根據延伸式乏晰派翠網路和時間派翠網路以作知識表示和推理的新方法

Abstract

在本論文中，我們提出一個延伸式乏晰派翠網路模式以塑模規則式系統中 之乏晰IF-THEN規則和乏晰IF-THEN-ELSE規則。我們也根據此延伸式乏晰 派翠網路模式提出了一個很有效率的模糊推理演算法以自動的作模糊推理 ，此演算法能使電腦以更有彈性且更有智慧的方式作推理。根據此模糊推 理演算法，我們在個人電腦上實作了一個模糊推理系統以自動的作模糊推 理。另外，我們也利用時間派翠網路來描述狀態和事件之間的關係，並根 據此時間派翠網路以作時間知識的表示和推理。我們亦根據時間派翠網路 提出一個時間推理演算法以作時間知識一致性的檢查。根據此時間推理演 算法，我們亦在個人電腦上實作了一個時間推理系統以檢查時間知識的一 致性。 In this thesis, we present an extended fuzzy Petri net model to model fuzzy IF-THEN rules and fuzzy IF-THEN-ELSE rules of rule- based systems. We also develop an efficient algorithm for performing vague reasoning automatically based on the extended fuzzy Petri net model. The proposed algorithm allows the computers to perform reasoning in a more flexible and more intelligent manner. Based on the proposed vague reasoning algorithm, we have implemented a vague reasoning system on a Pentium PC for performing vague reasoning automatically. Furthermore, we propose a method to describe the relationships between states and events for temporal knowledge representation and reasoning using time Petri nets. We also propose an algorithm to infer the consistency of the temporal knowledge based on the time Petri nets. Based on the proposed temporal reasoning algorithm, we have implemented a temporal reasoning system on a Pentium PC to check the consistency of the temporal knowledge.