多智能体实例检索控制模型的研究

Study On the Multi-Agent Case Retrieval Control Model

引言基于实例的推理(Case一Based Reason,ng,eBR)是基于利用以往求解类似问题的以实例形式表示的经验知识进行推理,从而获得当前问题求解结果的一种推理模式Llj,它克服了传统的基于规则推理系统的知识难于获取和推理的脆弱性等缺陷,已广泛应用于各种问题求解的领域。一般来说,CBR的推理过程包括问题特征的抽取、相应实例检索、实例的改写与调整、方案的评价和新实例的存贮等基本步骤〔,一2〕。

On the basis of analysis of traditional case retrieval control strategies, the paper proposed an autonomy intelligent agent which consists of the case with the knowledge, goal and ability, it can actively run in the case retrieval process. A multi-agent case retrieval control model which consists of the agent and its control process was proposed in the paper.