多Agent理论的对策论方法

人工智能研究的一个重要目标是设计出能代表其拥有者或用户的、有一定自主性的决策体Agent。为此通常是赋予每个Agent一种自身利益(偏好),Agent在行动中寻求自身的利益最大化。当考虑MAS时,Agent通常被赋予利益、理性决策能力及推断其它Agent的利益等的能力。

The Game Theoretical Methrod in Multi-Agent Theory

Multi-Agent Theory was presented as a computational model of distributed intelligence.Recent research was focused on the specification for describing agent,knowledge representation,and formalizing some concepts of cognitive science and phdosophy. There are logic method and game theoretical method in the research of multi-agent theory.Using non-classical logic to model the cognitive attribute is main research method,there are lots of paper to discuss this topic. This paper focuses on the game theoretical method.