Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 4.4 (ii) — (∑iDi)†Dj=(∑iDi∗Di)†Dj∗Dj(\sum_i D_i)^\dagger D_j = (\sum_i D_i^*D_i)^\dagger D_j^*D_j(∑i​Di​)†Dj​=(∑i​Di∗​Di​)†Dj∗​Dj​

Proved
HarmonicGames.Decomposition.lemma_4_4_ii

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

game-theoryharmonic-gameslinear-algebrap2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1potential-games

For a finite game with players M\mathcal MM and nonempty finite strategy sets EmE^mEm (hm=∣Em∣h_m = |E^m|hm​=∣Em∣), let Dm:C0→C1D_m : C_0 \to C_1Dm​:C0​→C1​ be the gradient along mmm-comparable profiles, D:C0M→C1D : C_0^{\mathcal M}\to C_1D:C0M​→C1​, Du=∑mDmumDu = \sum_m D_m u^mDu=∑m​Dm​um, δ0\delta_0δ0​ the gradient of the game graph, Π=diag⁡(Π1,…,ΠM)\Pi = \operatorname{diag}(\Pi_1,\dots,\Pi_M)Π=diag(Π1​,…,ΠM​) with Πm=Dm†Dm\Pi_m = D_m^\dagger D_mΠm​=Dm†​Dm​. Adjoints ∗{}^*∗ and Moore–Penrose pseudoinverses †{}^\dagger† are taken for the inner products (7) on C0C_0C0​, C1C_1C1​ and the unweighted sum inner product ⟨u,v⟩=∑m⟨um,vm⟩0\langle u,v\rangle = \sum_m\langle u^m,v^m\rangle_0⟨u,v⟩=∑m​⟨um,vm⟩0​ on C0MC_0^{\mathcal M}C0M​.

Lemma 4.4 (ii). For every player jjj,

(∑iDi)†Dj=(∑iDi∗Di)†Dj∗Dj.\Big(\sum_i D_i\Big)^\dagger D_j = \Big(\sum_i D_i^* D_i\Big)^\dagger D_j^* D_j .(i∑​Di​)†Dj​=(i∑​Di∗​Di​)†Dj∗​Dj​.

Since ∑iDi=δ0\sum_i D_i = \delta_0∑i​Di​=δ0​ and ∑iDi∗Di=Δ0\sum_i D_i^*D_i = \Delta_0∑i​Di∗​Di​=Δ0​, this rewrites the potential function of Theorem 4.1 as φ=Δ0†∑mΔ0,mum\varphi = \Delta_0^\dagger\sum_m \Delta_{0,m}u^mφ=Δ0†​∑m​Δ0,m​um, which involves only operators on C0C_0C0​.

Formalization Note The five identities of Lemma 4.4 are split into five theorem items; this one is part (ii). The paper's full statement is:

Lemma 4.4. The pseudoinverses of operators DmD_mDm​ and DDD satisfy the following identities: (i) Dm†=1hmDm∗D_m^\dagger = \frac{1}{h_m}D_m^*Dm†​=hm​1​Dm∗​, (ii) (∑iDi)†Dj=(∑iDi∗Di)†Dj∗Dj(\sum_i D_i)^\dagger D_j = (\sum_i D_i^*D_i)^\dagger D_j^*D_j(∑i​Di​)†Dj​=(∑i​Di∗​Di​)†Dj∗​Dj​, (iii) D†=[D1†;… ;DM†]D^\dagger = [D_1^\dagger;\dots;D_M^\dagger]D†=[D1†​;…;DM†​], (iv) Π=D†D\Pi = D^\dagger DΠ=D†D, (v) DD†δ0=δ0DD^\dagger\delta_0 = \delta_0DD†δ0​=δ0​.

The left-hand side is stated with the sum ∑iDi\sum_i D_i∑i​Di​, as printed (not with δ0\delta_0δ0​).

Preamble
import Mathlib
import Definitions.Def_HarmonicGames_Decomposition_Pinv
import Definitions.Def_HarmonicGames_Decomposition_Flows
import Definitions.Def_HarmonicGames_Decomposition_Games
Formal statement
namespace HarmonicGames.Decomposition

theorem lemma_4_4_ii (ι : Type) [Fintype ι] [DecidableEq ι] (E : ι → Type) [∀ m, Fintype (E m)]
    [∀ m, DecidableEq (E m)] [∀ m, Nonempty (E m)] (j : ι) :
    pinv (∑ i, Dm E i) ∘ₗ Dm E j =
      pinv (∑ i, LinearMap.adjoint (Dm E i) ∘ₗ Dm E i) ∘ₗ
        (LinearMap.adjoint (Dm E j) ∘ₗ Dm E j) := by sorry

end HarmonicGames.Decomposition
Source
Candogan, Menache, Ozdaglar, Parrilo, Flows and Decompositions of Games: Harmonic and Potential Games, arXiv:1005.2405v2, p. 15, Lemma 4.4 (ii)
Human review
  • Endorsed by Shuze Chen · Oct 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 7, 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