{"id":"arxiv/2610.34390","url":"https://arxiv.org/abs/2610.34390","tags":["math.OC","cs.LG"],"score":0,"title":"Controlled Block Proximal Linearized Method for Nonconvex and Nonsmooth Optimization","domain":"math.OC","source":"arxiv","authors":["Weifeng Yang"],"reviews":0,"subject":"math.OC","summary":"We consider multiblock nonconvex and nonsmooth optimization without assuming bounded iterates or coercivity of the original objective. We propose the Controlled Block Proximal Linearized (CBPL) method, which incorporates a control potential into block updates while retaining the original objective, variables, and constraints. Under explicit control design conditions and the KL property, the control yields boundedness, finite length, whole-sequence convergence of the original variables, and convergence of objective values. At the limit, one limiting subgradient of the original objective…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-25","reactions":{}}