Lemma 3, p. 12 — the group commutator of exp(iF), exp(iG) is exp(i·i[F,G]) up to O(δ³)
ProvedSKAlgorithm.Qudit.lemma_3group-commutatormatrix-exponentialp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1solovay-kitaev
There is a constant with the following property. Let and be Hermitian matrices (any ) with . Then
where .
The paper gives . With Lemma 2 this shows that a gate near the identity is approximated to third order by a group commutator of two gates closer to the identity.
Formalization Note. The constant is existential and is chosen before the dimension and the matrices: the paper's "some constant " is a number, not a function of . The value is not pinned. The exponential is NormedSpace.exp, and is kept in the paper's form exp (I • (I • (F * G - G * F))).
Preamble
import Mathlib import Definitions.Def_SKAlgorithm_Qudit_Model open scoped Matrix open QAlgorithms
Formal statement
namespace SKAlgorithm.Qudit
theorem lemma_3 :
∃ c₁ : ℝ, 0 < c₁ ∧ ∀ (d : ℕ) (F G : Matrix (Fin d) (Fin d) ℂ) (δ : ℝ),
F.IsHermitian → G.IsHermitian → specNorm F < δ → specNorm G < δ →
specNorm (NormedSpace.exp (Complex.I • F) * NormedSpace.exp (Complex.I • G) *
NormedSpace.exp ((-Complex.I) • F) * NormedSpace.exp ((-Complex.I) • G) -
NormedSpace.exp (Complex.I • (Complex.I • (F * G - G * F)))) ≤ c₁ * δ ^ 3 := by sorry
end SKAlgorithm.Qudit
Source
Dawson and Nielsen, The Solovay-Kitaev algorithm, arXiv:quant-ph/0505030v2, p. 12, §5.2, Lemma 3, Eq. (32)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.