§4.2, (13) — the regularized problem min (f(x) + ‖x − xⁿ‖² | x ∈ C) has a unique solution
ProvedMartinetReg.ConvexMin.step_existsUniqueconvex-optimizationhilbert-spacep2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1proximal-point
Under the hypotheses of §4.1 ( a real Hilbert space, nonempty, closed and convex, convex and lower semicontinuous with bounded sublevel sets ), for every the regularized problem
has exactly one solution .
This is the claim built into the definition (13) of the algorithm (" solution unique de …"): it guarantees that the sequence of §4.2 is well defined.
Formalization Note The paper applies (13) to , and the statement quantifies over exactly those . The weight on the squared distance is , as printed.
Preamble
import Mathlib import Definitions.Def_MartinetReg_ConvexMin_Setting
Formal statement
namespace MartinetReg.ConvexMin
/-- §4.2, (13), p. 157: for xⁿ ∈ C, the regularized problem min (f(x) + ‖x − xⁿ‖² | x ∈ C) has
exactly one solution xⁿ⁺¹ ∈ C. -/
theorem step_existsUnique {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℝ H]
[CompleteSpace H] (f : H → ℝ) (C : Set H) (hS : Standing f C) :
∀ xn ∈ C, ∃! y : H, IsProxStep f C xn y := by sorry
end MartinetReg.ConvexMin
Source
Martinet, Régularisation d'inéquations variationnelles par approximations successives, R.I.R.O. 4(R-3) (1970), p. 157, §4.2, (13)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.