Author: Frank Schneider
Date: 23:47:47 10/28/98
In the current issue of the ICCA journal there was an interesting article about programs able to play games that they have to learn by themselfes, given only the rules of a game. For some time there is also a discussion about what might be the next AI-game after chess (for example Go or Shogi). This year there is an interesting competition at our university: the task is to write a program that controls two players of a 4-player rugball (simplified rugby/american football) -team. Program 1 controls the quarterback and the receiver Program 2 controls two other attackers Program 3 controls 2 defenders Program 4 controls 2 defenders Every program computes the moves for it's players and then a 'referee' solves collisions and makes the moves (if possible). Although it is a 'two player zero-sum game' the two main differences to Chess, Shogi and Go are, that two programs have to cooperate and that the moves of the other players (even the partner) are unknown. It is possible to do a little tree-search, but brute-force is too slow since there are more than 1.000.000.000 possible combinations of moves of the 8 players. You can find more information (rules, referee-program, ...) in german here: http://www-i1.informatik.rwth-aachen.de/TdI98/ If somebody is interested in details I can translate the most important facts into english. I'm going to participate in that competition, wrote a program and experimented a little. Now I think that this game is really interesting, because on the one hand it is more similar to real-world-problems than chess but on the other hand it is not too complex. What do others think? Frank
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.