Rank-one resolvent inversion
ProvedQAlgorithms.Papers.Independent.Intakenext100.P41.resolventInversionLet , , with , and . Assume the §18 simple-spectrum, nonresonance and general-position hypotheses for both and , using eigenbases normalized by . Define where the resolvent exists. For every for which and are invertible,
This is the common-resolvent-domain form of the source rational identity. The hypotheses avoid assigning total-inverse values at poles; the two determinant conditions also imply that the scalar being inverted is nonzero.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P41
noncomputable section
open scoped BigOperators
open MeasureTheory Set
namespace QAlgorithms.Papers.Independent.Intakenext100.P41
/-- §22 inversion lemma, PDF p.28. Pointwise form on the common resolvent
set; the excluded poles are essential because Lean's matrix inverse is total. -/
theorem resolventInversion {d : ℕ} (A : BasicODEData d) (t : ℂ)
(hP : ((-t) • (1 : Matrix (Fin d) (Fin d) ℂ) - A.P).det ≠ 0)
(hR : (t • (1 : Matrix (Fin d) (Fin d) ℂ) -
(linearRankOneMatrix A.v A.phi - A.P)).det ≠ 0) :
resolventScalar (linearRankOneMatrix A.v A.phi - A.P) A.v A.phi t =
(resolventScalar A.P A.v A.phi (-t))⁻¹ := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P41
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be a natural number and let consist of a positive dimension , a matrix , a vector , a complex-linear functional with , families and , and bases and of . Put , where is the standard coordinate vector. Assume that , , and for every ; that every coordinate of in each of these two bases is nonzero; and that, for distinct and every integer , both and . For every satisfying and , one has
Both matrix inverses are of invertible matrices by the stated hypotheses; the outer inverse on the right is complex inversion, whose total convention assigns .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.