Proof of Theorem 3.8, p. 12 — max_{|S|≤K} f(S, (u₀, l)) = max{λ : max_{|S|≤K} (1/u₀) Σ l_i (r_i − λ) ≥ λ}
ProvedRobustMNL.SizeCon.known_weights_parametricassortment-optimizationfractional-programmingmultinomial-logitp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1
Let and , let be real revenues and . Write and for the MNL revenue of under the known weights . Then
This is the closing chain of the proof of Theorem 3.8: the parametric description of a size-constrained assortment problem with known weights. Combined with the parametric form of and the rectangular reduction, it identifies of a box with the known-weights optimum.
Formalization Note The "max" of the set of admissible is IsGreatest. Revenues are arbitrary reals. Products are 0-indexed (Fin n).
Preamble
import Mathlib import Definitions.Def_ChoiceCDLP_MNL_mnlObjective import Definitions.Def_RobustMNL_SizeCon_Model
Formal statement
namespace RobustMNL.SizeCon
/-- Proof of Theorem 3.8, p. 12, last chain: for positive weights `(u₀, l₁, …, lₙ)`,
`max_{|S| ≤ K} f(S, (u₀, l))` is the largest `λ` with
`λ ≤ max_{|S| ≤ K} (1/u₀) ∑_{i ∈ S} l_i (r_i − λ)`. -/
theorem known_weights_parametric {n : ℕ} (u₀ : ℝ) (l : Fin n → ℝ) (hu₀ : 0 < u₀)
(hl : ∀ i, 0 < l i) (r : Fin n → ℝ) (K : ℕ) :
IsGreatest
{lam : ℝ | lam ≤ (sizeFeasible n K).sup' (sizeFeasible_nonempty n K)
(fun S => (1 / u₀) * ∑ i ∈ S, l i * (r i - lam))}
(knownMax r K (u₀, l)) := by sorry
end RobustMNL.SizeCon
Source
Rusmevichientong, Topaloglu, Robust Assortment Optimization in Revenue Management Under the Multinomial Logit Choice Model, Operations Research (2012), doi:10.1287/opre.1120.1063, authors' manuscript of 20 Sep 2011, proof of Theorem 3.8, p. 12, last display chain
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.