洪水制御支援のためのメタ知識とその利用法に関する研究

Abstract

The aim of this study is to design meta-knowledge systems, which are important elements in hierarchical structuring method of knowledge-based systems for flood control supporting.First, the functions of a meta-knowledge system are clearly defined from the viewpoint of effective utilization of incomplete knowledge.Then an inference mechanism of the system is proposed based on Dempster & Shafer's probability theory.Reliability index of subordinate knowledge-based systems is also introduced as meta-knowledge.The inference system designed here is applied to the flood control supporting of Amagase dam reservoir for the verification of the theory.