I've used it as the basis for a rules based heads up no limit poker agent (game logic is also implemented in Prolog, GUI is Java/JS...needs a major rewrite though). Constraint satisfaction programming is pretty helpful.
It's mostly a rather simple agent but I'm currently working on a DSL to codify strategies more easily. I'm fully aware a rules based agent will never be strong but I'm mostly interested in figuring out how well experts can codify their strategies and if you can make them comparable (also looking into inductive reasoning and datamining strategy rules from sufficient hand samples of winning online players)
Mostly a toy project these days :)
[Edit: I'm actually going to rewrite it because it used Sicstus and I want to use an open implementation so I'll switch to SWI + while I'm at it I'll drop the Java stuff and go full JS/HTML. Pengines looks pretty good: http://www.swi-prolog.org/pldoc/package/pengines.html]
It's mostly a rather simple agent but I'm currently working on a DSL to codify strategies more easily. I'm fully aware a rules based agent will never be strong but I'm mostly interested in figuring out how well experts can codify their strategies and if you can make them comparable (also looking into inductive reasoning and datamining strategy rules from sufficient hand samples of winning online players)
Mostly a toy project these days :)
[Edit: I'm actually going to rewrite it because it used Sicstus and I want to use an open implementation so I'll switch to SWI + while I'm at it I'll drop the Java stuff and go full JS/HTML. Pengines looks pretty good: http://www.swi-prolog.org/pldoc/package/pengines.html]