Theorem 7 proof — multiplier recursion is a Bregman proximal step
ProvedBregmanPPA.IneqMult.proximal_identitymultiplier-methodp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1proximal-point
Under the standing conditions for problem (10), the Bregman function and the positive step sizes, every sequence satisfying recursion (11) is a Bregman proximal-point run for the subdifferential of the negative dual functional:
This identifies the operator to which Theorem 1 applies.
Formalization Note The multiplier sequence is explicitly kept inside . The dual functional is extended-real valued; the run's monotone conjugate uses real values justified by Lemma 3.
Preamble
import Mathlib import Definitions.Def_BregmanPPA_IneqMult_Run open Filter Topology
Formal statement
namespace BregmanPPA.IneqMult
/-- The last displayed equivalence in the proof of Theorem 7, p. 217: (11) is a
Bregman proximal point run for the subdifferential of the negative dual functional. -/
theorem proximal_identity {n m : ℕ} (C : Set (E n)) (f : E n → EReal)
(g : Fin m → E n → EReal) (S : Set (E m)) (h : E m → ℝ)
(c : ℕ → ℝ) (x : ℕ → E n) (p : ℕ → E m)
(hP : IsConvexProgram C f g)
(hd : ∃ q : E m, dualFn C f g q ≠ ⊥)
(hh : BregmanPPA.Convergence.IsBregmanFunction S h)
(hS : posOrthant m ⊆ S)
(him : posOrthant m ⊆ gradient h '' S)
(hc : ∀ k, 0 < c k)
(hrun : IsIneqMultiplierRun C f g S h c x p) :
BregmanPPA.Convergence.IsBregmanPPARun S h (BregmanPPA.Convergence.subdiffOp (fun q => -(dualFn C f g q))) c p := by sorry
end BregmanPPA.IneqMult
Source
Eckstein, Nonlinear proximal point algorithms using Bregman functions, with applications to convex programming, Math. Oper. Res. 18(1) (1993), p. 217, proof of Theorem 7, final displayed equivalence, https://doi.org/10.1287/moor.18.1.202
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.