{"id":"arxiv/2609.35946","url":"https://arxiv.org/abs/2609.35946","tags":["math.OC"],"score":0,"title":"Gradient-Free Methods for Stochastic Convex Optimization with Stochastic Functional Constraints","domain":"math.OC","source":"arxiv","authors":["Vadim Abronin","Alexander Gasnikov","Darina Dvinskikh"],"reviews":0,"subject":"math.OC","summary":"We develop accelerated gradient-free methods for stochastic convex optimization with constraints defined by expectations. Our batched primal-dual sliding method uses two-point evaluations sharing a random sample and guarantees expected objective error and expected maximum constraint violation at most $\\varepsilon$. It achieves $O(\\varepsilon^{-1/2})$ sequential oracle rounds for smooth data and $O(d^{1/4}/\\varepsilon)$ for nonsmooth data in dimension $d$, recovering the accuracy and dimension dependence of the corresponding unconstrained accelerated methods. The smooth rate is optimal in…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-09-28","reactions":{}}