Hi! We're just working on a nice Tetris clone, and we'd like to include a multiplayer mode, where one can play against the other one at the same time. Furthermore, there should be a computer opponent, if just one player is available and we also could use it for a demo mode. What do you think, which AI-technique would be the best for such a task? FSM, Genetic Algorithms, Fuzzy logic?!? Tell me your thoughts about it and if you got an URL about this stuff, please tell me! Thanks in advance, Have a nice day. Paralax Edited by - Paralax on 8/19/00 8:38:52 AM

Paralax-

It all depends on what the computer has to do right? I don''t think you''re going to need genetic algorithms for a mere Tetris clone. And since the idea of the game is pretty straightforward (unless you really changed the rules) there shouldn''t be too many opportunities to outsmart your opponent, so fuzzy logic isn''t needed either. An FSM should be enough, just program in the various states (positions) of the blocks (left, right, etc.) and let the AI decide where best to plunk it down. If you really want to get complicated, have an array store the positions and states of all the blocks already stacked so the AI can "search" for the best place to put the current peice that would help it in the long run.
Again, since I don''t know the rules of your game, this is the best suggestion i can make. Hope this helps.

******************************
"I do not fear computers, I fear the lack of them"

- Isaac Asimov

Drew Sikora
Napali Networks, Inc.