規則式方法應用於彈性製造系統的派工問題研究

Abstract

本論文將探討應用決策樹、決策表、IPO模式(Input-Process- Output Model)及彩色斐氏圖(Color Petri-Net)等一系列方法與程序來擷取、分析專家的知識、經驗，並使用智慧型決策支援系統(IDSS)提供的人機介面，轉換、模擬與測試知識規則及決策模型，將知識轉換過程的分析方法與程序予以明確化，讓使用者能因此按部就班，逐步分析、構建知識規則。 而本論文所使用的IDSS工具是應用網際網路技術、決策支援概念及Java Expert System Shell（JESS）推論引擎所發展的規則式推論編輯工具，其提供圖形化知識規則編輯器及決策元件封裝嵌入功能，使得建置、編輯知識規則與整合外部系統或裝置更加彈性。並以彈性製造系統(Flexible Manufacture System, FMS)作為測試、驗證載具，以實際驗證本研究所提出之分析方法與程序的可行性與實用性。

This thesis was described a method and procedure that apply decision tree, decision table, IPO model and Color Petri Net to analyze the domain experts knowledge and to transfer the knowledge rule and decision model via Intelligent Decision Support System (IDSS). The objective of developing the rule-based inference method is to provide an easy follow knowledge rule construction approach. This IDSS tool was constructed based on the web technology, decision support concept, and Java Expert System Shell (JESS) inference engine. It is to provide the visual editing interface and the decision object’s packaging & embedding function, so as to have the integration of knowledge rule inference ability and other systems are more flexible. The developed method was test and verified experimental results were given to illustrate the usefulness of the developed method via the FMS test platform.