Tuning Artificial Neural Networks Parameters Using an Evolutionary Algorithm

2008 ◽  
Author(s):  
Leandro M. Almeida ◽  
Teresa B. Ludermir
Author(s):  
Anthony M. Roy ◽  
Erik K. Antonsson ◽  
Andrew A. Shapiro

Genetic Programs that have phenotypes created by the application of genotypes comprising rules are robust and highly scalable. Such encodings are useful for complex applications such as controller design. This paper outlines an evolutionary algorithm capable of creating a controller for 2 DOF, path following robot. The controllers are embodied by Artificial Neural Networks capable of full functionality despite multiple failures.


Sign in / Sign up

Export Citation Format

Share Document