Theorem 2.3: an efficient set maximizing (2.27) is optimal for the choice-based model, and the largest optimal set in the efficient order is increasing in the remaining capacity x and in time t
ProvedRevenueManagement.choice_optimal_policydiscrete-choiceefficient-setsnested-policiesoperations-researchrevenue-management
In the choice-based model with a choice model, arrival probabilities that are positive in every period , and nonnegative prices:
(a) in every period with units remaining some efficient set maximizes over all offer sets, and ;
(b) for fixed and , every efficient optimal set at is matched by an efficient optimal set at with at least as large a purchase probability;
(c) for fixed and , every efficient optimal set in period is matched by an efficient optimal set in period with at least as large a purchase probability.
Since the efficient sets are indexed in increasing order of , (b) and (c) say that the largest optimal index is nondecreasing in and in .
Preamble
import Definitions.Def_RevenueManagement_singleResource
Formal statement
namespace RevenueManagement
theorem choice_optimal_policy {n : ℕ} (lam : ℕ → ℝ) (P : Finset (Fin n) → Fin n → ℝ)
(p : Fin n → ℝ) (T : ℕ) (hP : IsChoiceModel P) (hlam : ∀ t, 0 ≤ lam t ∧ lam t ≤ 1)
(hpos : ∀ t, 1 ≤ t → t ≤ T → 0 < lam t) (hp : ∀ j, 0 ≤ p j) :
(∀ t x, 1 ≤ t → t ≤ T → 1 ≤ x → ∃ S, IsEfficient P p S ∧ IsChoiceOptimal lam P p T t x S ∧
choiceValue lam P p T t x = choiceObj lam P p T t x S + choiceValue lam P p T (t + 1) x) ∧
(∀ t x x', 1 ≤ t → t ≤ T → 1 ≤ x → x ≤ x' → ∀ S, IsEfficient P p S →
IsChoiceOptimal lam P p T t x S → ∃ S', IsEfficient P p S' ∧
IsChoiceOptimal lam P p T t x' S' ∧ purchaseProb P S ≤ purchaseProb P S') ∧
(∀ t t' x, 1 ≤ t → t ≤ t' → t' ≤ T → 1 ≤ x → ∀ S, IsEfficient P p S →
IsChoiceOptimal lam P p T t x S → ∃ S', IsEfficient P p S' ∧
IsChoiceOptimal lam P p T t' x S' ∧ purchaseProb P S ≤ purchaseProb P S') := by sorry
end RevenueManagement
Source
Kalyan T. Talluri and Garrett J. van Ryzin, The Theory and Practice of Revenue Management, Kluwer/Springer 2004, DOI 10.1007/b139000, p. 69, Theorem 2.3 ('An optimal policy for (2.24) is to select a set k* from among the m efficient, ordered sets {Sk : k = 1, ..., m} that maximizes (2.27). Moreover, for a fixed t, the largest optimal index k* is increasing in the remaining capacity x, and for any fixed x, k* is increasing in time t')
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.