Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

(A.3) — one exact randomized BR step contracts ∥xk−x∗∥P2\|x_k - x^*\|_P^2∥xk​−x∗∥P2​ in conditional mean

Proved
LeiBR.Rand.eqA3

by mikedeng1 · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

conditional-expectationp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1randomized-algorithms

Let Assumptions 1 and 2 hold, μ>0\mu > 0μ>0, a=∥Γ∥a = \|\Gamma\|a=∥Γ∥, and let x∗x^*x∗ be a Nash equilibrium. Let xkx_kxk​ be an Fk\mathcal F_kFk​-measurable random profile with values in XXX. Let χi,k∈{0,1}\chi_{i,k} \in \{0,1\}χi,k​∈{0,1} be coins with P(χi,k=1)=pi∈(0,1]\mathbb P(\chi_{i,k} = 1) = p_i \in (0,1]P(χi,k​=1)=pi​∈(0,1], each independent of Fk\mathcal F_kFk​. Define the exact randomized step

ti,k=xi,k+χi,k(x^i(xk)−xi,k).t_{i,k} = x_{i,k} + \chi_{i,k}\big(\widehat x_i(x_k) - x_{i,k}\big).ti,k​=xi,k​+χi,k​(xi​(xk​)−xi,k​).

Then, almost surely,

E[∥tk−x∗∥P2 ∣ Fk]≤∥xk−x∗∥P2−(1−a2) ∥xk−x∗∥2,\mathbb E\big[\|t_k - x^*\|_P^2 \,\big|\, \mathcal F_k\big] \le \|x_k - x^*\|_P^2 - (1 - a^2)\,\|x_k - x^*\|^2,E[∥tk​−x∗∥P2​​Fk​]≤∥xk​−x∗∥P2​−(1−a2)∥xk​−x∗∥2,

where ∥⋅∥P\|\cdot\|_P∥⋅∥P​ is the weighted norm (A.2) and ∥xk−x∗∥2=∑i∥xi,k−xi∗∥2\|x_k - x^*\|^2 = \sum_i \|x_{i,k} - x_i^*\|^2∥xk​−x∗∥2=∑i​∥xi,k​−xi∗​∥2.

This is the one-step descent inequality of the randomized scheme for the exact best response, the starting point of both Lemma 4 and Lemma 5.

Formalization Note Only the final inequality of the paper's display (A.3) is stated; its intermediate equalities, which use (A.1), are not.

Preamble
import Mathlib
import Definitions.Def_LeiBR_Rand_Game
import Definitions.Def_LeiBR_Rand_ProxBR
import Definitions.Def_LeiBR_Rand_Assumption1
import Definitions.Def_LeiBR_Rand_SA
import Definitions.Def_LeiBR_Rand_Algorithm2
import Definitions.Def_LeiBR_Rand_Constants

open MeasureTheory ProbabilityTheory
Formal statement
namespace LeiBR.Rand

/-- (A.3), App. A, p. 31. Let `x_k` be an `F_k`-measurable `X`-valued random profile and let the
coins `χ_i ∈ {0,1}` satisfy `P(χ_i = 1) = p_i` and be independent of `F_k`. With
`t_{i,k} = x_{i,k} + χ_{i,k}(x̂_i(x_k) − x_{i,k})`,
`E[‖t_k − x*‖²_P | F_k] ≤ ‖x_k − x*‖²_P − (1 − a²)‖x_k − x*‖²` a.s. -/
theorem eqA3 {N : ℕ} {n : Fin N → ℕ} {d : ℕ} (G : Game N n)
    (μξ : Measure (EuclideanSpace ℝ (Fin d)))
    (ψ : Fin N → LeiBR.Sync.Profile n → EuclideanSpace ℝ (Fin d) → ℝ)
    (gψ : ∀ i, LeiBR.Sync.Profile n → EuclideanSpace ℝ (Fin d) → LeiBR.Sync.Strat n i) (M : Fin N → ℝ)
    (hA1 : Assumption1 G μξ ψ gψ M) (mu : ℝ) (hmu : 0 < mu)
    (xhat : LeiBR.Sync.Profile n → LeiBR.Sync.Profile n) (hxhat : IsProxBR G mu xhat)
    (hA2 : Assumption2 G mu) (xstar : LeiBR.Sync.Profile n) (hNE : G.IsNE xstar)
    {Ω : Type*} [mΩ : MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
    (F : Filtration ℕ mΩ) (k : ℕ) (p : Fin N → ℝ) (hp : ∀ i, 0 < p i) (hp1 : ∀ i, p i ≤ 1)
    (xk : Ω → LeiBR.Sync.Profile n) (hxk_meas : StronglyMeasurable[F k] xk)
    (hxk_feas : ∀ ω, G.Feasible (xk ω))
    (χ : Fin N → Ω → ℕ) (hχ1 : ∀ i ω, χ i ω ≤ 1) (hχ_meas : ∀ i, Measurable (χ i))
    (hχ_prob : ∀ i, P {ω | χ i ω = 1} = ENNReal.ofReal (p i))
    (hχ_indep : ∀ i, Indep (MeasurableSpace.comap (χ i) inferInstance) (F k) P) :
    P[fun ω => wnorm p ((fun i => xk ω i + (χ i ω : ℝ) • (xhat (xk ω) i - xk ω i)) - xstar) ^ 2 | F k]
      ≤ᵐ[P] fun ω => wnorm p (xk ω - xstar) ^ 2
        - (1 - contrFactor G mu ^ 2) * blockNorm (xk ω - xstar) ^ 2 := by sorry

end LeiBR.Rand
Source
Lei, Shanbhag, Pang & Sen, On Synchronous, Asynchronous, and Randomized Best-Response Schemes for Stochastic Nash Games, arXiv:1704.04578v2, p. 31, App. A, (A.1)–(A.3)
Human review
  • Endorsed by Shuze Chen · Oct 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 8, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me