This article delves into a style of program design that starts up without pre-ordained strategy-type knowledge. It features a program -- in this case a chess-type game called Hexapawn -- that can train itself using simple, straightforward techniques.
- Posted by: Gervase Gallant
- Posted on: February 24 2004 00:13 EST
While the example is a game, the technique could be applicable to any data-sensing component that where state conditions can be recognized. And even if this doesn't ring a bell, you can always try to beat the computer!