{"id":"arxiv/2609.29521","url":"https://arxiv.org/abs/2609.29521","tags":["math.OC"],"score":0,"title":"Envelopt: Constrained Convex Composite Optimization","domain":"math.OC","source":"arxiv","authors":["Alberto De Marchi","Dominique Orban"],"reviews":0,"subject":"math.OC","summary":"We introduce Envelopt, a globally convergent iterative framework for a broad class of structured optimization problems where a smooth objective is augmented by a nonsmooth convex regularizer composed with a smooth mapping, and the variables are subject to general smooth constraints. All smooth functions may be nonconvex. The method is akin to an augmented-Lagrangian method in which partial minimization with respect to a lifting variable results in smooth subproblems involving the Moreau envelope of the nonsmooth regularizer, and the original constraints are retained explicitly. Only the…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-08-25","reactions":{}}