Proposition 5.3: in the pricing-and-replenishment model, G_{t+1}(y, d) is jointly concave, Vₜ(x) is concave, and G_{t+1} is supermodular (its differences in d increase in y and in y increase in d)
ProvedRevenueManagement.replenishment_concave_supermodularIn the pricing-and-replenishment model (5.20) under its assumptions (affine random demand with bounded noise, revenue rate concave and continuous on , nonnegative ordering costs, convex nonnegative holding/backorder costs), for every period : (i) the continuation value is jointly concave in on ; (ii) the value function is concave on ; (iii)-(iv) has increasing differences, for and , , which is the supermodularity the book derives from its two partial-derivative statements (the derivatives need not exist for a concave function, so the statement is in difference form).
Formalization Note The book's is the continuation value of period ; the statement is indexed by the period whose continuation value is . The demand is affine in the rate; Assumption 5.1's convexity alone does not suffice (see the definition's note).
import Definitions.Def_RevenueManagement_dynamicPricing
namespace RevenueManagement
variable {Ω : Type*} [MeasurableSpace Ω]
theorem replenishment_concave_supermodular (M : ReplPricing Ω) (hM : M.IsModel) (t : ℕ)
(ht : 1 ≤ t) (htT : t ≤ M.T) :
ConcaveOn ℝ (Set.univ ×ˢ Set.Icc 0 M.dbar) (fun yd : ℝ × ℝ => M.contValue t yd.1 yd.2) ∧
ConcaveOn ℝ Set.univ (M.value t) ∧
(∀ y y' d d', y ≤ y' → d ∈ Set.Icc 0 M.dbar → d' ∈ Set.Icc 0 M.dbar → d ≤ d' →
M.contValue t y d' - M.contValue t y d ≤ M.contValue t y' d' - M.contValue t y' d) := by sorry
end RevenueManagement
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.