  


An inexact accelerated proximal gradient method for large scale linearly constrained convex SDP
K.F. Jiang (kaifengjiangnus.edu.sg) Abstract: The accelerated proximal gradient (APG) method, first proposed by Nesterov, and later refined by Beck and Teboulle, and studied in a unifying manner by Tseng has proven to be highly efficient in solving some classes of large scale structured convex optimization (possibly nonsmooth) problems, including nuclear norm minimization problems in matrix completion and $l_1$ minimization problems in compressed sensing. The method has superior worstcase iteration complexity over the classical projected gradient method, and usually has good practical performance on problems with appropriate structures. In this paper, we extend the APG method to the inexact setting where the subproblem in each iteration is only solved approximately, and show that it enjoys the same worstcase iteration complexity as the exact counterpart if the subproblems are progressively solved to sufficient accuracy. We apply our inexact APG method to solve large scale convex quadratic semidefinite programming (QSDP) problems of the form: $\min\{ \frac{1}{2}\inprod{x}{\mathcal{Q}(x)} + \inprod{c}{x}\mid \mathcal{A}(x) = b, x\succeq 0\}$, where $\mathcal{Q},\mathcal{A}$ are given linear maps and $b,c$ are given data. The subproblem in each iteration is solved by a semismooth NewtonCG (SSNCG) method with warmstart using the iterate from the previous iteration. Our APGSSNCG method is demonstrated to be efficient for QSDP problems whose positive semidefinite linear maps $\mathcal{Q}$ are highly illconditioned or rank deficient. Keywords: Inexact accelerated proximal gradient method, Semismooth Newton CG method, Convex quadratic SDP, Nearest correlation matrix estimation Category 1: Convex and Nonsmooth Optimization (Convex Optimization ) Category 2: Linear, Cone and Semidefinite Programming (Semidefinite Programming ) Citation: SIAM J. Optimization, to appear. Download: [PDF] Entry Submitted: 09/09/2011 Modify/Update this entry  
Visitors  Authors  More about us  Links  
Subscribe, Unsubscribe Digest Archive Search, Browse the Repository

Submit Update Policies 
Coordinator's Board Classification Scheme Credits Give us feedback 
Optimization Journals, Sites, Societies  