以網路服務分群輔助動態商業流程整合之研究

Clustering Web Services for Dynamic Business Process Integration

陳宇軒(Yu-Hsuan Chen)

指導教授 ： 吳忠敏

國立臺北科技大學/管理學院/商業自動化與管理研究所/碩士(2005年)

摘要

企業的商業流程跨越許多商業夥伴，流程中的參與成員眾多，如何整合跨企業之商業流程為關鍵的議題之一，商業流程可分成「企業對企業」與「企業對消費者」兩種。本研究以企業對消費者之流程作為整合的對象。服務導向架構下的流程整合，可將流程中的活動封裝成一個網路服務，以網路服務組合的方式形成更貼近商業流程的複合服務。
現今網路服務數量繁多，各自獨立運作且分散於不同的組織中，造成服務消費者在瀏覽與使用服務時的不便，因此須將服務集中於一個平台上，透過本研究提出「以分群輔助動態網路服務組合」之機制，將平台內的服務予以整合，而不採用事先定義靜態流程模板的方式。研究方法係透過群組技術中的鏈結能量演算法，將網路服務予以群聚，再找出群內服務之間的前後順序，可依服務消費者點選之服務，自動地推薦同群體內的所有前服務與後服務。服務消費者與平台的互動下，可減少服務搜尋的時間，並動態產生客製化流程。
本研究以旅遊產業為背景，建立服務整合平台，平台內註冊了三十八個旅遊相關的網路服務，經分群演算法的運算，分為二群。實際應用本研究發展之動態網路服務組合機制，提供企業對消費者之整合式電子商務服務。

關鍵字

網路服務分群 ； 動態網路服務組合 ； 網路服務

並列摘要

The business process involves a multitude of partners and participants. How to integrate various business processes across corporations is thus a key issue for today’s top management. General business process falls into two categories: “business to business” and “business to consumer”. This research studies the “business to consumer” category, trying to find out how to integrate the business process under a service-oriented architecture by transforming an activity in a process into a web service, making the business process more streamlined by the technique of Web service composition.
Nowadays, there are often too many web services on the internet, existing discreetly in and operated independently by different organizations, making it difficult and inconvenient for the service requester to browse. To make it easy for the requester to make use of various services offered, we can pack all the services into a platform, using the techniques of Clustering Web Services for Dynamic Business Process Integration as developed by this study to integrate the services, thus dispensing the need to pre-define all the static process templates of all the web services. The method employed by this study is that by using Bond Energy Algorithm, the web services available at the platform are first clustered according to their characters, and then precedence orders of the services within each cluster are determined. The system will automatically record the services a requester browsed and then recommend to him or her the related pre- and after-services. This interaction between the requester and the platform creates a customized browsing process that greatly reduces the time the requester spends in searching for his/her desired services.
This study uses the traveling industry as a model on which a platform comprising 38 related traveling web services grouped into two clusters through clustering algorithm is built. By making use of the dynamic web service composition mechanism developed by this study, one can obtain a new integrated “business to consumer” e-commerce service system.

並列關鍵字

Web Service Clustering ； Dynamic Web Service Composition ； Web Services

參考文獻

1. Arabie, P. and L. J. Hubert, “The bond energy algorithm revisited,” IEEE Transactions on Systems, Man and Cybernetics, vol. 20, no. 1, 1990, pp. 268-274.

7. Choi, I., H. Jeong, M. Song and Y. U. Ryu, “IPM-EPDL: An XML-based executable process definition language,” Computers in Industry, vol. 56, no. 1, 2005, pp. 85-104.

8. Chandrasekaran, S., J. A. Miller, G. S. Silver, B. Arpinar and A. P. Sheth, “Performance analysis and simulation of composite Web services,” The International Journal of Electronic Commerce and Business Media (EM), vol. 13, no. 2, 2003, pp. 120-132.

9. Chen, P. C., The trip planner: Exploring an ontology-based Web service composition and brokering framework, unpublished master’s thesis, Department of Information Management, National Taiwan University, Taipei, Taiwan, 2004.

11. Da Costa, L. A. G., P. F. Pires and M. Mattoso, “Automatic composition of Web services with contingency plans,” Proceedings – IEEE International Congress on Web Services (ICWS 2004), San. Diego, CA, USA, 2004, pp. 454-461.