Artificial Intelligence Algorithm for the Game of NoGo based on Value Evaluation

The game of NoGo is one of the chesses for machine game, and is a variation of go-chess. NoGo and its related rules are firstly introduced with demonstration on recent researches on artificial intelligence for the game. Then a model function of value evaluation for arbitrary chessboard is proposed through analysis of characteristics of NoGo, and one or more candidate points with the highest evaluation value shall be given. Hence, an optimal solution can be given if the candidate point is sole, otherwise, the optimal solution can be given via scattering rule or the algorithm of Monte Carlo tree search (MCTS) according to different cases. Finally, the game result between the proposed algorithm by this paper and the famous open source software (OSS)-OASE-NOGO is obtained, and the proposed algorithm gives 95% of win rate, which, comparing with previous researches, proves to be feasible and efficient.