Ε-subgradient algorithms for bilevel convex optimization*ES Helou was supported by FAPESP grants 2013/07375-0 and 2013/16508-3 and CNPq grant 311476/2014-7. LEA Simões was supported by FAPESP grants 2011/02219-4 and 2013/14615-7. (5th April 2017)

• Record Type: Journal Article
• Authors: Helou, Elias S; Simões, Lucas E A
• Is Part Of: Inverse problems. Volume 33:Number 5(2017:May)
• Publication Date: 2017-04-05
• Subjects: ϵ-subgradient methods -- bilevel optimization -- tomographic image reconstruction; Inverse problems (Differential equations) -- Periodicals; 515.357
• DOI: 10.1088/1361-6420/aa6136
• Languages: English
• ISSNs: 0266-5611

Abstract

Abstract: This paper introduces and studies the convergence properties of a new class of explicit ϵ-subgradient methods for the task of minimizing a convex function over a set of minimizers of another convex minimization problem. The general algorithm specializes to some important cases, such as first-order methods applied to a varying objective function, which have computationally cheap iterations. We present numerical experimentation concerning certain applications where the theoretical framework encompasses efficient algorithmic techniques, enabling the use of the resulting methods to solve very large practical problems arising in tomographic image reconstruction.