Eq. (7):
ProvedPricingInvCensored.eq_7operations-researchp2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1probability
Let be a probability measure on with finite mean, let , and let . Then
With and for prices , the left side is the unobservable part of (Eq. (6)), and the right side splits it into Part A, known up to , and Part B, estimable from censored leftovers.
Differences from the source:
- generalization: the noise law of Assumption 1 and for → any probability measure on with finite mean and any ; reason: the identity holds for every such law and pair, and the source's case is with , .
Preamble
import Mathlib open MeasureTheory Set
Formal statement
namespace PricingInvCensored
/-- **Eq. (7)**, Chen, Wang & Zhou, SSRN 3750413 (2022-10-19), p.15. For a noise law `µ` with
finite mean and `z ≤ z′`,
`E[(ε − z)⁺] − E[(ε − z′)⁺] = E[(ε − z)⁺ − (ε − z′)⁺]
= (z′ − z) Pr[ε ≥ z] − E[(z′ − ε) 1{z ≤ ε ≤ z′}]`.
**Formalization Note.** Stated for every probability measure on `ℝ` with integrable identity and
every `z ≤ z′`; the paper's case is `µ` of Assumption 1 and `z = z_p < z′ = z_{p′}`. -/
theorem eq_7 (μ : Measure ℝ) [IsProbabilityMeasure μ] (hμ : Integrable (fun x : ℝ => x) μ)
(z z' : ℝ) (hzz : z ≤ z') :
(∫ ε, max (ε - z) 0 ∂μ) - (∫ ε, max (ε - z') 0 ∂μ) =
∫ ε, (max (ε - z) 0 - max (ε - z') 0) ∂μ ∧
∫ ε, (max (ε - z) 0 - max (ε - z') 0) ∂μ =
(z' - z) * (μ (Ici z)).toReal - ∫ ε in Icc z z', (z' - ε) ∂μ := by sorry
end PricingInvCensored
Source
Chen, Wang & Zhou, Optimal Policies for Dynamic Pricing and Inventory Control with Nonparametric Censored Demands, SSRN 3750413 (2022-10-19), Eq. (7), p.15
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.