• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2012, Vol. 34 ›› Issue (2): 139-145.

• 论文 • Previous Articles     Next Articles

Research on the Least Common Subsumer in Description Logic εLN

ZHANG Wei1,2,CAO Fasheng2,YU Quan1,2,WANG Ju3   

  1.  (1.Department of Mathematics,Qiannan Normal College for Nationalities,Duyun 558000;2.Center for Logic,Language and Cognition,Bijie University,Bijie 551700;3.School of Computer Science and Information Engineering,Guangxi Normal University,Guilin 541004,China)
  • Received:2010-05-10 Revised:2010-08-28 Online:2012-02-25 Published:2012-02-25

Abstract:

Nonstandard inference is a new inference service in description logic,which was proposed at the turn of the century. It mainly includes the most specific concept,the least common subsumer,matching problem,concept rewriting and so on. Nonstandard inference aims to deal with the construction and management of the knowledge base and some other problems which researchers have to face in applications. Nonstandard inference can support the bottomup construction method,as the description logic is widely applied in other areas,the knowledge base construction and management is more and more critical and inevitable,but the standard inference almost cannot provide any help. So the nonstandard inference becomes a main issue which researchers are focusing on in recent years. In this paper the fundamentality of nonstandard inference in description logic, especially the current research progress and existing problem of the LCS(least common subsumer) inference in description logic are analyzed. Aiming to the insufficiency of the LCS inference which cannot handle existential restriction and number restriction at the same time, the LCS inference in description logic εLN is studied. Firstly the εLNdescription tree and its homomorphism are defined, a subsumption algorithm is presented by computing the homomorphism between two description trees. Then the product of description trees is defined and the LCS inference algorithm is presented by computing the product of description trees. The complexity of the LCS inference algorithm in εLN is polynomial.

Key words: description logic;description tree;homomorphism;least common subsumer