Applications of gauge duality in robust principal component analysis and semidefinite programming (Q341322): Difference between revisions

The nonlinear gauge optimization problem is to minimize a closed gauge function \(\kappa\) over a closed convex set \(X \subset \mathbb{R}^n\), i.e. \(\min_x\{\kappa(x)|x\in X\}\). Its nonlinear gauge dual problem is defined as minimizing the polar function \(\kappa^\circ\) over the antipolar set \(X'\), i.e. \(\min_y\{\kappa^\circ(y)|y\in X'\}\), where \(\kappa^\circ(y)=\inf\{\mu>0|\langle x,y\rangle \leq \mu\kappa(x) \text{ for all } x\}\). The authors present new theoretical results on applying the gauge duality theory [\textit{R. M. Freund}, Math. Program. 38, 47--67 (1987; Zbl 0632.90054)] to robust principal component analusis and general semidefinite programming [\textit{M. P. Friedlander} and \textit{I. Macêdo}, SIAM J. Sci. Comput. 38, No. 3, A1616--A1638 (2016; Zbl 1342.90115)]. For each considered problem, they give its gauge dual problem, characterize the optimality conditions for the primal-dual gauge pair and finally they validate a way to recover a primal optimal solution from a dual one.