Evolving Modular Programs By Extracting Reusable Functions Using Significance Testing
| dc.contributor | Loeppert, Anthony | en_US |
| dc.date.accessioned | 2007-08-23T01:56:40Z | |
| dc.date.accessioned | 2011-08-24T21:40:28Z | |
| dc.date.available | 2007-08-23T01:56:40Z | |
| dc.date.available | 2011-08-24T21:40:28Z | |
| dc.date.issued | 2007-08-23T01:56:40Z | |
| dc.date.submitted | December 2005 | en_US |
| dc.description.abstract | Genetic programming is an automatic programming method that uses biologically inspired methods to evolve programs. Genetic programming, and evolutionary methods in general, are useful for problem domains in which a method for \emph{constructing} solutions is either not known or infeasible, but a method for \emph{rating} solutions exists. In order to address more complex problem domains, techniques exist to extract functions (modules) automatically during a GP search. This work describes a method to identify useful automatically extracted functions from a GP search to assist subsequent GP searches within the same problem domain, using significance testing. Functions classified as beneficial augment the programmer supplied function set and accelerate the learning rate, by seeding the initial population of a subsequent GP search. | en_US |
| dc.identifier.uri | http://hdl.handle.net/10106/406 | |
| dc.language.iso | EN | en_US |
| dc.publisher | Computer Science & Engineering | en_US |
| dc.title | Evolving Modular Programs By Extracting Reusable Functions Using Significance Testing | en_US |
| dc.type | M.S. | en_US |