| Summary: | 碩士 === 中華大學 === 資訊管理學系碩士班 === 98 === Quality abnormal (QA) issue is a usual and general phenomenon while manufacturing process implementation. To assist and accelerate engineers deal with such QA issue, engineers indeed require more information. Additionally, some previous experience for QA issues processing would be very usually however such experience usually bundle of sophisticated engineers as potential knowledge. Therefore, this research proposed an Intelligent and Interactive abnormal diagnosis mechanism (IADRM) to assist engineers easily to find out the causal results for QA issue. IADRM enable even inexperience engineer process QA issues as sophisticated engineer as possible and improve effective for QA issues processing.
There are three goals of IADRM. The first one, IADRM can intelligence figure symptoms of QA issue out and provide some reference case(s) for QA treatment suggestion, Then, such knowledge would be cumulated and store in business case base for further reference. Such retain case(s) back process enable business knowledge management implementation. Finally, a feedback diagnosis process enable IADRM reinforce self and improves the diagnosis of quality abnormal of products in enterprise.
IADRM is implemented by Case-based Reasoning (CBR) approach. Prior to the IADRM implementation, a lot of QA issues are collected and analyzed by domain experts to formulate business abnormal case base. Additionally, three similarity evaluate methodologies are adopted to measure the similarity between case(s) and Target QA issue, including: totally identical, partial similarity and frequently useage.
Finally, IADRM is implemented and validated its feasibility and effective via filed study in H company in Taiwan. Two experiments are implemented. The first one testified three abnormal features (three QA issues) for IADRM feasibility validation. As the experiment result show, IADRM is appropriate for all situations. Additionally, as the second one experimental result demonstrated, IADRM can simulate human thought to recommend the good diagnosis and are better then human randomly input criterion.
Finally, as the interview analyze result show, most users satisfy with IADRM that enable them reduce diagnosis time for abnormal process. The IADRM exactly provides knowledge management to the novices and employees through interactive case-based reasoning. IADRM also reduces the diagnosis time on process abnormal cases. The research can adds the automatic process to evaluation mechanism and think about data mining to create rule-based to enhance IADRM.
|
|---|
