At a heuristic equilibrium the retailer over-estimates the profit of every narrower assortment
ProvedAssortSearch.HeurEq.overestimate_narrowerassortment-planningconsumer-searchmultinomial-logitp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Under the hypotheses of the mission (, , , monotone margins whenever , concave and increasing on with ), let be a heuristic equilibrium of iterative no-search assortment planning in the independent assortment model. Then for every assortment narrower than ,
where is the true expected profit and the profit the no-search model predicts with the estimates .
This combines Theorem 7 with the non-negativity of estimated variant profits, and is the comparison on which the proof of Theorem 8 rests.
Formalization Note is added as in the positivity milestone. The per-variant version is implied and not posed separately.
Preamble
import Mathlib import Definitions.Def_AssortSearch_HeurEq_Model
Formal statement
namespace AssortSearch.HeurEq
open RetailVariety.Structure
/-- Proof of Theorem 8, pp. 23–24: at a heuristic equilibrium `x*` the retailer over-estimates
the expected profit of every assortment narrower than `x*`:
`π(x) ≤ π(x | v̂(x*))` for `1 ≤ x < x*`. -/
theorem overestimate_narrower {n : ℕ} (lam : ℝ) (m : Fin n → ℝ) (c : ℝ → ℝ)
(v : Fin n → ℝ) (v0 : ℝ)
(hv : ∀ i, 0 < v i) (hanti : Antitone v) (hv0 : 0 < v0) (hlam : 0 < lam)
(hm : ∀ j k : Fin n, v k ≤ v j → m k ≤ m j)
(hcc : ConcaveOn ℝ (Set.Icc 0 1) c) (hcm : MonotoneOn c (Set.Icc 0 1))
(hc0 : 0 ≤ c 0)
(xstar : ℕ) (heq : IsHeuristicEquilibrium lam m c v v0 xstar)
(x : ℕ) (hx1 : 1 ≤ x) (hx : x < xstar) :
trueProfit lam m c v v0 x ≤
estProfit m c (estPref lam v v0 xstar) (estPrefNoPurchase lam v v0 xstar) x := by sorry
end AssortSearch.HeurEq
Source
Cachon, Terwiesch & Xu, Retail Assortment Planning in the Presence of Consumer Search, working paper (Dec. 20, 2002), pp. 23–24 (PDF 25–26), proof of Theorem 8, π(x) ≤ π(x|v̂(x*)) for 1 ≤ x < x*
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.