首页 | 本学科首页   官方微博 | 高级检索  
     

铁路计算机联锁软件制作专家系统的研究
引用本文:张杰. 铁路计算机联锁软件制作专家系统的研究[J]. 铁道标准设计通讯, 2019, 0(11): 165-171
作者姓名:张杰
作者单位:中国铁道科学研究院集团有限公司通信信号研究所
摘    要:铁路计算机联锁软件制作的关键数据,由手工输入,智能化水平不高。提出专家系统的思想,借助专家系统工具CLIPS,实现联锁站场特征的自动识别,数据的自动生成。将蓝图信息规范化,统一为Excel文件,解决了联锁知识的概念化及形式化问题。采用面向对象的知识表示方法,借鉴ID3的分类算法,建立知识库。分类算法提高了规则的合法化及程序执行的效率。在站场特征识别问题上,先讨论了基于二叉树的实现方法,又叙述了直接利用CLIPS消息函数的方法,比较可得,消息函数的方法简洁,具有优越性。总结3类VC++和CLIPS交互技术,说明CLIPS嵌入到VC++是可行的。经检验,该专家系统可以初步解决计算机联锁数据标准化、自动化的问题。

关 键 词:专家系统  铁路计算机联锁  CLIPS  知识表示  知识库  辅助设计

Research on Expert System for Railway Computer Interlocking Software Production
Affiliation:,Signal and Communication Research Institute, China Academy of Railway Sciences Corporation Limited
Abstract:The key data of the production of computer interlocking software is manually input with low intelligence. In this paper, the idea of expert system is put forward. With the help of CLIPS, the automatic recognition of interlocking station features and the automatic generation of data are realized. The blueprint information is standardized and unified into Excel files, which solves the problem of conceptualization and formalization of interlocking knowledge. Using object-oriented knowledge representation method and ID3 classification algorithm, a knowledge base is established. The classification algorithm improves the legalization of rules and the efficiency of program execution. On station feature recognition, the method based on binary tree is discussed firstly, and the method of directly using CLIPS message function is described. The analysis and comparison show that the method of message function is simple and superior. The three types of VC++ and CLIPS interaction technologies are summarized and CLIPS embedded into VC++ is proved applicable. After testing, the expert system can initially solve the problem of standardization and automation of computer interlock data.
Keywords:expert system  railway computer interlocking  CLIPS  knowledge representation  knowledge base  auxiliary design
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号