Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 2.4(1) / Proposition 2.9: ξ∗=Y/(2η)\xi^*=Y/(2\eta)ξ∗=Y/(2η) is optimal given any version μ∗\mu^*μ∗ of E[ξt∗∣Ft0]\mathbb E[\xi^*_t\mid\mathcal F^0_t]E[ξt∗​∣Ft0​]

Open
MFGLiquidation.Nash.proposition_2_9

by tianyipeng · Nov 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

fbsdemean-field-gameoptimal-control

Let (X,Y,Z)(X,Y,Z)(X,Y,Z) be a solution of the conditional mean-field FBSDE (2.3) for the representative player's data (T,W~,X,κ,λ,η)(T,\widetilde W,\mathcal X,\kappa,\lambda,\eta)(T,W,X,κ,λ,η), in the class of Theorem 2.4:

(X,Y,Z)∈Hα×LF2([0,T]×Ω;R)×LF2([0,T−]×Ω;Rm),(X,Y,Z)\in\mathcal H_\alpha\times L^2_{\mathbb F}([0,T]\times\Omega;\mathbb R)\times L^2_{\mathbb F}([0,T-]\times\Omega;\mathbb R^m),(X,Y,Z)∈Hα​×LF2​([0,T]×Ω;R)×LF2​([0,T−]×Ω;Rm),

and assume the standing Assumption 2.3. Put ξt∗=Yt/(2ηt)\xi^*_t=Y_t/(2\eta_t)ξt∗​=Yt​/(2ηt​) (the strategy (2.2)) and let μ∗\mu^*μ∗ be any F0\mathbb F^0F0-progressive version of

μt∗=E[Yt2ηt ∣ Ft0].\mu^*_t=\mathbb E\Big[\frac{Y_t}{2\eta_t}\,\Big|\,\mathcal F^0_t\Big].μt∗​=E[2ηt​Yt​​​Ft0​].

Then ξ∗\xi^*ξ∗ is an optimal control for the representative player given μ∗\mu^*μ∗: ξ∗∈AF(X)\xi^*\in\mathcal A_{\mathbb F}(\mathcal X)ξ∗∈AF​(X) and, for every admissible ξ∈AF(X)\xi\in\mathcal A_{\mathbb F}(\mathcal X)ξ∈AF​(X),

J(X,ξ∗;μ∗)≤J(X,ξ;μ∗)a.s.,J(\mathcal X,\xi^*;\mu^*)\le J(\mathcal X,\xi;\mu^*)\qquad\text{a.s.},J(X,ξ∗;μ∗)≤J(X,ξ;μ∗)a.s.,

where J(X,ξ;μ)=E[∫0T(κsXsξμs+ηsξs2+λs(Xsξ)2) ds ∣ X]J(\mathcal X,\xi;\mu)=\mathbb E\big[\int_0^T(\kappa_sX^\xi_s\mu_s+\eta_s\xi_s^2+\lambda_s(X^\xi_s)^2)\,ds\,\big|\,\mathcal X\big]J(X,ξ;μ)=E[∫0T​(κs​Xsξ​μs​+ηs​ξs2​+λs​(Xsξ​)2)ds​X] is the conditional cost of the representative player.

This is item 1 of Theorem 2.4 (proved in Proposition 2.9), restated over the definitions of the Nash-equilibrium chapter (Section 3) so that it applies to each player iii of the NNN-player game, whose data (T,(W0,Wi),Xi,κi,λi,ηi)(T,(W^0,W^i),\mathcal X^i,\kappa^i,\lambda^i,\eta^i)(T,(W0,Wi),Xi,κi,λi,ηi) is a representative-player data set. It is the inequality "J1(ξ;μ∗)≥J1(ξ∗,1;μ∗)J^1(\xi;\mu^*)\ge J^1(\xi^{*,1};\mu^*)J1(ξ;μ∗)≥J1(ξ∗,1;μ∗) by Proposition 2.9" used in the proof of Theorem 3.3 (p. 23).

Formalization Note. The solution is any solution in the class of Theorem 2.4 (unique by that theorem), so the statement is Theorem 2.4(1) for the solution at hand. It coincides with the optimality conjunct of MFGLiquidation.Equilibrium.theorem_2_4, over the chapter-3 copies of the definitions (MFGLiquidation.Nash.*).

Preamble
import Mathlib
import Definitions.Def_MFGLiquidation_Nash_Setting
import Definitions.Def_MFGLiquidation_Nash_Game

open MeasureTheory ProbabilityTheory Filter Topology Peng1990.SMP
open scoped ENNReal NNReal
Formal statement
namespace MFGLiquidation.Nash

theorem proposition_2_9
    {Ω : Type*} [MeasurableSpace Ω] {k : ℕ} (P : Measure Ω) [IsProbabilityMeasure P]
    (D : Data Ω k) (hD : D.Standing P) (hA : D.Assumption23 P hD)
    (X Y : ℝ≥0 → Ω → ℝ) (Z : Fin (k + 1) → ℝ≥0 → Ω → ℝ)
    (hsol : SolvesFBSDE23InClass hD X Y Z)
    (μ : ℝ≥0 → Ω → ℝ)
    (hμ : IsCondExpVersion (filtF0 hD) P D.T (fun t ω => Y t ω / (2 * D.η t ω)) μ) :
    IsOptimal hD μ (fun t ω => Y t ω / (2 * D.η t ω)) := by sorry

end MFGLiquidation.Nash
Source
Fu, Graewe, Horst, Popier, A Mean Field Game of Optimal Portfolio Liquidation, arXiv:1804.04911v3, p. 9, Theorem 2.4 (1); p. 15, Proposition 2.9; used on p. 23 in the proof of Theorem 3.3

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