Journal of the Korean Institute of Gas (한국가스학회지)
- Volume 2 Issue 1
- /
- Pages.53-58
- /
- 1998
- /
- 1226-8402(pISSN)
- /
- 2713-6922(eISSN)
A Study on the Automatic Synthesis of Signed Directed Graph Using Knowledge-based Approach and Loop Verification
지식 기반 접근법과 Loop 검증을 이용한 부호운향그래프 자동합성에 관한 연구
- Lee Sung-gun (Dept. of Chem. Eng., Pusan National University) ;
- An Dae-Myung (Dept. of Chem. Eng., Pusan National University) ;
- Hwang Kyu Suk (Dept. of Chem. Eng., Pusan National University)
- Published : 1998.03.01
Abstract
By knowledge-based approach, the SDG(Signed Directed Graph) is automatically synthesized, which is commonly used to represent the causal effects between process variables. Automatic synthesis of SDG is progressed by two steps : (1)inference step uses knowledge base and (2)verification step uses Loop-Verifier. First, Topology and Knowledge Base are constructed by using the information on equipment. And then, Primary-SDG is synthesized by Character Pattern Matching between Variable-Relation-Representation generated by using Topology and Variable-Tendency-Data contained in Knowledge Base. Finally, a modified SDG is made after the Primary-SDG is verified by Loop-Verifier.
화학공정 변수간의 관계를 표현하는 방법으로, 지식기반 접근법을 이용하여 부호유향그래프(signed directed graph, SDG)를 자동합성하였다. SDG의 자동합성은 지식베이스를 이용한 추론과정 및 Loop 검증의 두 단계를 거쳐 수행된다. 먼저, 공정내 장치를 중심으로 장치간의 결합관계를 Topology로 표현하고, Topology 정보를 이용한 공정 변수관계 표현 및 지식베이스의 공정경향 데이타를 문자 패턴 매칭하여 Primary-SDG를 자동으로 합성한다. 생성된 Primary-SDG를 Loop 검증기의 추론을 통하여 검증, 수정하여 SDG를 자동합성하였다.