摘要
混凝土泵送技术的复杂性和模糊性一直都是靠领域专家的经验来解决的。神经网络在专家系统中的应用是在人工智能方面的重大发展,它克服了传统专家系统在知识获取和表达上的薄弱环节,由于模拟了大脑的思维性,推理机制比产生式推理优越许多。基于神经网络的专家系统只需要领域专家解决问题的实例或范例来训练神经网络,使在同样输入的情况下神经网络能够得到与专家给出的解答尽可能相同的输出,而且解答的精度可用它与专家解答之间的均方差的准则来度量。文中介绍了BP神经网络在专家系统中的知识获取和表示以及推理机制问题。
Concrete pump technology depends on experts in the special field to resolve complexity and faintness at all times in expert system.Network expert system is a great development at artificial intelligence and expert system.It conquers many difficulties in obtaining and expressing knowledge.Owing to simulating cerebra's thinking,its reasoning power is better than tradition's.Expert system based on neural network only need examples that is about the expert field to train the neural network.This expert system can get same output alike the genuine expert if there are same inputs.The precision of the expert system's outputs is resemble as the real field expert.This article is about network expert system's obtaining and expressing knowledge and reasoning power.
出处
《微机发展》
2004年第7期115-117,共3页
Microcomputer Development
基金
国家"十五"科技攻关项目资助(2001BA307B01-02-01)
关键词
神经网络
泵送专家系统
人工智能
BP算法
neural network
pumpcrete technique
expert system
knowledge base
BP arithmetic
