整合行動代理人及本體論於雲端環境上的服務發現

Integrating Mobile Agent and Ontology for Services Discovery on Cloud Environment

張哲祥(Che-Hsiang Chang)

指導教授 ： 張玉山

國立臺北大學/電機資訊學院/資訊工程學系/碩士(2011年)

摘要

隨著網路的快速發展，雲端運算與我們生活息息相關，舉凡我們使用的軟體、硬體、設施、
平台…等都已經有雲端化的趨勢。而我們在這樣一個服務琳瑯滿目的環境裡，要如何快速及準確
找到適合我們的服務，是一個很迫切的需要。
在這篇論文中，實作了一個雲端服務發現的平台，並且整合行動代理人技術與本體論的概念。
整合行動代理人可將這項平台封裝為一項行動服務，給使用者隨時隨地去使用，查詢他們想要的服
務；而本體論則用於將服務分類，將使用者想要的服務進行描繪，得其定義及描述，進而根據使用
者的查詢語句，分析語意得知想要查詢的服務為何。
預先定義的本體論有二，其一為以服務為導向結構的本體論，以此將現有雲端服務分類，從基
礎三個概念SaaS、PaaS、IaaS進而延伸劃分；其二是知識本體，用來處理使用者輸入的自然語言去
分析比對。在系統實作中設計了一些代理人如推理代理人、發現代理人、檢索代理人及過濾代理人，去完成整個查詢流程。
最後進行了一些實驗，分別評估每個代理人的執行時間及討論如何使得推理能夠更加準確，從
增加推理規則及擴充分類去實驗；比較部分跟現有的服務搜尋系統及資訊檢索比較。比較的結果因
為我們學習參考文獻的方法，從查詢到分析進而得到結果有了更完整的架構，實驗結果也可判斷出
如何有效讓這個系統更加完善，使得推理出的結論更讓人滿意。

關鍵字

代理人 ； 本體論 ； 雲端 ； 服務

並列摘要

With the rapid development of Internet, cloud computing is very close our life, the software, hardware, infrastructure and platform we use all be cloudify. And we such in this environment of various services how do we quickly and accurately find a suitable service is very urgent need.
In this paper, we integrate mobile agent and ontology to implement a system of discovering cloud service, so that we can package this platform with a mobile service, and provide to user for query a cloud service they want anytime anywhere; in addition to the ontology is classify the service, depict the definition and description of it, and then analyze the sentence to know what is the service user query.
Thus, we have two pre-defined ontology, the one is SOA ontology, which is used to classify existing service, extend and divide from three based concept as SaaS, PaaS, and IaaS; the second one is knowledge ontology, which is used to handle the nature language of user query. In this prototype we design multi-agent like Reasoning Agent, Discovering Agent, Retrieving Agent, and Filtering Agent to finish the whole flow.
Final we have some experiment, individually estimate the execute time of each agent, and we have some discussion about how do we reason more accurate from add inference rules and expend the class; and we still have some comparison between existing service discovery system and information retrieval from references. And we study and learn the advantage of these researches, let we have the complete architecture from query to analyze; the result of experiment is know some method to improve our system, and get the conclusion after reasoning is satisfying.

並列關鍵字

Agent ； Ontology ； Cloud ； Service

參考文獻

[6] Brain Meets Brawn: Why Grid and Agents Need Each Other (2004)

by Ian Foster , Carl Kesselman , Nicholas Jennings

[7] Cloud Computing and Grid Computing 360-Degree Compared, Foster, I.; Yong Zhao; Raicu, I.; Lu, S.; Dept. of Comput. Sci., Univ. of Chicago, Chicago, IL, USA

[12] T. Finin, et al., “KQML as an Agent Communication Language,

[13] FIPA 97 Part 2 Version 2.0: Agent Communication Language Specification,