A Military Chess Game Tree Algorithm Based on Refresh Probability Table

Computer game is divided into complete information game and incomplete information game. As a kind of incomplete information game, military chess game has many shortcomings in game-tree algorithm. In view of the fact that it is difficult to develop the game-tree because the specific information of the opponent's chess pieces can not be obtained accurately in military chess game, this paper proposes an algorithm to expand the game tree by refreshing the probability table formed by the rules of military chess, the result of the opponent's chess game and the result of the collision of the two opponents' chess pieces. The algorithm of the refresh probability table is used to provide the score support of the expanded game tree for the military chess game, thereby ensuring the expansion of the game tree. The initial values and weights in the refresh probability table algorithm are optimized by machine learning to improve the speed and accuracy of the refresh probability table, and the game tree is optimized by the minimax algorithm and the Alpha-beta pruning algorithm. Experiments show that the refresh probability table algorithm based on the game tree improves the winning rate in the game and has achieved good results.