{"id":"arxiv/2603.20751","url":"https://arxiv.org/abs/2603.20751","tags":["math.OC"],"score":0,"title":"Local Convergence Analysis of ADMM for Nonconvex Composite Optimization","domain":"math.OC","source":"arxiv","authors":["Xiyuan Xie","Lihua Yang","Qia li"],"reviews":0,"subject":"math.OC","summary":"In this paper, we study the local convergence of the standard ADMM scheme for a class of nonconvex composite optimization problems motivated by applications in signal processing and machine learning. The problems are constrained by a closed convex set, while their objective is the sum of a continuously differentiable, possibly nonconvex, smooth term and a polyhedral convex nonsmooth term composed with a linear mapping. Motivated by recent works of Rockafellar, we first provide an elementary proof of a local strong convexity property of the Moreau envelope of polyhedral convex functions on…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-03-21","reactions":{}}