Published on:
14 March 2024
Primary Category:
Optimization and Control
Paper Authors:
Juyoung Jeong,
David Sossa
Extends commutation principles to nonsmooth objectives
Local minimizers operator commute with subdifferential elements under regularity
Local maximizers operator commute with subdifferential elements
Principles allow optimizing shifted strictly convex functions
Simplifying and summarizing key insights on commutation principles for nonsmooth optimization
This paper expands on previous work establishing 'commutation principles' connecting the structure of solutions to optimization problems over Euclidean Jordan algebras with spectral functions/sets. It shows these principles can be extended to problems with nonsmooth objective functions. Specifically, with mild regularity assumptions, local minimizers/maximizers operator commute with elements of certain generalized subdifferentials of the objective.
Key insights on non-harmonic pseudo-differential operators
Self-adjoint operators approximable by commuting ones
Key results on compactness of commutators
Symmetrization for general nonlocal linear elliptic problems
Existence of strong solutions for free discontinuity problems in linear elasticity
Symmetrizations reveal key results for nonlocal equations
No comments yet, be the first to start the conversation...
Sign up to comment on this paper