Output details
11 - Computer Science and Informatics
University of Kent
Output title
Semantically driven crossover in genetic programming
Type
E - Conference contribution
Name of conference/published proceedings
IEEE Congress on Evolutionary Computation, 2008. CEC 2008. (IEEE World Congress on Computational Intelligence).
Volume number
-
Issue number
-
First page of article
111
ISSN of proceedings
-
Year of publication
2008
URL
-
Number of additional authors
1
Additional information
<24> This paper was the first to introduce semantic methods in genetic programming. We show that by avoiding semantically-redundant crossover, the speed of learning by these algorithms can be accelerated. We provide a substantial analysis of results against other established methods, and statistical techniques are used to determine which problem types the new method works well on. We followed this with further semantic methods in papers for CEC2009 and GPEM Journal, 2009. This work has led to a number of papers by other research groups (e.g. UCD, TU Poznan) and a PhD at Dublin has extended the work.
Interdisciplinary
-
Cross-referral requested
-
Research group
I - Computational Intelligence Group
Citation count
40
Proposed double-weighted
No
Double-weighted statement
-
Reserve for a double-weighted output
No
Non-English
No
English abstract
-
