A generalized MATLAB-based distributed-computing optimization tool

Author

Lysiak, Keith ; Polendo, Jason

Author_Institution

Southwest Res. Inst., San Antonio, TX, USA

fYear

2005

fDate

3-7 April 2005

Firstpage

170

Lastpage

173

Abstract

A distributed computing optimization architecture was developed by SwRI for use with numerical solvers. This optimization toolbox is very generalized and can be easily adapted to use with any numerical solver having a well-described input and output file system. In combination with the use of MATLAB, this makes it a very flexible and powerful analysis and development tool. The tool employs an optimization routine, input file modification routines to generate numerous solver input files, and output extraction routines to extract the pertinent solver output data. Specifically, advanced development has been done using the numerical electromagnetic code (NEC) ante wire/plate with dielectrics (WIPL-D), genetic algorithm optimization toolbox (GAOT) and CONDOR^® to demonstrate proof-of-concept.

Keywords

distributed algorithms; genetic algorithms; mathematics computing; software tools; generalized MATLAB-based distributed-computing optimization tool; genetic algorithm optimization; input file system; numerical electromagnetic code; output file system; Computer architecture; Computer languages; Data mining; Dielectrics; Distributed computing; File systems; Genetic algorithms; MATLAB; National electric code; Wire;

fLanguage

English

Publisher

ieee

Conference_Titel

Wireless Communications and Applied Computational Electromagnetics, 2005. IEEE/ACES International Conference on

Print_ISBN

0-7803-9068-7

Type

conf

DOI

10.1109/WCACEM.2005.1469554

Filename

1469554