Lemma 6, p. 21 — problem (8) has an optimal solution that is a nested-by-revenue assortment with at most one fractional product
ProvedNestedLogitVariants.Synergistic.lemma_6Fix a nest of an instance with fully-captured nests and a real number , and assume . The maximization problem
has an optimal solution of the form
for some . In particular the maximum in (8) is attained.
Up to one fractional component, a nested-by-revenue assortment solves the inner problem of (7); this is why the nested-by-revenue assortments are almost sufficient for problem (7).
Formalization Note The standing assumptions are those of the model (see the Model definition) and of §4: every nest is fully captured, , and for some nest (p. 16). Products are indexed from (Fin n), so is a Fin n. The value of the objective at is (real power and ). The condition is the range of the page.
import Mathlib import Definitions.Def_NestedLogitVariants_Synergistic_Relaxation
namespace NestedLogitVariants.Synergistic
/-- Lemma 6, p. 21: for every nest `i` and every `x`, problem (8) has an optimal solution `z*` of the
form `z*_1 = ⋯ = z*_{k−1} = 1`, `z*_k ∈ [0, 1]`, `z*_{k+1} = ⋯ = z*_n = 0` for some product `k`. -/
theorem lemma_6 {ι : Type*} [Fintype ι] [DecidableEq ι] {n : ℕ} (I : Instance ι n)
(hI : I.Standing) (hfc : ∀ i, I.vnp i = 0) (hsyn : ∃ i, 1 < I.γ i)
(hn1 : 0 < n) (i : ι) (x : ℝ) :
∃ z ∈ NestedLogitVariants.LP.box n, (∀ z' ∈ NestedLogitVariants.LP.box n, F8 I i z' x ≤ F8 I i z x) ∧ ∃ k : Fin n, IsFracPrefix z k := by sorry
end NestedLogitVariants.Synergistic
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.