pax

Past research in artificial intelligence was primarily engaged in deterministic games which provide perfect information, for instance to mention chess or othello among many others. However, the Çrealë world is characterized by a state of imperfect information and absence of determinism. For this reason, the present work shall investigate a game which is quite different from those mentioned above. First of all, the nescience about the ohter takersë cards is a situation in which imperfect information prevails. The indeterministic character of the game most of all accrues due to the random mixing of the cards which gives a stochastic feature to the nature of Poker. Those attributes turn it into a game which is close to reality. This field of research in artificial intelligence represents a huge challenge. The occuring difficulties during a game are intuitively capable by a human player, but, however, constitutes in a highly complex problem for automated computation. The main goal of this thesis is to develop an artificial intelligence being able to contest in a Poker game with more than two players. Its core shall be a strategy enabling the artificial agent to maximize its profit in any situation of the game. This optimization problem will be approched by game theoretical decision making.

Strategiefindung für ein Spiel mit unvollständiger Inform...

Strategiefindung für ein Spiel mit unvollständiger Information (Texas Hold'em Poker)

actions