静态排列知识库专家系统时间估计模型的研究

Time Assessing Model in Expert Systems of Natural Sequence Knowledge Base

在一些复杂系统中，专家系统是作为决策环节，服从于整个系统的运行，因而研究专家系统的推理时间就有其必要性．专家系统的推理时间与其推理模式、知识库结构、计算机性能以及其支持语言等若干因素有关．针对知识库结构，研究了自然排列知识库专家系统的推理时间问题，利用时间齐次马尔柯夫链为专家系统的知识库进行建模，并给出了相应的时间估计数学模型及计算量的确定问题。

In some comprehensive systems, expert systems, as a decisive loop, consist in the whole systems, so much more studies on inference time are necessary for them. Inference time in expert systems relates to its inference mode, knowledge base architecture, computer performance, and provided language. This paper takes the knowledge base architecture as an example to study the corresponding problem of inference time in expert systems of natural sequence knowledge base. At last, temporally homogeneous Markov chain is used to model for knowledge base. The mathematics model, the computation method and the determination of the computing quantity are also given, which can provide proofs for the comprehensive system controlling and real time simulation.