论文部分内容阅读
本文针对多Agent系统中Agent之间的盲目交互可能产生的效率低下问题。提出了一种基于概念树结构的多Agent合作求解模型。在这个模型中,各Agent基于自己的领域知识构造出概念树,通过Agent之间的合作,对概念树从根节点开始使用证据理论实现逐层聚焦,逐步缩小求解范围。为此,本文基于模态、逻辑和关系概念提出了一种面向可能解集的证据理论表示,并探讨了在多Agent环境下应用证据理论可能导致的若干问题。
In this paper, we aim at the inefficiencies that may arise from the blind interactions between agents in a multi-agent system. A multi-agent cooperative solution model based on concept tree is proposed. In this model, each Agent constructs a conceptual tree based on its own domain knowledge. Through the cooperation among Agents, the concept tree is gradually layer-focused using the evidence theory from the root node to gradually reduce the scope of the solution. Therefore, based on the concepts of modal, logic and relation, this paper presents a proof theory representation for possible solution sets and discusses some possible problems that may arise when using evidence theory in multi-agent environments.