Nearly Optimal Linear Convergence of Stochastic Primal-Dual Methods for Linear Programming
arxiv(2021)
摘要
There is a recent interest on first-order methods for linear programming
(LP). In this paper,we propose a stochastic algorithm using variance reduction
and restarts for solving sharp primal-dual problems such as LP. We show that
the proposed stochastic method exhibits a linear convergence rate for solving
sharp instances with a high probability. In addition, we propose an efficient
coordinate-based stochastic oracle for unconstrained bilinear problems, which
has 𝒪(1) per iteration cost and improves the complexity of the
existing deterministic and stochastic algorithms. Finally, we show that the
obtained linear convergence rate is nearly optimal (upto log terms) for a
wide class of stochastic primal dual methods.
更多查看译文
关键词
linear programming,optimal linear convergence,primal-dual
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要