Fejér processes in theory and practice: recent results (Q735923)

This article is a survey on Fejér mappings and related iterative methods, as treated in the monographs [\textit{I. I. Eremin} and \textit{V. D. Mazurov}, Nonstationary processes of mathematical programming. (Russian). Moskva: ``Nauka'' (1979; Zbl 0447.90051); \textit{V. V. Vasin} and \textit{I. I. Eremin}, Operators and iterative processes of Fejér type. Theory and applications. Berlin: Walter de Gruyter (2009; Zbl 1245.65061)]. The authors describe applications of these concepts to structured systems of linear and convex inequalities, operator equations, and inconsistent (unsolvable) linear and quadratic problems. The first section introduces the concept of Fejér mappings and processes, and in Section 2 the use of these concepts for systems of inequalities is illustrated, pointing out to the application to problems for which there is no a-priori information. Section 3 describes the connection between Fejér processes and randomization, while Sections 4 and 5 are devoted to the use of Fejér mappings for solving Linear and Quadratic Programming problems, respectively. Section 6 and 7 are devoted to the use of the Fejér process for solving improper (unsolvable) mutually dual LP problems, including issues of Decomposition and Parallel Computations.