Publication Details

Evolutionary Design of Reconfiguration Strategies to Reduce the Test Application Time

ŠIMÁČEK, J.; SEKANINA, L.; STAREČEK, L. Evolutionary Design of Reconfiguration Strategies to Reduce the Test Application Time. Evolvable Systems: From Biology to Hardware. Lecture Notes in Computer Science. Berlin: Springer Verlag, 2010. p. 214-225. ISBN: 978-3-642-15322-8.
Czech title
Evoluční návrh rekonfiguračních strategií pro redukci počtu testovacích vektorů
Type
conference paper
Language
English
Authors
Šimáček Jiří, Ing., Ph.D.
Sekanina Lukáš, prof. Ing., Ph.D. (DCSY)
Stareček Lukáš, Ing., Ph.D. (RG EHW)
Keywords

evolutionary algorithm, reconfiguration, testing, digital circuit

Abstract

Recently, a method has been presented that allows a significant test application time reduction if some of gates of a digital circuit are reconfigured before test is applied. Selection of the gates for reconfiguration was performed using a very time consuming deterministic recursive search algorithm. In this paper, a new method is proposed for selection of the gates in order to reduce the test application time. The method utilizes an evolutionary algorithm which is able to discover very competitive reconfiguration strategies while the time of optimization is considerably reduced with respect to the original algorithm. Moreover, the user can easily balance the trade off between the number of test vectors and amount of logic that has to be reconfigured. Experimental results are reported for the ISCAS85 benchmark suite.

Published
2010
Pages
214–225
Proceedings
Evolvable Systems: From Biology to Hardware
Series
Lecture Notes in Computer Science
Volume
6274
ISBN
978-3-642-15322-8
Publisher
Springer Verlag
Place
Berlin
BibTeX
@inproceedings{BUT34847,
  author="Jiří {Šimáček} and Lukáš {Sekanina} and Lukáš {Stareček}",
  title="Evolutionary Design of Reconfiguration Strategies to Reduce the Test Application Time",
  booktitle="Evolvable Systems: From Biology to Hardware",
  year="2010",
  series="Lecture Notes in Computer Science",
  volume="6274",
  pages="214--225",
  publisher="Springer Verlag",
  address="Berlin",
  isbn="978-3-642-15322-8",
  url="https://www.fit.vut.cz/research/publication/9300/"
}
Files
Back to top