知识系统决策规则的动态更新策略及等价矩阵计算

专家系统是航天领域中人工智能技术应用的重要形式。学习能力是人工智能、机器学习的本质特征之一。用属性及其取值表述对象的知识系统通常需要从大量样本中获取知识并以决策规则的形式加以表达。当样本集合发生变化时,则需对决策规则集进行动态更新。本文分析了增加和移除学习样本两种情况下调整决策规则的方法,提出了知识系统决策规则动态更新的策略和等价矩阵计算方法,通过算例进行了仿真和验证。结果表明:决策规则的动态更新方法和等价矩阵计算可以解决知识系统的动态学习问题,相较决策规则的全局获取运算,减少了计算量。

Dynamic Updating Strategy and Equivalent Matrix Computation of Decision Rules in Knowledge System

Expert system is an important application of artificial intelligence technology in aerospace field. Learning ability is one of the essential characteristics of artificial intelligence and machine learning. Knowledge systems that express objects by attributes and their values usually acquire knowledge from a large number of samples and express it in the form of decision rules. When the sample set changes, the decision rule set needs to be updated dynamically. In this paper, the methods of adjusting decision rules in the case of adding and removing learning samples are analyzed, and the strategy of dynamic updating decision rules in knowledge system and the method of equivalent matrix calculation are put forward. The simulation and verification are carried out through an example. The dynamic updating method of decision rules and equivalent matrix calculation can solve the dynamic learning problem of knowledge system. Compared with the global acquisition operation of decision rules, it reduces the amount of calculation.