Proof of Theorem 2.9, p. 9 — (A + B)⁻¹(0) has at most one element when A is strongly monotone and B monotone
ProvedFRBSplitting.Linear.zeroSet_subsingletonmonotone-inclusionp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1strong-monotonicity
Let be a real inner product space, -strongly monotone with , and monotone. Then the zero set
contains at most one point.
In Theorem 2.9 this gives the uniqueness of the limit: the zero set is assumed nonempty, so it is a singleton, and the iterates converge to its only element.
Formalization Note. The paper obtains uniqueness at the end of the proof of Theorem 2.9, from the convergence of a single run to every zero. The statement here needs neither the method nor Lipschitz continuity nor maximality, and is stated without them (a stronger statement).
Preamble
import Mathlib import Definitions.Def_ThreeOpSplitting_Accel_MonotoneOperators import Definitions.Def_FRBSplitting_Linear_Setting open InnerProductSpace ThreeOpSplitting.Accel
Formal statement
namespace FRBSplitting.Linear
theorem zeroSet_subsingleton {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℝ H]
(A : H → Set H) (B : H → H) (m : ℝ)
(hm : 0 < m) (hAs : IsStronglyMonotoneOp m A)
(hBm : IsMonotoneFun B) :
(FRBSplitting.Weak.zeroSet A B).Subsingleton := by sorry
end FRBSplitting.Linear
Source
Malitsky & Tam, A Forward-Backward Splitting Method for Monotone Inclusions Without Cocoercivity, arXiv:1808.04162v4, p. 9, proof of Theorem 2.9, last sentence
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.