透過您的圖書館登入
IP:3.141.47.221
  • 學位論文

運用語意對映輔助電子化工作流程建置

Applying Semantic Matching to Assist the Process Construction of Electronic Workflow

指導教授 : 項衛中

摘要


在運用網路服務的環境中,流程開發者透過服務仲介者UDDI(Universal Description, Discovery, and Integration)來協助尋找合適的網路服務,且利用關鍵字搜尋,但此搜尋方式不容易搜尋到所需要的網路服務,而其內容資訊也無法讓使用者判斷搜尋到的網路服務是否合適。現今在企業描述電子化工作流程的技術規範中,以BPEL(Business Process Execution Language)較成熟且被廣泛使用,但無專屬的搜尋機制。本研究為了解決此問題,提出開發改良的UDDI系統,並提供工作流程的語意資訊讓使用者能了解工作流程的內容,以此系統註冊及搜尋BPEL所描述的工作流程,且在系統中使用語意對映提高搜尋結果的準確度。 本研究參考OWL-S(Web Ontology Language for Services)的規範並將其Profile的描述格式加以延伸,讓使用者能定義BPEL的語意資訊。在搜尋工作流程時,系統將依據使用者所輸入的需求與已註冊工作流程的語意資訊加以對映比較,將所得的相似度結果作為使用者選擇工作流程的參考建議。 本研究所改良的UDDI系統具有註冊、修改、搜尋工作流程的功能,文中以三個範例分別說明系統操作的方式。為了驗證使用語意對映可以提升搜尋功能,本研究針對語意對映與關鍵字兩種搜尋方式做比較,利用對映品質指標Recall及Precision計算各別搜尋結果的績效,經由二個範例的搜尋結果比較,得知語意對映的搜尋方式有較好的對映品質。因此本研究所改良的UDDI系統,可以提高搜尋所需要工作流程的準確度,並且輔助使用者完成電子化工作流程的建置。

關鍵字

BPEL UDDI 語意對映 OWL-S 工作流程

並列摘要


Users may find available Web Services through the service broker UDDI (Universal Description, Discovery, and Integration) to develop workflows. In general, the UDDI system supports users to search Web Services by keywords. The searching result usually does not provide sufficient information for users to decide which Web Services can fulfill the requirements. Although BPEL (Business Process Execution Language) is widely used in electronic workflows, there is no well-designed system to help users to find BPEL services. This research proposes to establish an improved UDDI system which can register and search BPEL descriptions and use semantic mapping to enhance the search results. This research extended the format of OWL-S (Web Ontology Language for Services) profile description so that users can define semantic information of workflows. Moreover, this research calculated semantic similarity between the required BPEL and the registered workflows. Users could choose the preferred workflow based on the similarity results. In this thesis, three examples were utilized to show the functions of the improved UDDI system such as registering, modifying and searching the semantic information of workflows. This research used two examples to compare the performance of semantic mapping method and the keyword method with the Recall and Precision indices. The results showed that the semantic mapping method had better searching performance than the keyword method; therefore, this improved UDDI system can help workflow developers to build workflow more efficiently.

並列關鍵字

OWL-S workflow semantic matching BPEL UDDI

參考文獻


[24] 蔡宗諺,(2009),「運用限制式校正距離與語意相似度比較模組化產品物料清單」,
[23] 葉榕真,(2008),「在資料對映中運用機器學習的初探」,中原大學工業工程學系碩
[22] 陳信豪,(2007),「BPEL之語意特性資料」,國立中央大學資訊工程研究所碩士論文。
[21] 陳信志,(2007),「運用混合式綱要對應方法增進資料交換中綱要對應的效能」,中
[14] Paolucci Massimo, Takahiro Kawamura, Terry R. Payne, Katia Sycara. (2002). Importing

延伸閱讀