For the current REF see the REF 2021 website REF 2021 logo

Output details

11 - Computer Science and Informatics

Aston University

Return to search Previous output Next output
Output 25 of 68 in the submission
Article title

Genetic algorithms for discovery of matrix multiplication methods

Type
D - Journal article
Title of journal
IEEE transactions on evolutionary computation
Article number
6151102
Volume number
16
Issue number
5
First page of article
749
ISSN of journal
1089-778X
Year of publication
2012
Number of additional authors
2
Additional information

<22> More efficient matrix multiplication methods would have huge impact on practical implementation of matrix calculations used in mathematical software packages. For square matrices of size 3, there are no known practical methods reaching the theoretical lower bound. The new algorithm discovered by genetic algorithms is one of very few approximate methods. Discovery of an exact method would imply that multiplication of large matrices would switch to use this instead of the algorithm used since 1968. The authors were invited to talk at the Genetic and Evolutionary Computation conference (2011), after being shortlisted for the Human-Competitive Results competition.

Interdisciplinary
-
Cross-referral requested
-
Research group
B - Computer Science Research Group
Citation count
0
Proposed double-weighted
No
Double-weighted statement
-
Reserve for a double-weighted output
No
Non-English
No
English abstract
-