RUS  ENG
Full version
JOURNALS // Izvestiya of Saratov University. Mathematics. Mechanics. Informatics // Archive

Izv. Saratov Univ. Math. Mech. Inform., 2012 Volume 12, Issue 3, Pages 109–117 (Mi isu322)

This article is cited in 1 paper

Computer science

The parallel variant of conditional optimization algorithm with Box complex method

A. N. Savin

Saratov State University, Chair of Discrete Mathematics and Information Technologies

Abstract: This article presents the results of the adaptation algorithm for global extremum searching with presence explicit and implicit constraints complex method created by Box for systems of the distributed and parallel computing. The optimal count of nodes of computing system from the point of view of reliability a global extremum finding and time of its search is defined. At the same time, the run time was several times smaller than on one computer and reliability of searching the global extremum is much higher due to the simultaneous researching of the target function in several areas of the factor space.

Key words: global optimization, conditional optimization, Box complex method, multiextreme criterion function, parallel computing system.

UDC: 681.3.06+681.322

DOI: 10.18500/1816-9791-2012-12-3-109-117



Bibliographic databases:


© Steklov Math. Inst. of RAS, 2024