(17) — first-order condition: an optimal price is a weighted average of
ProvedProductFraming.Pricing.eq_17Fix a feasible framing on pages of capacity , a page-count law on , qualities and . Let be an optimal price vector for this framing, i.e. for every . Then for every displayed product , with page , that some consumer sees, i.e. ,
Thus each optimal price is the constant plus a weighted average of the revenues of the consumers who see the product. Since the weights of all products on one page agree up to a common factor, this gives the within-page part of Theorem 5.
Formalization Note. Prices are finite, so whenever , and the paper's "priced out" branch () does not arise; a priced-out product is the same as an undisplayed one. The hypothesis is added: it is exactly the condition that the denominator is positive (so Lean's convention never applies); a product that no consumer ever sees has an arbitrary price at an optimum, and the paper does not consider that case. The statement is about every maximizer; it does not assert that one exists.
import Mathlib import Definitions.Def_ProductFraming_Pricing_MNL import Definitions.Def_ProductFraming_Pricing_Model open Finset
namespace ProductFraming.Pricing
theorem eq_17 {n m p : ℕ} (a : Fin n → ℝ) (β : ℝ) (lam : ℕ → ℝ)
(hβ : 0 < β) (hm : 1 ≤ m) (hp : 1 ≤ p) (hlam : ProductFraming.Nest.IsPageLaw m lam)
(f : Fin n → ℕ) (hf : ProductFraming.Trunc.Feasible m p f) (r : Fin n → ℝ)
(hr : ∀ r' : Fin n → ℝ, totalRevenue m a β lam f r' ≤ totalRevenue m a β lam f r)
(i : Fin n) (hi : 1 ≤ f i) (hreach : 0 < ProductFraming.Nest.tail m lam (f i)) :
r i = (∑ l ∈ Icc (f i) m, lam l * mnl a β r (ProductFraming.Nest.consideration f l) i *
(1 / β + revenue a β r (ProductFraming.Nest.consideration f l))) /
(∑ l ∈ Icc (f i) m, lam l * mnl a β r (ProductFraming.Nest.consideration f l) i) := by sorry
end ProductFraming.Pricing
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.