This table is intended to be a comprehensive list of evolutionary algorithm software frameworks that support some flavour of genetic programming. Each entry lists the language the framework is written in, which program representations it supports and whether the software still appears to be being actively developed or not.

If you know of any other genetic programming software that has been omitted from this list, then please leave a comment with details.

Framework License Lang Tree GE CFG-GP CGP GEP Active?
AForge.NET LGPL .NET
Beagle LGPL C++
DEAP LGPL Python
DGPF LGPL Java
DRP GPL Ruby
Discipulus Commercial
ECF ? C++
ECJ AFL Java
EO LGPL C++ ? ? ? ? ?
EpochX LGPL Java
ep4js Apache JScript
Eva2 LGPL Java
Evogen MIT Flex ? ? ? ? ?
GAlib as specified C++
Genetik LGPL Java
GenPro Apache Java ?
GEVA GPL Java ?
GPC++ GPL C++ ?
GPalta GPL Java ?
GPE AFL .NET ? ? ? ? ?
GPLAB ? MATLAB ?
GPTIPS GPL MATLAB
Groovy GPL Java ?
JAGA GPL Java
Java GAlib ? Java ? ? ? ? ?
JCLEC GPL Java ? ? ?
JEF LGPL Java ? ? ? ?
JGAP LGPL/MPL Java ?
jGE GPL Java ?
JRGP GPL Java
lil-GP ? C ?
LAGEP GPL C++ ?
MicroGP GPL C++
PerlGP GPL Perl ?
PMDGP GPL C++ ?
PonyGE GPL Python
PushGP ? C++ ?
PySTEP MIT Python ?
Pyevolve PSF Python
Pyro ? Python ? ? ? ? ?
RMIT GP ? C++ ?
SmallGP GPL C++
TinyGP ? Java
Watchmaker Apache Java

The program representations that are listed are: tree (standard tree GP as described by Koza, including STGP), GE (Grammatical Evolution), CFG-GP (Whigham‘s Context-Free Grammar GP), CGP (Julian Miller’s Cartesian GP) and GEP (Gene-expression Programming). Other unlisted representations may also be supported by the listed software.

A framework is considered to be active still if it has had a release within the past 9 months. If its commit history is visible then it will only be considered active if there have been commits within the past 6 months.

6 Responses to “Which Genetic Programming software? A comparison”

  1. Do you think there is a commercial opportunity for EpocX or is it more of an academic area at the moment?

    • @Michael – EpochX is released under an LGPL license so it can be used within commercial applications, but its primary purpose is for research. A purely commercial GP framework would probably put a little more emphasis on speed and less on extensibility and run statistics.

  2. Just some info: Pyevolve uses the standard Tree GP, but is different in the sense that this tree is compiled to Python bytecode before evaluate the entire fitness points. The license is PSF, the same as Python and is GPL compatible.

  3. GPTIPS is released under the GPL license. It’s still actively being developed, albeit slowly….

Leave a Reply

(required)

(required)

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Spam protection by WP Captcha-Free

© 2012 TC33.org Suffusion theme by Sayontan Sinha