Proof of Theorem 2, "Continuity of utility in all strategies", p. 12 — x_ij, s_ij and u_i are continuous in α
ProvedMultPacing.Existence.theorem_2_continuitycontinuityp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1pacingsmoothed-game
Fix a pacing game, and any , and consider the -smoothed pacing game. For every buyer and good , the allocation , the payment and the utility are continuous functions of the whole profile of pacing multipliers on the strategy space:
This is one of the hypotheses of the pure Nash equilibrium existence theorem (Debreu, Glicksberg, Fan) that proves Theorem 2.
Formalization Note Continuity is ContinuousOn on the cube Set.univ.pi (fun _ => Set.Icc 0 1) in with the product topology.
Differences from the source:
- generalization: the bullet is argued under Theorem 2's hypotheses (, and Definition 3's ) → every real ; reason: the two budget branches of the utility agree at for every , so continuity never uses a bound on , and the statement specializes to the source's case.
Preamble
import Mathlib import Definitions.Def_MultPacing_Existence_Smoothed
Formal statement
namespace MultPacing.Existence
theorem theorem_2_continuity {N G : Type*} [Fintype N] [DecidableEq N] [Fintype G]
(P : PacingGame N G) (ε H : ℝ) (hε : 0 < ε) :
(∀ i j, ContinuousOn (fun α => smAlloc P ε α i j) (Set.univ.pi fun _ => Set.Icc (0 : ℝ) 1)) ∧
(∀ i j, ContinuousOn (fun α => smPay P ε α i j) (Set.univ.pi fun _ => Set.Icc (0 : ℝ) 1)) ∧
(∀ i, ContinuousOn (fun α => smUtility P ε H α i) (Set.univ.pi fun _ => Set.Icc (0 : ℝ) 1)) := by sorry
end MultPacing.Existence
Source
Conitzer, Kroer, Sodomka & Stier-Moses, Multiplicative Pacing Equilibria in Auction Markets, arXiv:1706.07151v5, p. 12 (PDF 12), proof of Theorem 2, bullet "Continuity of utility in all strategies"
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.