| The research on intelligent agent has been one of the focus of Artificial Intelligence,with the development of computer network, communication technology and so on.Especially,the idea of creating an intelligent agent,which can act and reason autonomously in the real world,has been researched by many experts and scholars.One of the central problems is that endowing agents with the capability of autonomously reasoning about actions and achieving the goals based on those formal descriptions of actions and environment properties.Based on the analysis of related literatures about fluent calculus and FLUX,the paper extends the GoFlux to ConFlux.Simultaneously,this paper combines the representation of knowledge of other agents,imports the communication actions and improves on the knowledge update axiom. Finally,an intelligent scheduling plan is applied to validate the above those.The main research is as follows:(1) Give a systematical research of the formal theory of Fluent Calculus and FLUX,the implementation language of fluent calculus. Account for dynamics and indetermination of MAS,fluent calculus and FLUX are adopted to formally describe and implement the environment, actions and goal-oriented reasoning of MAS. (2) Aim at GoFlux can't process the concurrence,concurrence with different priorities,interrupt and so on,this paper extend GoFlux to ConFlux based on the GoFlux and absorption of ConGolog semantics in order to implement the concurrent and intersect multi-agent program.(3) Based on the fluent calculus,the representation of knowledge of other agents is combined,communication actions are imported to expand the intersection feature of ConFlux,and how communication actions influence on agents each other is discussed.A request/server cooperation model is put forward based on the above theory and given the formalized semantics.(4)Finally,an intelligent scheduling plan is applied to validate the above theory,and it has some practical value. |
PDF Full Text Request