Hey, so let me first explain my game. I've built a rudimentary checkers game, and I wanted to include a computer player. The most I could think of is a rudimentary system of immediately checking the closest positions, and calculation estimated moves. In my initial ALPHA testing, the system is very basic and unable to properly play against a user.
Does anyone have any advice on how I could essentially implement AI learning? Something like ANNs. Basically, the AI will use datastores to let the data be stored globally. Starting with only the fundamental basics (where you are/are not allowed to go), and it essentially takes the games it plays with our testers (and later average players) and uses that data to improve its 'thinking'.
If anybody knows where to start, has any sudocode I could use, or something along those lines, because I bearly know where to start, because this is a really complex and complicated process.
Thanks for any help! |