Theorem 1 — convergence or unboundedness of the Bregman proximal-point run
ProvedBregmanPPA.Convergence.theorem1Let be maximal monotone on a finite-dimensional real inner-product space. Let be a Bregman function with zone , with . Let be positive and bounded below by a positive constant, and let be an infinite Bregman proximal-point run. Assume either (C1) , or (C2) for a proper lower semicontinuous convex function . Then:
This is the paper's central convergence and divergence alternative for nonlinear proximal-point algorithms.
Formalization Note The space is a finite-dimensional real inner-product space, representing . The run explicitly keeps every iterate in and expresses (3) through the equivalent inclusion (4). The standing inclusion differs from (C1), . Unbounded means the range of the full sequence is not bounded.
import Mathlib import Definitions.Def_BregmanPPA_Convergence_Model open ThreeOpSplitting.Convergence InertialFB.IFB Filter Topology
namespace BregmanPPA.Convergence
variable {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℝ H] [FiniteDimensional ℝ H]
/-- Theorem 1: convergence to a zero, or unboundedness when there are no zeros under (C1). -/
theorem theorem1 (T : H → Set H) (S : Set H) (h : H → ℝ) (c : ℕ → ℝ) (x : ℕ → H)
(hT : IsMaximalMonotone T) (hh : IsBregmanFunction S h)
(hdom : dom T ⊆ closure S) (hc : ∀ k, 0 < c k)
(hcinf : ∃ ε : ℝ, 0 < ε ∧ ∀ k, ε ≤ c k)
(hx : IsBregmanPPARun S h T c x)
(hC : closure (dom T) ⊆ S ∨ ∃ f : H → EReal,
IsProperFn f ∧ IsConvexFn f ∧ LowerSemicontinuous f ∧ T = subdiffOp f) :
((zer T).Nonempty → ∃ z : H, z ∈ zer T ∧ Tendsto x atTop (𝓝 z)) ∧
(zer T = ∅ → closure (dom T) ⊆ S → ¬ Bornology.IsBounded (Set.range x)) := by sorry
end BregmanPPA.Convergence
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.