IWPAAMS2007-02: Efecto de la Hipótesis de Lamarck en la Evolución Gramatical Guiada por Refuerzo (IWPAAMS2007-02: The Role Of The Lamarck Hypothesis In The Grammatical Evolution Guided By Reinforcement)

J. M. Mingo (jmingo@inf.uc3m.es), R. Aler (aler@inf.uc3m.es)


Universidad Carlos III de Madrid, Spain
This paper appears in: Revista IEEE América Latina

Publication Date: Oct. 2008
Volume: 6,   Issue: 6 
ISSN: 1548-0992


Abstract:
Grammatical Evolution is an evolutionary algorithm able to develop programs in any language, defined by a grammar. The evolutionary process may be improved if we let the individuals learn during their lifetime. With this aim, the Grammatical Evolution Guided by Reinforcement, an algorithm which merges evolution and learning, was created. Grammatical Evolution Guided by Reinforcement uses a Lamarckian mechanism for replacing the original genotypes when a successful learning has occurred. This paper explores the role of the Lamarckian hypothesis. At the same time, Grammatical Evolution Guided by Reinforcement is tested in a new domain: autonomous navigation in a Kephera robot simulation.

Index Terms:
Grammatical Evolution, Reinforcement Learning, Lamarck Effect   


Documents that cite this document
This function is not implemented yet.


[PDF Full-Text (441)]