Constituent grammatical evolution

Constituent Grammatical Evolution
Constituent Grammatical Evolution (CGE) is a proposed variation by Georgiou and Teahan (see section Publications) of Grammatical Evolution which incorporates the concepts of constituent genes and conditional behavior-switching. CGE targets agent-oriented problems, that is problems involving the evolution and emergence of the behaviour of an agent. Such well-known problems are artificial ant and maze searching.
It has been demonstrated, as claimed by the inventors (see links in the next section), that CGE outperforms Grammatical Evolution in four benchmark problems when the search spaces as defined by the original definitions of the problems are used: Santa Fe Trail, Los Altos, Hills, Hampton Court Maze, and Chevening House Maze. According to Georgiou and Teahan (2011) CGE achieves higher success rate and finds better solutions than Grammatical Evolution in all these benchmarks.
The algorithm was fist presented in the Twenty-Second International Joint Conference on Artificial Intelligence (Barcelona, July 2011).
CGE Description and Benchmarking
# [http://ijcai-11.iiia.csic.es/files/slides/1320-CGE%20-%20IJCAI%20Presentation%20v2.0.ppt Constituent Grammatical Evolution presentation in IJCAI 2011]
#
Web Site
The official web site of CGE is under construction.
Publications
* Georgiou, L. and Teahan, W. J. (2011) “Constituent Grammatical Evolution”. In Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence (IJCAI 2011), July 16-22, 2011, Barcelona, Catalonia (Spain), 1261-1268.
* Georgiou, L. and Teahan, W. J. (2010) “Grammatical Evolution and the Santa Fe Trail Problem”. In Proceedings of the International Conference on Evolutionary Computation (ICEC 2010), October 24-26, 2010, Valencia, Spain, 10-19.

Main Menu

See Also

Main Menu