Inferring Rival´s Private Information in Game Circumstance

It is important for having the initiative in competition to infer rival´s private information accurately. The rival´s decision-making model is described by influence diagrams with game relation, and the criterion that rival will take equilibrium strategy in game and hope to gain global maximum expected utility in multi-steps decision is followed. After transforming influence diagrams into Bayesian network, the rival´s private information can be inferred conversely once the rival´s decision-making behaviors are observed. The experiment results show that proposed method is effective.