Lemma 2: reduction to a canonical instance
ProvedMinBuying.lemma2_canonical_reductionFor an independent triangular instance of products normalized to optimal uniform revenue one, and every , there exists a canonical instance with additional items plus one terminal item. Its designated cap-pricing revenue is at least the original optimal differentiated revenue minus , and its optimal uniform revenue is at most one.
import Definitions.Def_MinBuying_Triangular
namespace MinBuying
/-- Lemma 2, PDF p.20: normalized triangular instances admit canonical
cap-pricing approximations, with at most `n` nonterminal items. -/
theorem lemma2_canonical_reduction (n : ℕ) (hn : 0 < n)
(I : TriangularInstance n) (hNorm : UPRevenue (triangularJoint I) = 1)
(ε : ℝ) (hε : 0 < ε) :
∃ J : CanonicalInstance, J.count ≤ n ∧
DPRevenue (triangularJoint I) - ε ≤
Revenue (canonicalLaw J) (canonicalPrices J) ∧
UPRevenue (canonicalLaw J) ≤ 1 := by
sorry
end MinBuying
Read-back
What the Lean code literally says, in plain math · gpt-6-astra
For real , let be the image of Lebesgue measure restricted to under . For a measure on a finite real coordinate space and a price vector , let be the minimum of the numbers over those indices satisfying , with when there is no such index, and put . Write and , where the suprema are taken in and is the constant price vector. For every integer and positive real arrays , let and assume . Then, for every real , there exist a count with , a terminal cap , and arrays and for , strictly decreasing in their caps in the sense that whenever , such that, with and , one has and . The output count counts the nonterminal coordinates, so the output law has coordinates, and is permitted. There is no ordering hypothesis on the input caps. The output parameters may depend on , the approximation inequality is weak, and the output uniform supremum is required only to be at most .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.