Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Remark 3.6, (3.8b), (3.8d), (3.8f) — against φ2∗\varphi_2^*φ2∗​ the state stays in D2\mathcal D_2D2​ and player 2 acts on {M2V2=V2}\{\mathcal M_2V_2=V_2\}{M2​V2​=V2​} with impulse δ2\delta_2δ2​

Proved
ImpulseGames.Verification.remark3_6_player_two

by mikedeng1 · Oct 4, 2026 · Mathlib 0df444a (Lean v4.33.1)

impulse-controlp2o-batch-pfp1ap2o-gran-per-chapterp2o-plan-paperp2o-v1quasi-variational-inequalitiesstochastic-differential-games

Assume the hypotheses of Theorem 3.3 on V1,V2,δ1,δ2V_1,V_2,\delta_1,\delta_2V1​,V2​,δ1​,δ2​: (3.1) with δi\delta_iδi​ continuous on SSS, the QVI system (3.3), and the regularity conditions (ii)–(iii). Let x∈Sx\in Sx∈S, let φ1=(C1,ξ1)\varphi_1=(\mathcal C_1,\xi_1)φ1​=(C1​,ξ1​) be a strategy of player 1 such that (φ1,φ2∗)∈Φx(\varphi_1,\varphi_2^*)\in\Phi_x(φ1​,φ2∗​)∈Φx​, where φ2∗=(D2,δ2)\varphi_2^*=(\mathcal D_2,\delta_2)φ2∗​=(D2​,δ2​), and let X,τS,τ2,n,δ2,nX,\tau_S,\tau_{2,n},\delta_{2,n}X,τS​,τ2,n​,δ2,n​ be the corresponding controlled process, exit time and controls of player 2 on an admissible realization. Then, for every ω\omegaω:

  1. for every s∈ ]0,τS[s\in\,]0,\tau_S[s∈]0,τS​[ that is not an intervention time,
(M2V2−V2)(Xs)<0;(3.8b)(\mathcal M_2V_2-V_2)(X_s)<0 ;\tag{3.8b}(M2​V2​−V2​)(Xs​)<0;(3.8b)
  1. for every n≥1n\ge1n≥1 with τ2,n<τS\tau_{2,n}<\tau_Sτ2,n​<τS​,
δ2,n=δ2(X(τ2,n)−),(3.8d)\delta_{2,n}=\delta_2\big(X_{(\tau_{2,n})^-}\big),\tag{3.8d}δ2,n​=δ2​(X(τ2,n​)−​),(3.8d) (M2V2−V2)(X(τ2,n)−)=0.(3.8f)(\mathcal M_2V_2-V_2)\big(X_{(\tau_{2,n})^-}\big)=0.\tag{3.8f}(M2​V2​−V2​)(X(τ2,n​)−​)=0.(3.8f)

These identities are what Step 1 of the proof of Theorem 3.3 uses: player 1's payoff can be compared with V1V_1V1​ only because, against φ2∗\varphi_2^*φ2∗​, the state never leaves D2\mathcal D_2D2​ and player 2 intervenes only on {M2V2=V2}\{\mathcal M_2V_2=V_2\}{M2​V2​=V2​} with her optimal impulse.

Formalization Note. The page states (3.8b) for "every s≥0s\ge0s≥0" and (3.8d), (3.8f) for every τ2,k<∞\tau_{2,k}<\inftyτ2,k​<∞. M2V2\mathcal M_2V_2M2​V2​ is defined only on SSS and the controls are only meaningful before the end of the game, so the statement uses s∈ ]0,τS[s\in\,]0,\tau_S[s∈]0,τS​[ off the intervention times and τ2,n<τS\tau_{2,n}<\tau_Sτ2,n​<τS​; Lemma 2.3 explains why s=0s=0s=0 and intervention instants must be excluded. The symmetric statements (3.8a), (3.8c), (3.8e) for player 1 are not stated separately.

Preamble
import Mathlib
import Definitions.Def_ImpulseGames_Verification_Game
import Definitions.Def_ImpulseGames_Verification_QVI

open MeasureTheory ProbabilityTheory Filter Topology
open scoped ENNReal NNReal
Formal statement
namespace ImpulseGames.Verification

/-- Remark 3.6, (3.8b), (3.8d), (3.8f): under the hypotheses of Theorem 3.3, for every strategy
`φ_1 = (C1, ξ1)` with `(φ_1, φ_2^*) ∈ Φ_x` and every sample path of an admissible realization,
`(𝓜_2 V_2 - V_2)(X_s) < 0` at every `s ∈ ]0, τ_S[` that is not an intervention time, and at every
intervention of player 2 before `τ_S`, `δ_{2,n} = δ_2(X_{(τ_{2,n})^-})` and
`(𝓜_2 V_2 - V_2)(X_{(τ_{2,n})^-}) = 0`. -/
theorem remark3_6_player_two {d k : ℕ} {Ω : Type*} [m : MeasurableSpace Ω]
    (P : Measure Ω) (𝔽 : Filtration ℝ≥0 m) (W : ℝ≥0 → Ω → (Fin k → ℝ))
    (hW : IsBrownianBasis P 𝔽 W) (G : Game d k) (hG : G.Standing)
    (V : Player → State d → ℝ) (δ : (i : Player) → State d → G.Imp i)
    (hV : VerificationHypotheses G V δ) (x : State d) (hx : x ∈ G.S)
    (C1 : Set (State d)) (ξ1 : State d → G.Imp .one) (Y : ℕ → ℝ≥0 → Ω → State d)
    (hY : IsAdmissible P 𝔽 W G x ((starProfile G V δ).update .one C1 ξ1) Y) (ω : Ω) :
    (∀ s : ℝ≥0, 0 < s →
      (s : ℝ≥0∞) < exitTime G x ((starProfile G V δ).update .one C1 ξ1) (pathOf Y ω) →
      (s : ℝ≥0∞) ∉ interventionTimes G x ((starProfile G V δ).update .one C1 ξ1) (pathOf Y ω) →
      interventionOp G V δ .two
          (ctrl G x ((starProfile G V δ).update .one C1 ξ1) (pathOf Y ω) s)
        - V .two (ctrl G x ((starProfile G V δ).update .one C1 ξ1) (pathOf Y ω) s) < 0) ∧
    (∀ n : ℕ, 1 ≤ n →
      interTime G x ((starProfile G V δ).update .one C1 ξ1) (pathOf Y ω) .two n <
        exitTime G x ((starProfile G V δ).update .one C1 ξ1) (pathOf Y ω) →
      impulse G x ((starProfile G V δ).update .one C1 ξ1) (pathOf Y ω) .two n =
          δ .two (preState G x ((starProfile G V δ).update .one C1 ξ1) (pathOf Y ω) .two n) ∧
        interventionOp G V δ .two
            (preState G x ((starProfile G V δ).update .one C1 ξ1) (pathOf Y ω) .two n)
          - V .two (preState G x ((starProfile G V δ).update .one C1 ξ1) (pathOf Y ω) .two n)
          = 0) := by sorry

end ImpulseGames.Verification
Source
Aïd et al. (2020), Math. Oper. Res. 45(1), accepted manuscript, Remark 3.6, (3.8b), (3.8d), (3.8f) (p. 10)
Human review
  • Endorsed by Shuze Chen · Oct 4, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 4, 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