Syntactic and semantic measures to evaluate similarity of risk scenarios in manufacturing systems

Abstract

Manufacturing systems are subject to predictable and unpredictable occurrences of disturbances, which may alter pre-set organization, degrade performance and generate significant risks (direct and indirect consequences) that need to be addressed. In the literature, ontologies were used to capture past occurrences of disturbance and risk scenarios and capitalize reaction decisions in order to enable reuse of this experience in case of future occurrences of similar disturbances and risks. Unfortunately, existing works do not suggest similarity measures that take advantage of semantic and syntactic similarities between new and stored scenarios. This article fills in this gap by suggesting such similarity measures. A retrieval algorithm is developed to compare new disturbance and risk scenarios with stored ones, and to evaluate their similarity in terms of nature and severity of disturbances and risks. A case study shows competitive and promising results.