Accelerated Primal Dual Method for a Class of Saddle Point Problem with Strongly Convex Component

arXiv1906.07691MaRDI QIDQ6320701FDOQ6320701

Publication date: 18 June 2019

Abstract: This paper presents a simple primal dual method named DPD which is a flexible framework for a class of saddle point problem with or without strongly convex component. The presented method has linearized version named LDPD and exact version EDPD. Each iteration of DPD updates sequentially the dual and primal variable via simple proximal mapping and refines the dual variable via extrapolation. Convergence analysis with smooth or strongly convex primal component recovers previous state-of-the-art results, and that with strongly convex dual component attains full acceleration

O (1 / k^{2})

in terms of primal dual gap. Total variation image deblurring on Gaussian noisy or Salt-Pepper noisy image demonstrate the effectiveness of the full acceleration by imposing the strongly convexity on dual component.

Mathematics Subject Classification ID

Convex programming (90C25) Large-scale problems in mathematical programming (90C06) Optimality conditions and duality in mathematical programming (90C46) Minimax problems in mathematical programming (90C47)

This page was built for publication: Accelerated Primal Dual Method for a Class of Saddle Point Problem with Strongly Convex Component

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6320701)