My view is that GEP is one more tool in the GP arsenal. I wouldn't say it is just an alternate
encoding. It's probably different enough to be labelled a new approach to the same general
GP ideas ... generate some models, test their applicability (fitness), do some sort of evolution of the
models, etc.
It appears to be quite useful for many problems. Is it better than other approaches? I'm
not sure. All I can say is that it's another tool to try with your problems and see for yourself.
The encoding of expressions is different than others have used and because all expressions
generated are 'valid' expressions it appears to run more quickly in many cases. This allows
you to make more runs with your data and this could lead to improved results.
Our view here is that if there is a tool that might work for our data/problem types then let's
try it to see if it has merit. We will generally try several tools with our data.
Bob Orchard
National Research Council Canada Conseil national de recherches Canada
Institute for Information Technology Institut de technologie de l'information
1200 Montreal Road, Building M-50 M50, 1200 chemin Montréal
Ottawa, ON, Canada K1A 0R6 Ottawa (Ontario) Canada K1A 0R6
(613) 993-8557
(613) 952-0215 Fax / télécopieur
[log in to unmask]
Government of Canada | Gouvernement du Canada
-----Original Message-----
From: ECJ Evolutionary Computation Toolkit [mailto:[log in to unmask]] On Behalf Of Magrath, Shane
Sent: Sunday, July 22, 2007 7:22 PM
To: [log in to unmask]
Subject: FW: ADF Examples/Tutorials
Thanks to Bob and Tiberiu for their reply.
1) I've downloaded the book on GEP and have started reading it. Without wanting to be controversial, is GEP something truly novel or is it an alternate encoding for GP? I'm no expert in this, moreover anything which promises a big speedup is all good to me... more reading required on my part, I guess.
2) I've briefly looked at the Lawnmower problem and the Parity problem. I don't have a copy of Koza-II, so can someone give a short explanation of the problem?
Many thanks.
Shane
________________________________
From: ECJ Evolutionary Computation Toolkit [mailto:[log in to unmask]] On Behalf Of Orchard, Bob
Sent: Wednesday, 18 July 2007 10:07 PM
To: [log in to unmask]
Subject: Re: ADF Examples/Tutorials
Have you tried the GEP contribution to ECJ. It supports the Gene Expression Programming
variant of genetic programming developed by Dr. Candida Ferreira. It can be simpler to use
and I believe is faster and produces quite good results in most cases ... There are
quite a few examples included.
also Ferreira's book 'Gene Expression Programming; Mathematical Modeling by an Artificial Intelligence'
is excellent.
Bob Orchard
National Research Council Canada Conseil national de recherches Canada
Institute for Information Technology Institut de technologie de l'information
1200 Montreal Road, Building M-50 M50, 1200 chemin Montréal
Ottawa, ON, Canada K1A 0R6 Ottawa (Ontario) Canada K1A 0R6
(613) 993-8557
(613) 952-0215 Fax / télécopieur
[log in to unmask]
Government of Canada | Gouvernement du Canada
-----Original Message-----
From: ECJ Evolutionary Computation Toolkit [mailto:[log in to unmask]] On Behalf Of Magrath, Shane
Sent: Tuesday, July 17, 2007 11:29 PM
To: [log in to unmask]
Subject: ADF Examples/Tutorials
I've been gradually building up my skills with ECJ Genetic Programming with some success. However, I wondering if there are some examples available of how to use the ADF features of ECJ?
A tutorial would be outstanding - I'm even happy to write it once I get competent in it :)
Thanks in advance
Shane Magrath
IMPORTANT: This email remains the property of the Australian Defence Organisation and is subject to the jurisdiction of section 70 of the CRIMES ACT 1914. If you have received this email in error, you are requested to contact the sender and delete the email.
|