Applying the Modified Delphi Method to Implement the Relational Operators in Management Ontology—The Case of Master Production Schedule Planning Ontology

• Main Authors: Wan-Li Chiou, 丘婉儷
• Other Authors: Tai-Yue,Wang
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/88742250006834886805

碩士 === 國立成功大學 === 資訊管理研究所 === 94 === Different ontologies are specifically applicable to different domains in order to help the users to share their domain knowledge as the knowledge management is blooming in these days. However, most researchers usually create classes for the concepts of the domain, define the properties of the classes, and set the relations among concepts indirectly via these properties when developing their ontology. They seldom define the relations directly among classes, instances, and ontologies. In this research, we apply Modified Delphi Method to define the relational operators among knowledge items directly. These relational operators, on one hand, will link the relationship among knowledge items. And the relational operators, on the other hand, provide the foundation of management ontology. We assume that there exists different ontology under different management levels. And different relational operators are needed for different ontologies in each management levels. Thus, the relational operators of high level ontology can be used in lower level ontology. The Master Production Schedule Planning is used as the production management domain to develop the relational operators of the ontology. These relational operators can improve the integration of ontology, help the domain experts to build more complete ontology, and make the users understand the domain knowledge more quickly. Specifically, we use Protégé 3.1.1 as an ontology-developing environment and Java is used as a program-developing tool to implement the relational operator plug-in for protégé. When the knowledge engineers develop Master Production Schedule Planning ontology, they can use the plug-in to implement relations among concepts or instances. And users can understand these relations thoroughly.