Example 3.4, p. 8 — for V = ∏[l_i, u_i], Z*(V) = max_{S⊆A} f(S, (u₀, l₁, …, lₙ))
ProvedRobustMNL.SizeCon.rectangular_unconstrainedLet be a rectangular uncertainty set (, ) and real revenues. Then the optimal value of the (unconstrained) robust assortment problem equals the optimal revenue under the known parameter vector :
In words, the adversary facing a box picks the largest no-purchase weight and the smallest product weights, and the robust problem is an ordinary MNL assortment problem with these weights. It is the special case of Theorem 3.8, which the paper establishes independently in Example 3.4.
Formalization Note The paper states the chain ; the last expression is by the definition of as an optimal assortment under , so the Lean statement keeps the first equality, the mathematical content. The minimum is a real infimum over the box; and say the intervals are nonempty. Revenues are arbitrary reals. Products are 0-indexed (Fin n).
import Mathlib import Definitions.Def_ChoiceCDLP_MNL_mnlObjective import Definitions.Def_RobustMNL_SizeCon_Model
namespace RobustMNL.SizeCon
/-- Example 3.4, p. 8: over the rectangular set `∏_{i=0}^n [l_i, u_i] ⊆ ℝⁿ⁺¹₊₊`,
`Z*(V) = max_{S ⊆ 𝒜} min_{v ∈ V} f(S, v) = max_{S ⊆ 𝒜} f(S, (u₀, l₁, …, lₙ))`. -/
theorem rectangular_unconstrained {n : ℕ} (l₀ u₀ : ℝ) (l u : Fin n → ℝ) (hl₀ : 0 < l₀)
(hlu₀ : l₀ ≤ u₀) (hl : ∀ i, 0 < l i) (hlu : l ≤ u) (r : Fin n → ℝ) :
RobustMNL.Static.Zstar (box l₀ u₀ l u) r =
(Finset.univ : Finset (Finset (Fin n))).sup' Finset.univ_nonempty
(fun S => RobustMNL.Static.rev r S (u₀, l)) := by sorry
end RobustMNL.SizeCon
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.