Affiliation:
1. Chongqing College of Humanities, Science & Technology
2. Southwest University
3. Department of Basic, Chongqing Telecommunications Polytechnic College
Abstract
Abstract
The linearized alternating direction methods of multipliers (L-ADMM) for solving convex minimization problems with two separable blocks in the objective functions is efficient. And its extended version (\(m \geqslant 3\)) is convergent under some mild conditions. Recently, The L-ADMM inspires much attention in analyzing its theoretical convergence rate. However, the research on its convergence rate is still in its infancy. In this paper, we consider the convergence rate of L-ADMM when solving the convex optimization problems that the subdifferentials of the underlying functions are piecewise linear multifunctions. Based on the error bound, we establish the linear convergence rate.
Publisher
Research Square Platform LLC
Reference20 articles.
1. R. Glowinski and A. Marrocco, Approximation par elements finis dordre un et resolution par penalisation-dualite dune classe de problemes nonlineaires, M2AN Math. Model. Numer. Anal, (1975), pp. 41–76.
2. A dual algorithm for the solution of nonlinear variational problems via finite element approximations;Gabay D;Comput. Math. Appl.,1976
3. Splitting algorithms for the sum of two nonlinear operators;Lions PL;SIAM J. Numer. Anal.,1979
4. Ergodic convergence to a zero of the sum of monotone operators in Hilbert space;Passty GB;J. Math. Anal. Appl.,1979
5. The linearized alternating direction method of multipliers for Dantzig selector;Wang X;SIAM J. Sci. Comput.,2012