人工知能
Online ISSN : 2435-8614
Print ISSN : 2188-2266
人工知能学会誌(1986~2013, Print ISSN:0912-8085)
並列サーカムスクリプションのパラメーター消去手法
岩沼 宏治原尾 政輝
著者情報
解説誌・一般情報誌 フリー

1990 年 5 巻 3 号 p. 324-332

詳細
抄録

Predicates treated as variables, i. e. parameters, in parallel circumscription make it difficult to compute parallel circumscription. Therefore it is very important to develop powerful methods for eliminating parameters. In this paper, we present some parameter elimination methods in parallel circumscription. At first, we show two fundamental theorems for eliminating parameters. Next, we give two parameter elimlnation methods, which are directly derivable from those theorems. One of them is used for eliminating a parameter which occurs only positively or only negatively in a condition sentence of parallel circumscription. Another is a method for eliminating a parameter by replacing it with a formula expressing the minimal or maximal extension of the parameter. These elimination methods are systematic, and can deal with parameters such that they are quantified with both existential and universal quantifiers in the conditlon sentence. The class of parallel circumscription which can be translated into predicate circumscription by these methods is larger than the one by well known Lifschitz's method.

著者関連情報
© 1990 人工知能学会
前の記事 次の記事
feedback
Top