Contraction of the PGD update and monotonicity of executed mappings
OpenPGDRevision20261113.pgd_mapping_contractionbregman-geometryconvex-optimizationproximal-gradient
Let be convex and differentiable with an -Lipschitz gradient, where . Let be proper, closed and -strongly convex, with . Fix with , and define
The update map obeys the global contraction bound
Consequently, along any successive updates ,
Thus the best executed mapping and the terminal executed mapping have the same norm. The penalty need not have a non-singleton domain for this contraction statement.
Related results. Fixed-penalty exact curve · Optimal constant stepsizes
Preamble
import Definitions.Def_PGDRevision20261113_Euclidean open PGDRevision20261113
Formal statement
theorem PGDRevision20261113.pgd_mapping_contraction : MappingContractionClaim := by sorry
Source
Author manuscript: Exact Residual Bounds and Optimal Geometry Selection for Proximal Gradient Methods, revised reading edition, 13 November 2026, Lemma 2.2, p. 3; source file sections/02_exact_curve.tex, label lem:lastmapping.