[LRUG] Battleships last night

[James Cumming]

This is and interesting bit of research too. Would have been nice to have the time to build a player too.


http://thevirtuosi.blogspot.com/2011/10/linear-theory-of-battleship.html

 
James Cumming, CFA
+44 7799 554468


________________________________
From: Chris Lowis <chris.lowis at gmail.com>
To: London Ruby Users Group <chat at lists.lrug.org>
Sent: Tuesday, 11 October 2011, 9:57
Subject: Re: [LRUG] Battleships last night

I never found the time to implement a player, sadly, but I did really
enjoying reading this paper:

http://www.cores2.com/files/FinalResearchPaper.pdf

In it the author describes a learning algorithm to optimise ship
placement by learning from its opponent over the course of many games.
I suspect that wouldn't have worked very well in a best of three match
though.

I also enjoyed the monte-carlo simulation approach to generating the
ship placement - a little similar to Gunter.

For the attacking strategy the paper describes a state-machine which
again using a genetic algorithm to learn the best time to switch
between states to effectively hunt and sink the enemy ships.

It would be fun if the DRB server was left running and we could submit
new or improved bots over time :)

Cheers,

Chris
_______________________________________________
Chat mailing list
Chat at lists.lrug.org
http://lists.lrug.org/listinfo.cgi/chat-lrug.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lrug.org/pipermail/chat-lrug.org/attachments/20111011/74839f54/attachment-0003.html>