A Knowledge-Based Helpdesk System For TFT-LCD Factory CIM

• Main Authors: TZU-CHE HUNG, 洪子哲
• Other Authors: Sheng-Tun Li
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/87261159671089916764

碩士 === 國立成功大學 === 工業與資訊管理學系專班 === 95 === TFT-LCD is one of the most important industries in Taiwan. Up to 2006 year's end, Taiwan is the most major large-size TFT-LCD production area in the whole world. CIM (Computer Integrated Manufacturing) System using the most advanced computer and network technology to manage and control the whole production activities in a TFT-LCD factory. The Purpose of CIM system is to integrate the functions of material control and information flows to improve the production efficiency, lower the production cost, and to optimize the product quality. Since the helpdesk is the only service contact window of the CIM system for 7 days 24 hours, and the agent member in helpdesk will not increase as the CIM system becomes more and more complex. How to help the helpdesk agent to find out system problem quickly and solve it correctly becomes a very important and critical issue. In this article we present an approach to develop a knowledge-based system for TFT-LCD Factory CIM Helpdesk, using Knowledge Acquisition (KA) technologies such as VPA (Verbal Protocol Analysis) and PBG (Problem Behavior Graph) processes to collect and analysis knowledge elements from experts of CIM subsystems. Then we developed different inference methods including Forward-Chaining and Backward-Chaining to adapt the different functional requirements for helpdesk agents to solve their problems. Through these processes, the knowledge classes and rules were successfully modeled and the knowledge-based helpdesk system was successfully verified by real test cases in the TFT-LCD factory CIM system.