基于智能体的混合知识自适应推理控制

A Mixed Knowledge Adaptive Reasoning and Controlling Based on Agent

本文将传统的混合知识表示法推广 ,建立了一个将一般知识、案例知识、模型知识、模型知识及神经网络知识有机集成的结构 .该种知识表示结构有助于知识的搜索、匹配和推理控制 ,解决了复杂问题的知识表示 .为适应推理方法的需要 ,将各种不同的推理方法有机融合与集成 ,提出一种基于智能体的自适应推理控制结构 ,该种自适应推理控制结构对于改进解决复杂问题的效果以及提高解决复杂问题的效率具有重要意义 .

In this paper ,a structure of mixed knowledge is given for describing general knowledge, case knowledge, model knowledge and neutral network knowledge,which is of importance not only in knowledge's search,marching,reasoning and control,but also extending traditional mixed knowledge describing approach.At the same time,in the need of adapting reasoning method,an adaptive reasoning and control architecture based on agent is presented.This architecture can greatly improve the effect to complicated problem as well as enhance its efficiency.