The present study discusses the relationships between two independently developed models of games with incomplete information, hypergames (Bennett, 1977) and Bayesian games (Harsanyi, 1967). The authors first show...The present study discusses the relationships between two independently developed models of games with incomplete information, hypergames (Bennett, 1977) and Bayesian games (Harsanyi, 1967). The authors first show that any hypergame can naturally be reformulated in terms of Bayesian games in an unified way. The transformation procedure is called Bayesian representation of hypergame. The authors then prove that some equilibrium concepts defined for hypergames are in a sense equivalent to those for Bayesian games. Furthermore, the authors discuss carefully based on the proposed analysis how each model should be used according to the analyzer's purpose.展开更多
基金supported by Grant-in-Aid for Japan Society for the Promotion of Science(JSPS) Fellows, No.21-9482
文摘The present study discusses the relationships between two independently developed models of games with incomplete information, hypergames (Bennett, 1977) and Bayesian games (Harsanyi, 1967). The authors first show that any hypergame can naturally be reformulated in terms of Bayesian games in an unified way. The transformation procedure is called Bayesian representation of hypergame. The authors then prove that some equilibrium concepts defined for hypergames are in a sense equivalent to those for Bayesian games. Furthermore, the authors discuss carefully based on the proposed analysis how each model should be used according to the analyzer's purpose.