Moriyoshi Kushida

MODAL LOGIC TO EVALUATE A KNOWLEDGE-BASED BRIDGE RATING SYSTEM

Knowledge acquisition has been recognized as a major bottleneck in the development of effective knowledge-based systems in recent years. Therefore, many researchers are investigating knowledge-acquisition methodologies to overcome this problem. In this article, a flexible method for expressing empirical knowledge related to bridge rating acquired from questionnaires filled out by bridge rating experts is proposed by relating the knowledge to a membership function based on fuzzy-set theory. In this method, a membership function is regarded as a possibility distribution based on a theory of possibility, and an apparent possibility distribution expressed by an extended II function is transformed into possibility measure distribution and necessity measure distribution utilizing a concept of modal logic. Furthermore, the proposed method has been applied to a questionnaire survey conducted to evaluate the serviceability of an existing bridge. The validity of the method has been verified through comparison between the subjective rating results acquired from application of the method and objective rating results obtained from nondestructive tests conducted for evaluating the serviceability of the same bridge.

A STUDY OF SYSTEM RELIABILITY IMPROVEMENT FOR NEURO-FUZZY BRIDGE RATING EXPERT SYSTEM

本研究は著者等が開発したコンクリート橋診断ニューロ・ファジィエキスパートシステムのさらなる実用化を目指して, エキスパートシステム技術自体が直面している種々の問題点を検討し, システムの信頼性を高めることを目的としている. まず, 従来のアンケート調査による教師データ獲得手法を見直し, 教師データに内在する不要なあいまいさを排除する手法を提案する. また, 学習機能を保有するシステムの最大の欠陥である推論過程のブラックボックス化を防止する支援機能を新たに開発し, システム自体の推論機構の信頼度を向上させる. そして, 現システムの知識ベースを提案する手法により洗練した上で, 実橋の耐用性診断を行い, その診断結果および目視調査結果を比較する事で提案手法の妥当性を検証する.