以物件為基礎之即時診斷型專家系統

Abstract

在一個動態的系統中，由於組成元件彼此之間的交互作用，常會導致系統發生不可預期的故障。而知識庫系統的出現，正適合做為故障診斷的最佳工具。但傳統的專家系統或者物件導向的方法，已不適用於處理即時與同步系統的診斷問題，在本研究中，我們提出一個以物件為基礎的即時診所型專家系統來解決這些問題。這個診斷方法論整合了物件庫及裴氏圖的概念來表現系統的靜態與動態的關係，使得其推論機制在診斷物件間能更有效的運作。最後，本研究將根據上述方法實作此系統。

In a dynamic system, components which are interactive with other components will be faulty in an unexpected way and lead to breakdown of the system. Knowledge-based systems have emerged as the most appropriate approach for fault diagnosis. However, it has become apparent that a conventional expert system or an object-oriented knowledge-based approach is not going to be adaptive in real-time and concurrent systems. In this research, we present an alternative, an object-based real-time diagnostic expert system, to solve these problems. This diagnostic methodology integrates object-based and Petri Net concepts to represent both the static and dynamic relationships. The diagnostic reasoning effectively and efficiently alternates between diagnostic objects for complete diagnosis. All these ideas will be implemented in this research.