Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Strict feasibility, duality, and both optima attained

Open
QAlgorithms.Papers.Independent.Intakenext100.P98.blockSdp_strictFeasibility_attainment

by Gohan · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let X=CnX=\mathbb C^nX=Cn, Y=CmY=\mathbb C^mY=Cm with n,m≥1n,m\ge1n,m≥1, and let Φ:L(X)→L(Y)\Phi:L(X)\to L(Y)Φ:L(X)→L(Y) be complex-linear with a Stinespring pair A,B:X→Y⊗ZA,B:X\to Y\otimes ZA,B:X→Y⊗Z of minimal environment dimension dim⁡Z=rank⁡J(Φ)\dim Z=\operatorname{rank}J(\Phi)dimZ=rankJ(Φ). Use the concrete block triple (Ψ,C,D)(\Psi,C,D)(Ψ,C,D) defined on PDF page 7, with Ψ(P)=diag⁡(Tr⁡PXX,Tr⁡Y(PYZ,YZ−APXXA∗))\Psi(P)=\operatorname{diag}(\operatorname{Tr}P_{XX},\operatorname{Tr}_Y(P_{YZ,YZ}-AP_{XX}A^*))Ψ(P)=diag(TrPXX​,TrY​(PYZ,YZ​−APXX​A∗)), C=diag⁡(0,BB∗)C=\operatorname{diag}(0,BB^*)C=diag(0,BB∗) and D=diag⁡(1,0)D=\operatorname{diag}(1,0)D=diag(1,0). There are operators P0≻0P_0\succ0P0​≻0 and S0≻0S_0\succ0S0​≻0 such that

D−Ψ(P0)≻0,Ψ∗(S0)−C≻0.D-\Psi(P_0)\succ0,\qquad\Psi^*(S_0)-C\succ0.D−Ψ(P0​)≻0,Ψ∗(S0​)−C≻0.

The primal supremum α\alphaα and dual infimum β\betaβ are equal, finite, and both attained: there exist primal and dual feasible P,SP,SP,S with

Re⁡Tr⁡(C∗P)=α=β=Re⁡Tr⁡(D∗S).\operatorname{Re}\operatorname{Tr}(C^*P)=\alpha=\beta=\operatorname{Re}\operatorname{Tr}(D^*S).ReTr(C∗P)=α=β=ReTr(D∗S).

This includes the zero map with its zero-dimensional minimal environment. The result supplies both optimality certificates for this particular SDP.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P98
Formal statement
noncomputable section

open scoped Matrix ComplexConjugate ComplexOrder

namespace QAlgorithms.Papers.Independent.Intakenext100.P98

/-- PDF pp.8,10--11: strict feasibility on both sides, strong duality, and
attainment of both finite optima for the concrete p.7 triple. An empty minimal
environment is allowed; no minimum eigenvalue of an empty matrix is used. -/
theorem blockSdp_strictFeasibility_attainment {n m r : ℕ} (hn : 0 < n) (hm : 0 < m)
    (Phi : (Matrix (Fin n) (Fin n) ℂ →ₗ[ℂ] Matrix (Fin m) (Fin m) ℂ)) (A B : (Matrix (Fin m × Fin r) (Fin n) ℂ))
    (hpair : stinespringPairRepresents Phi A B) (hminimal : r = (watrousChoiMatrix Phi).rank) :
    (∃ P : (Matrix (PrimalIndex n m r) (PrimalIndex n m r) ℂ),
      P.PosDef ∧ (stinespringSdpBound r - stinespringSdpMap A P).PosDef) ∧
    (∃ S : (Matrix (DualIndex r) (DualIndex r) ℂ),
      S.PosDef ∧ (blockAdjoint A S - stinespringSdpObjective B).PosDef) ∧
    blockPrimalValue A B = blockDualValue A B ∧
    ∃ (P : (Matrix (PrimalIndex n m r) (PrimalIndex n m r) ℂ)) (S : (Matrix (DualIndex r) (DualIndex r) ℂ)),
      BlockPrimalFeasible A P ∧ BlockDualFeasible A B S ∧
      (realTracePairing (stinespringSdpObjective B) P : EReal) = blockPrimalValue A B ∧
      (realTracePairing (stinespringSdpBound r) S : EReal) = blockDualValue A B := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P98

end
Source
Semidefinite programs for completely bounded norms, arXiv:0901.4709v2, PDF p. 8, §3, unnumbered: strict feasibility, strong duality and attainment
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

