| This paper discusses a man-machine game for Android mobile devices which is one kind of Tractor and its improvements of performance. Tractor is an incomplete information game, and it is also a2v2zero-sum game. The traditional Tractor uses the decision tree to decide the next step which will be made. This paper improves the decision tree of Tractor game, and then makes the next step more efficient and increases winning probability based on Bayesian formula and heuristic information.(1)This paper improves the decision tree used in traditional Tractor game, and uses probability theory and Bayesian formula for the process of man-machine game to infer the pros and cons of the brand strategy. One situation may have a variety of brand strategies, according to the order, color and type of the card, it determines (he decision tree and the pros and cons of each brand strategy based on probability theory and Bayesian formula, then determine the best brand strategy.(2) Tractor is an incomplete information game. Determining brand strategy is mainly based on the level ol understanding ol incomplete information. This paper increases the level of understanding ol incomplete information to increase winning probability by heuristic information. The information which need to be known mainly contain color, pair, score and quantity of card. According to the rules of the game and the habits of dealing card, it analyzes the existing incomplete information to get more information to increase the level of understanding of incomplete information.(3)Port the Tractor game which based on probability theory. Bayesian formula and heuristic information to mobile devices on the Android operating system, then analyze and compare with the experimental results of the traditional Tractor game. The experiment results show that the Tractor game based on improved decision tree, probability theory. Bayesian formula and heuristic information makes elllcient computation and higher chance of winning. |
PDF Full Text Request