斐氏圖軟體撰碼問題中自動規範測試系統的設計與實作

Abstract

對自動化製造系統而言，發展它的控制軟體過程可分為設計與實作兩階段。對設計階段而言，連續兩製造狀態的因果關係可用斐氏圖來建模，而所有可到達狀態的集合則稱為設計可達圖。相似地，對實作階段而言，對應的斐氏圖可透過程式語言來撰寫，並自然產生其實作可達圖。又測試設計可達圖與實作可達圖的等效性是必要的。換言之，設計可達圖可用來產生測試序列並用來測試實作可達圖的規範。 本研究中，透過設計行為的可達圖產生一條測試序列。假如設計可達圖是歐氏有向圖，則測試序列可直接透過弗勒希演算法來產生。假如設計可達圖不是歐氏有向圖，則建構運輸模式來增加最少的額外邊並形成新的歐氏有向圖，再透過Fleury’s演算法來產生測試序列。可達圖中的生產循環也可藉由符號矩陣的對角線元素偵測之。又本文以爪哇語言實作出此規範測試系統來驗證這些構想的可行性。

For a given automated manufacturing system, the development of its control software usually involves a design phase and an implementation phase. In the design phase, the causal relation between two sequential manufacturing states can be modeled by a Petri net, and a directed network consisting of all the reachable states is named its designed reachability graph. Similarly, in the implementation phase, the corresponding Petri net is coded through a proper programming language, and its implemented reachability graph is generated naturally. Also it is necessary to test the equivalent conformance of both designed and implemented reachability graphs. In this thesis, the designed reachability graph is used to generate a testing sequence for testing the conformance of the implemented reachability graph. If the designed reachability graph is an Eulerian digraph, then its testing sequence can be directly generated by Fleury’s algorithm. If not, then a transportation model is constructed to add minimal extra edges for forming a new Eulerian digraph to which the Fleury’s algorithm can be applied. Also the cycles on the reachability graph can be detected from the diagonal elements of its powered symbol matrix. In addition, a Java-based conformance testing system has been implemented for showing the feasibility of this approach.