§7, p.51 — , maximized at with value
ProvedBilateralProfitLB.expected_profit_formulabilateral-tradelower-boundsonline-learningp2o-gran-per-chapterp2o-plan-paperp2o-v1posted-pricesregret-bounds
Let , , and let denote expectation when the valuation pair has law (uniform seller, buyer with survival function , independent). For all prices :
- ;
- the same holds for the expected-profit function of §2;
- ;
- : for every buyer price the best seller price is .
This reduces the choice of posted prices on these instances to the choice of the buyer price for .
Differences from the source:
- reading: the profit formula for " and the family " → ; reason: the sentence follows the introduction of both, and is used in Claim 3.4
- encoding: "the best seller price comes from setting , for an expected profit of " → items 3 and 4, with defined as ; reason: the definition of on p.15
Preamble
import Mathlib import Definitions.Def_BilateralProfitLB_Construction
Formal statement
namespace BilateralProfitLB
open MeasureTheory BilateralProfit
/-- **§7, p.51 (PDF 53), display before Claim 3.4**, Brüstle, di Gregorio, Fusco, Leonardi
(manuscript, sha256 e91e8127…): under `𝒟_j`, `E^j[Profit(p, q)] = (q − p) p S_j(q)`; "for any
buyer price `q`, the best seller price comes from setting `p = q/2`, for an expected profit of
`h^j(q) = q²/4 · S_j(q)`".
Stated for every `K ≥ 1`, `j ≤ K` (`j = 0`: the baseline `𝒟_0`) and prices `p, q ∈ [0, 1]`:
1. `E^j[Profit(p, q)] = (q − p) p S_j(q)`, the expectation of the realized profit;
2. the same for the expected-profit function `µ(p, q) = (q − p) P(p, q)` of §2;
3. `h^j(q) = q²/4 · S_j(q)`;
4. `E^j[Profit(p, q)] ≤ h^j(q)`.
**Formalization Note.** `h^j(q)` is defined as `µ(q/2, q)` under `𝒟_j` (§3.5, p.15), so item 3
is a statement, not a definition. -/
theorem expected_profit_formula (K : ℕ) (hK : 1 ≤ K) (j : ℕ) (hj : j ≤ K) (p q : ℝ)
(hp : p ∈ Set.Icc (0 : ℝ) 1) (hq : q ∈ Set.Icc (0 : ℝ) 1) :
(∫ v, profit v (p, q) ∂(law K j)) = (q - p) * p * S K j q ∧
muP (law K j) (p, q) = (q - p) * p * S K j q ∧
h K j q = q ^ 2 / 4 * S K j q ∧
(∫ v, profit v (p, q) ∂(law K j)) ≤ h K j q := by sorry
end BilateralProfitLB
Source
Brüstle, di Gregorio, Fusco, Leonardi, Regret Bounds for Profit Maximization in Bilateral Trade: Learning the Best Posted-Price Mechanism (manuscript, STOC 2027 submission, sha256 e91e81279505b765c7c96c219d674c0fbbbd6ca47a00da8fe228d498ca25ac87), §7, p.51 (PDF 53), display and sentence before Claim 3.4; h^j: §3.5, p.15 (PDF 17)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.