A hypothetical reasoning for knowledge bases with preferentially ordered hypotheses
Yasuyuki Watanabe, Ikuo Tahara · Systems and Computers in Japan · 1997
When we have to reason with an incomplete knowledge base that consists of facts and hypotheses, we first generate a set of facts and acceptable hypotheses, called a scenario, and then reason deductively from this scenario. In this case, which hypotheses to choose is an important problem. This paper considers a knowledge base with preferentially ordered hypotheses and proposes a method of generating scenarios based on this preferential partial order. The notion of conflict sets is introduced to represent inconsistency among hypotheses, and the scenario of the knowledge base is defined by these conflict sets and the preference relation. The semantics of the scenario is discussed, and the consistency of the scenario is shown. Finally, an algorithm for determining whether or not a ground sentence is deducible from the scenario is presented. © 1997 Scripta Technica, Inc. Syst. Comp Jpn, 28(5): 52–59, 1997