedAssortment_IB_PenaltyFunction
Formal statement
namespace PersonalizedAssortment.IB
/-- §A.3, p. 1549: for a penalty function `Ψ`, a positive integer `c` and an integer `0 ≤ I < c`,
`(1/c) ∑_{k=I+1}^{c} Ψ(k/c) ≤ 1/c + ∫_{(I+1)/c}^1 Ψ(y) dy`. -/
theorem riemann_sum_bound (Ψ : ℝ → ℝ) (hΨ : IsPenaltyFunction Ψ) (c : ℕ) (hc : 0 < c)
(I : ℕ) (hI : I < c) :
1 / (c : ℝ) * ∑ k ∈ Finset.Ioc I c, Ψ ((k : ℝ) / (c : ℝ)) ≤
1 / (c : ℝ) + ∫ y in ((I : ℝ) + 1) / (c : ℝ)..1, Ψ y := by sorry
end PersonalizedAssortment.IB
Source
Golrezaei, Nazerzadeh, Rusmevichientong, Real-Time Optimization of Personalized Assortments, Management Sci. 60(6), 2014, https://doi.org/10.1287/mnsc.2014.1939, p. 1549, Section A.3, proof of Theorem 1, display after 'Because Psi(1) = 1 and Psi is increasing, we have'
Read-back
What the Lean code literally says, in plain math · claude-opus-5-5
Let satisfy:
- maps into ;
- is nondecreasing on ;
- and .
No concavity, continuity or differentiability is assumed. Let be a natural number with , and let be a natural number with . The statement asserts
where runs over the integers .
Degenerate cases.
- Since , the lower limit lies in , and every point in the sum lies in .
- If , the sum has the single term and the integral is over . The claim then reads .
- If , the hypotheses force , which is the same case.
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.