§5, p. 220 — the saddle operator K(x, p) = ∂ₓl × (−∂̂ₚl) is maximal monotone
ProvedBregmanPPA.ProxMult.K_maximallagrangianmaximal-monotonep2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1saddle-point
Consider a problem of the form (10): closed, nonempty and convex, proper lower semicontinuous convex and finite on . Let be its Lagrangian and
the set-valued map on . Then is maximal monotone.
The paper derives this from the fact that is a closed saddle function, convex in and concave in , citing Rockafellar's Convex Analysis, Corollary 37.5.2. Maximality is the hypothesis under which Theorem 1 is applied to .
Formalization Note is WithLp 2 (E n × E m). Monotonicity and maximality are the published IsMonotoneOp/IsMaximalMonotone (graph-maximality).
Preamble
import Mathlib import Definitions.Def_BregmanPPA_Convergence_Model import Definitions.Def_BregmanPPA_IneqMult_Program import Definitions.Def_BregmanPPA_ProxMult_Saddle open ThreeOpSplitting.Convergence
Formal statement
namespace BregmanPPA.ProxMult
/-- §5, p. 220: for a problem of the form (10), the operator
`K(x, p) = ∂ₓ l(x, p) × (−∂̂ₚ l(x, p))` on `ℝⁿ⁺ᵐ` is maximal monotone. -/
theorem K_maximal {n m : ℕ} (C : Set (BregmanPPA.IneqMult.E n)) (f : BregmanPPA.IneqMult.E n → EReal) (g : Fin m → BregmanPPA.IneqMult.E n → EReal)
(hP : BregmanPPA.IneqMult.IsConvexProgram C f g) :
IsMaximalMonotone (opK C f g) := by sorry
end BregmanPPA.ProxMult
Source
Eckstein, Nonlinear proximal point algorithms using Bregman functions, with applications to convex programming, Math. Oper. Res. 18(1) (1993), p. 220, §5 (K is maximal monotone [26, Corollary 37.5.2])
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.