摘要
目的讨论基于模糊产生式规则的故障诊断专家系统与神经网络相结合的问题,把推理网络同神经网络联系起来,使它能转换成神经网络.方法在转换中,把模糊产生式规则前提的置信度归结到神经网络的输入信息学习样本,把规则的置信度归结到神经网络的输出信息学习样本,并给出实例的具体实现过程.结果解决了模糊规则专家系统向神经网络转换问题,实现了基于神经网络的模糊快速推理诊断和知识自动获取.结论经实例验证,该方法可靠有效.利用神经网络的并行处理和自学习能力,能避免传统模糊推理的冲突,低效率和知识获取的瓶颈问题.
Aim\ The integration of the fuzzy produced rules based fault diagnosis expert system and neural networks, and conversion from the rules based reasoning networks to neural networks were discussed Methods In the converting process, the certainty factors of the conditions and the rules were respectively merged in the inputting and the outputt ing information of the learning samples of the neural network The details of the actual example were described Results Conversion from the rules based reasoning networks to neural networks was accomplished The fuzzy quick reasoning diagnosis and the automatic knowledge acquisition based on neural networks were realized Conclusion By the test of the actual example, it is shown that the method is effective and reliable, and solves the conflict in the fuzzy rules based reasoning Neural networks are combined with fuzzy rules based expert system effectively
出处
《北京理工大学学报》
EI
CAS
CSCD
1998年第1期81-86,共6页
Transactions of Beijing Institute of Technology
关键词
神经网络
模糊产生式规则
故障诊断
专家系统
neural network
fuzzy produced rule
certainty factor
fault diagnosis
reasoning conflict