Let n,m,rn,m,rn,m,r be natural numbers with n>0n>0n>0 and m>0m>0m>0, let Φ:Mn(C)→Mm(C)\Phi:M_n(\mathbb C)\to M_m(\mathbb C)Φ:Mn​(C)→Mm​(C) be any complex-linear map, and let A,B:Cn→Cm⊗CrA,B:\mathbb C^n\to\mathbb C^m\otimes\mathbb C^rA,B:Cn→Cm⊗Cr satisfy Φ(X)=Tr⁡r(AXB†)\Phi(X)=\operatorname{Tr}_{r}(AXB^\dagger)Φ(X)=Trr​(AXB†) for every XXX, and r=rank⁡J(Φ)r=\operatorname{rank}J(\Phi)r=rankJ(Φ), where J(Φ)(a,i),(b,j)=Φ(Eij)abJ(\Phi)_{(a,i),(b,j)}=\Phi(E_{ij})_{ab}J(Φ)(a,i),(b,j)​=Φ(Eij​)ab​, EijE_{ij}Eij​ is the matrix unit, and (Tr⁡rT)ab=∑q=0r−1T(a,q),(b,q)(\operatorname{Tr}_{r}T)_{ab}=\sum_{q=0}^{r-1}T_{(a,q),(b,q)}(Trr​T)ab​=∑q=0r−1​T(a,q),(b,q)​. For a matrix PPP on Cn⊕(Cm⊗Cr)\mathbb C^n\oplus(\mathbb C^m\otimes\mathbb C^r)Cn⊕(Cm⊗Cr), with diagonal blocks X,YX,YX,Y, and a matrix SSS on C⊕Cr\mathbb C\oplus\mathbb C^rC⊕Cr, with upper-left scalar zzz and lower-right block ZZZ, set

RA(P)=(1−tr⁡X00Tr⁡m(AXA†−Y)),QA,B(S)=(zIn−A†(Im⊗Z)A00Im⊗Z−BB†),R_A(P)= \begin{pmatrix} 1-\operatorname{tr}X&0\\ 0&\operatorname{Tr}_{m}(AXA^\dagger-Y) \end{pmatrix}, \qquad Q_{A,B}(S)= \begin{pmatrix} zI_n-A^\dagger(I_m\otimes Z)A&0\\ 0&I_m\otimes Z-BB^\dagger \end{pmatrix},RA​(P)=(1−trX0​0Trm​(AXA†−Y)​),QA,B​(S)=(zIn​−A†(Im​⊗Z)A0​0Im​⊗Z−BB†​),

where (Tr⁡mT)ij=∑a=0m−1T(a,i),(a,j)(\operatorname{Tr}_{m}T)_{ij}=\sum_{a=0}^{m-1}T_{(a,i),(a,j)}(Trm​T)ij​=∑a=0m−1​T(a,i),(a,j)​. Define ppp to be the extended-real supremum of Re⁡tr⁡(BB†Y)\operatorname{Re}\operatorname{tr}(BB^\dagger Y)Retr(BB†Y) over P⪰0P\succeq0P⪰0 with RA(P)⪰0R_A(P)\succeq0RA​(P)⪰0, and ddd to be the extended-real infimum of Re⁡z\operatorname{Re}zRez over S⪰0S\succeq0S⪰0 with QA,B(S)⪰0Q_{A,B}(S)\succeq0QA,B​(S)⪰0, embedding each real objective into the extended reals. The theorem asserts all of the following: there exists PPP with P≻0P\succ0P≻0 and RA(P)≻0R_A(P)\succ0RA​(P)≻0; there exists SSS with S≻0S\succ0S≻0 and QA,B(S)≻0Q_{A,B}(S)\succ0QA,B​(S)≻0; p=dp=dp=d; and there exist matrices P∗,S∗P_*,S_*P∗​,S∗​, with corresponding diagonal blocks X∗,Y∗X_*,Y_*X∗​,Y∗​ and z∗,Z∗z_*,Z_*z∗​,Z∗​, such that

P∗⪰0,RA(P∗)⪰0,S∗⪰0,QA,B(S∗)⪰0,Re⁡tr⁡(BB†Y∗)=p,Re⁡z∗=d,P_*\succeq0,\quad R_A(P_*)\succeq0,\quad S_*\succeq0,\quad Q_{A,B}(S_*)\succeq0,\quad \operatorname{Re}\operatorname{tr}(BB^\dagger Y_*)=p,\quad \operatorname{Re}z_*=d,P∗​⪰0,RA​(P∗​)⪰0,S∗​⪰0,QA,B​(S∗​)⪰0,Retr(BB†Y∗​)=p,Rez∗​=d,

with the final equalities understood in the extended reals. Here ⪰0\succeq0⪰0 and ≻0\succ0≻0 mean Hermitian positive semidefinite and Hermitian positive definite, respectively. No positivity or channel assumption is imposed on Φ\PhiΦ, and r=0r=0r=0 is allowed, using empty blocks and sums.

Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 31, 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