The Study Of Subclasses Computer Game And Evaluation Based On Multi-automata Compound

The meaning of Computer Game is play chess like people thinking. In recent years, as the "flies" of artificial intelligence, Computer Game has become one of the main research object of artificial intelligence and computer control workers increasingly.The Chinese Chess Modeling has been a hotspot of Computer Game research in recent decades.The multi-automata compound model is expounded in this theses. According to the theory of discrete event dynamical systems(DEDS), a automata model of Chinese Chess is established in this thesis. According to mlti-automata compound model and subclasses of Computer Game, multi-automata compound model of Chinese Chess is established and Chinese Chess rule is described.Game trees search is expounded detailed in this thesis.The traditional static evaluation used in Chinses Chess game on computer currently could not reflect the information of situation accurately, so a method of dynamic evaluation is proposed. Moreover, by introducing parameters in the traditional method of static evaluation and the test of competitive Chinese Chese with an important step in the game, extent and correlation of parameters were calculated.So that there is similar judgement ability whit Chinese Chese expert.Two method that calculate extent and correlation of parameters were designed in this thesis. One is based on current situation, another is based on all situation of the same layer in Game trees.The result of parameters of two method is calculated and tested in competitive Chinese Chese.The impression is all right in accuracy of evaluation get improved, which has a significant meaning to the Computer Intelligence development.