形式化智能世界的建模过程

各种网络化实体或智能实体的嵌入将物理世界转变为智能世界,增加了基于物理世界统一抽象模型的推理复杂性.基于Saitta和Zucker定义的知识重构与抽象模型,提出一种新的表示智能世界的方法.通过感知重构过程,根据构成实体之间的通信关系自动构建3个可区分的、相互关联的子模型,即物理世界、网络化世界和虚拟世界,3个子模型之间的关系构成了智能世界的一个集成化模型.着重介绍了感知重构过程的形式化表示和推理机制.实际搭建的智能世界设计诊断推理实验表明,与基于知识重构和抽象一般物理世界模型相比,基于提出的智能世界模型的推理过程可以将系统故障限制在一个子模型中,从而较大程度地缩小了诊断的搜索空间.

Formalizing Modeling Process of Intelligence World

Various networked objects or smart objects are embedded in physical world to transfer it into an intelligent world,which increases the reasoning complexity based on the unified abstraction model framework.This paper presen-ted a new method to represent such an intelligent world based on the knowledge reformulation and abstraction model proposed by Saitta and Zucker.We automatically constructed three distinguishable and interrelated sub-models,i.e.,the model of physical world,networked world and virtual world,according to the communicational ways of the compositive entities by a perception reformulation process.The relationships between the three sub-models are constructed to make them form an integrated model of the intelligent world.We focused on the formalized representation of the perception reformulation and reasoning mechanism in this paper.We built practical intelligent worlds and set up diagnosis reaso-ning experiments,showing that comparing to the knowledge reformulation and abstraction model, the reasoning process of the proposed modeling method can limit malfunctions to one of the sub-models to reduce the diagnosis space.