Lemma 7, p. 57 — for w_i(τ) = mτ + a with m, a > 0, r(u, i, w, σ) is strictly quasi-convex in u whenever ∆_{ji} ≤ 0
ProvedDriverSurge.IC.lemma_7affine-pricingp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1quasi-convexitysurge-pricing
Let the payment in state be positive affine, with and . For every policy (with and -integrable on ) such that
the function is strictly quasi-convex on : there is such that its derivative is strictly negative on and strictly positive on .
With Lemma 5 this gives the surge-state policy form under positive affine pricing.
Formalization Note Lean state 0 is the paper's non-surge state 1 and Lean state 1 the surge state 2; other i is the state .
Differences from the source:
- reading: "strictly quasi-convex in " → the definition of p. 63: the derivative in is strictly negative on and strictly positive on for some ; reason: the paper's own definition (proof of Lemma 7, p. 63), with the turning point allowed at or
- encoding: the formulas use and without saying they are finite → and are assumed -integrable on ; reason: the quantities , , are undefined otherwise (Lean's integral would be the junk value 0); only undefined cases are excluded
- generalization: D.3.1's standing assumptions (p. 56): has a bounded density, and state 2 dominates state 1 ( with for all ) → not assumed; reason: the claim is a property of as a function of that holds for all trip-length laws and payments, so it specializes to the source
Preamble
import Mathlib import Definitions.Def_DriverSurge_IC_Pricing
Formal statement
namespace DriverSurge.IC
open MeasureTheory Set
/-- Lemma 7 (p. 57). If `w_i(τ) = mτ + a` with `m, a > 0`, then for every policy `σ` (with `w_j` and `τ`
integrable on `σ_j`, `σ_i`) where `Δ_{ji} ≤ 0`, `r(u, i, w, σ)` is strictly quasi-convex in `u > 0` (its
derivative is strictly negative up to a point and strictly positive above it). -/
theorem lemma_7 (P : Primitives) (w : Fin 2 → ℝ → ℝ) (i : Fin 2) (m a : ℝ) (hm : 0 < m)
(ha : 0 < a) (hwi : w i = fun τ => m * τ + a) (σ : Fin 2 → Set ℝ) (hσ : IsPolicy σ)
(hw : ∀ k, IntegrableOn (w k) (σ k) (P.F k))
(hτ : ∀ k, IntegrableOn (fun τ : ℝ => τ) (σ k) (P.F k))
(hΔ : stateRate P (other i) (w (other i)) (σ (other i)) - stateRate P i (w i) (σ i) ≤ 0) :
StrictQuasiconvexIoi (rFun P w σ i) := by sorry
end DriverSurge.IC
Source
Garg, Nazerzadeh, Driver Surge Pricing, arXiv:1905.07544v4, p. 57, Lemma 7 (proof p. 63)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.