標題: 以模糊邏輯建構的品質導向網際網路服務選擇模型
A QoS-Based Web Service Selection Model Using Fuzzy Logic
作者: 林謂立
Lin, Wei-Li
羅濟群
趙國銘
Lo, Chi-Chun
Chao, Kuo-Ming
資訊管理研究所
關鍵字: 品質服務;網際網路服務;模糊邏輯;服務決策;QoS;Web Service;Fuzzy Logic;Service Selection
公開日期: 2008
摘要: 在本研究中,我們進行了兩階段以網際網路服務品質為導向的網際網路服務選擇模型研究 - QCMA模型(QCMA: QoS Consensus Moderation Approach) 與 FMG-QCMA模型 (Fuzzy Multi- Groups-Based QCMA)。第一階段的QCMA模型著重以辨識網際網路參與者對網際網路服務品質感知的相似度,進而確認這些參與者是否具高相似度,並根據已確認之高相似度參與者對網際網路服務品質因素偏好優先順序,決定這個高相似性群體對於網際網路服務選擇之決策模型。第二階段的FMG-QCMA模型著重於思考不同之消費者其差異過大的性格背景偏好差異,進而研究結合多維度的網際網路服務品質因素結構的分群演算法,建立更有效率的多群組架構網際網路選擇的決策模型。同時,在分群結構之中,因相似度些微低於分群比對的相似合格度,就被裁定網際網路服務品質意見為不相似,本研究也提出了模糊邊界的概念,將近似合格邊界的網際網路品質意見,納入為該群組之模糊相似意見,進而更明確的掌握分群有效性,避免分群失真現象。

本研究之模型適合任一種網際網路服務應用之目標客群分析,如線上旅行社、網際網路商城、網路拍賣會等網際網路應用服務。
In this research, two stages of modeling for QoS-aware selection of web service were established – QCMA (QoS Consensus Moderation Approach) and FMG-QCMA (Fuzzy Multi- Groups-Based QCMA). QCMA was proposed as the first stage in order to indentifying a group of participants by their high similarity and obtaining the group preference over all QoS attributes. FMG-QCMA was proposed as second stage in order to thinking over the distinct background and preference over QoS attributes among all web service participants. For this purpose a more efficient multi-attributes-based multi-groups clustering approach was studied for developing multi-groups-based QoS-aware selection model of web service. Also, the concept of fuzzy boundary, which is used for preventing possible omission of some opinions that should be treated as “similar” to group centre but cannot beyond the threshold distance defined in clustering criterion, was thought over.

The models in the research can be applied to “target customers analysis” on any web service application such as e-tourist agency, e-mall or e-auction.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT009234803
http://hdl.handle.net/11536/77185
Appears in Collections:Thesis


Files in This Item:

  1. 480302.pdf

If it is a zip file, please download the file and unzip it, then open index.html in a browser to view the full text content.