GeneThello: Genetic Othello Files

GeneThello
---------

GeneThello (read jə-ˈne-ˈthe-lō), is an acronym for genetic othello, an othello
(reversi) playing program which based on Genetic Algorithm (GA). In principle
GeneThello consist of an othello program and a genetic algorithm system.


Release Notes
-------------

The current version is 0.7

June 1, 2011, 0.7 is out.

If you find a bug please post it at:
https://sourceforge.net/tracker/?group_id=510590&atid=2075672


General Notes
-------------

GeneThello is hosted at sourceforge.net, with downloads and mailing
list through sourceforge.net

homepage: 	      http://genethello.blogspot.com/
user manual:          http://genethello.sourceforge.net/manual/
javadoc:              http://genethello.sourceforge.net/javadoc/
projectpage: 	      http://sourceforge.net/projects/genethello/
download:             http://sourceforge.net/projects/genethello/files
develop:              http://sourceforge.net/projects/genethello/develop

For a list of people who have helped work on this project please read the 
Authors file

GeneThello 0.7 is released under the GPL version 3, text is in LICENSE


Running GeneThello
-------------------

Generally, you can run GeneThello in two modes:

  'net.sf.genethello.ga.Evolver' to evolve, or
  'net.sf.genethello.gui.GenethelloGUI' to play.

For example, go to your installation directory and type

  'java -cp Genethello.jar net.sf.genethello.ga.Evolver'
  'java -cp Genethello.jar net.sf.genethello.gui.GenethelloGUI'

If you use your own server, see the options using..

  'java -cp Genethello.jar net.sf.genethello.ga.Evolver --help'
  'java -cp Genethello.jar net.sf.genethello.gui.GenethelloGUI --help'


Evolving GeneThello
------------------------

GeneThello is improved by evolving the population using genetic algorithm. For
chromosome with 3 features (type 'oe_fmc', 'oe_dmc', and 'oe_dmp'), the typical
population size is about 352. These individuals are matched in a tournament,
each playing twice with black and white disc alternately, resulting in large
number of match around 352^2 - 352 = 123552 matches, about 6 hours to finish in
my Intel CORE 2 Duo notebook.. :)

For chromosome with 4 features (type 'oe_dmcp'), the typical population size
will be 448, and number of match will be 448^2 - 448 = 200256 matches, it may
take around 9 - 10 hours to finish a tournament.

The result of tournament is reflected in Elo Rating of each player which is
then used as fitness value of each individual. Top rated individuals are then
crossed over to give new improved springs, which are subject to the tournament
of next generation. And the process is repeated again.

To evolve GeneThello, use

  'java -cp Genethello.jar net.sf.genethello.ga.Evolver'

This will connect to server to load the latest generation, evolve the
population, and save back the result to server. Or you may use

  'java -cp Genethello.jar net.sf.genethello.ga.Evolver --nosleep'

This will evolve GeneThello without sleep between matches, such that it will
use your full CPU power. The tournament will take less time to finish, but your
computer will be slowed down to some extent. The typical time above is measured
without using '--nosleep'.


User Manual
-----------

User manual is written using HelpSetMaker an open source for authoring help
document available at

http://www.cantamen.com/helpsetmaker.php

The source code for user manual could be found under /src/helpset directory.


Thanks
------

Thanks to all JGAP development team which create a great open source
for genetic algorithm used by GeneThello. Find JGAP at..
http://jgap.sourceforge.net/
    
Thanks to EVERYONE who has provided ideas and input, without you,
GeneThello would be nothing.


---------------------------------------
if you have any questions please email
bowo prasetyo praNzOjp at gSmaPil dAMot com
(take NO SPAM for email address)