专家系统的实时性分析

Real Time Analysis on Expert Systems

在一些复杂控制系统中 ,专家系统作为一个决策控制器 ,满足整个系统的性能需要 ,因此研究专家系统的推理时间是必要的。专家系统的推理时间与其推理模式、知识库结构等因素有关。针对知识库结构 ,通过专家系统的推理时间研究了实时性问题 ,利用时间齐次马尔可夫链为专家系统知识库进行建模 ,并给出了相应的时间估计模型及其排列准则。

In some complex control systems, expert systems, as a decisive controller, consist in the whole systems, so much more studies on inferring time are necessary for them. Inferring time in expert systems relates to its inference mode, knowledge-base architecture, computer performance, and provided languages. This paper takes the knowledge-base architecture as an example to study the corresponding problem of the alignment. At last, Temporally homogeneous Markov chain is used to modeling for knowledge-base. Time assessing mathematics model is given, which can provide proof for the complex control system.