- 學位論文
以本體知識模型為基礎的產品配方組合--以羽絨業為例
A Production Recipe Intelligence System Based On Ontology--for Down Industry
摘要
羽絨產品是多元素組合而成的混合物，主要的成份是羽絨跟羽毛，其品質與價值取決於羽絨的含量比例。然而，羽絨原料是天然的保暖素材，其貨源品質常受到氣候、產地或是突發性的傳染病如禽流感等不可抗力的因素影響，因此，羽絨產品的調合比例，需因時制宜因應當時廠內存貨配置、品質狀況及最佳成本考量等因素設計適當的配方。現行羽絨配方作業需仰賴人工進行調配及計算，其產業知識缺乏邏輯性的管理，導致產能規劃僵直、存貨配置不當及品質控管不佳的現象。本研究針對上述問題，提出以知識本體(Ontology)為核心的專家系統，利用知識工程方法，將此議題建立為一般化的知識模型，並將羽絨產品的成份及影響的變因納入分析的要項，透過關鍵性問題分析(CQs)，建立具邏輯定義的知識單元，並利用知識本體明確表達領域知識以及概念間關係，建置出配方表的配方組合及專家經驗的模型，進行知識推導，以求得羽絨配方產品的建議，做為廠內的分享平台，以逹到產業智識傳承的效用。
並列摘要
Product of down is a kind of mixtures consisting of down and feather in different grades. The proportion of down and feather is very critical. Also, stocks on site, quality of stocks, and the cost of production will be the most important factors for designing the recipe of production. Now, the production recipes all depend on the practical experiences of management staffs in the factory. Based on the collected data of production recipes, it reveals the issues of the production recipe design without a common sharing knowledge base. Lacking of the knowledge support during the design stage of recipe, it may directly lead poor design and cause further issues in production, such as inflexible plan of production, over/insufficient stocks, and poor quality management of the product. Enterprise need take care of these issues and these issues will generate extra costs, time, and some other enterprise resources. These issues are the risks of management and business. This research is to provide a solution for the issue of recipe design of down industry. By using the methodology of knowledge engineering, a knowledge base will be built based on the core of ontology. The research will illustrate how to use expert system to facilitate the design of production recipe. With providing requirements of product quality, the expert system will be used to find the optimal proportion of down and feather for the product. Management staffs can use expert system as a platform to increase the efficiency of communication and knowledge transfer.