行動代理者 (Mobile agent) 是可在一異質網路系統 (Heterogeneous network system) 內從一主機 (Host) 遷移 (Migrate) 至另一主機併和其他代理者及分散資源系統(Distributed resources) 溝通及交互作用(Interact)之軟體程式。它們特別適合用來發展分散式及網路上的應用系統。

　　由於行動代理者被發送到行動代理者網路執行工作時，常常會與使用者或其他行動代理者通訊，來完成使用者所交待的工作，而要與行動代理者通訊，我們必須知道行動代理者的所在位置，由於行動代理者的移動具備有自主性(autonomy)，因此行動代理者的所在位置為不可預測的，所以我們需要一套搜尋機制來得到行動代理者的所在位置。

　　本篇論文主要提出一個具一般性且有效率的行動代理者搜尋機制 (Hybrid-Proxy)， 本機制參考行動代理者搜尋機制中著名的兩個方法，Home-Proxy和Forward-Proxy的方法，設計一套系統以擷取兩個方法之優點、以互補其各自之缺點，並利用叢集式拓樸的叢集概念(Cluster Concept)來協助行動代理者在移動時選擇所留的移動資訊為位置資訊(location update message)或是足跡資訊(forward point)，當想要與某ㄧ行動代理者通訊的使用者或是其他行動代理者，便可根據行動代理者所留下的移動資訊，尋跡找到該行動代理者。

　　本文另實作出一個模擬器，模擬真實行動代理者執行環境會發生的事件，然後實行三種搜尋機制，以蒐集模擬的結果來比較三種機制的優劣，並分出析本文所提出的搜尋機制適用於何種情形。

　　Mobile agents are software programs that can move from host to host and interact with other agents and distributed resources in a heterogeneous network. They are therefore particularly effective for developing distributed applications. Mobile computing is also gaining more and more popularity in the recent years as professionals discover the benefits of being able to connect to the vast of distributed information resources at all times without any spatial constraints.

　　After a mobile agent is launched for remote operations, it often needs to communicate with his owner user or other mobile agents to accomplish the jobs scheduled by its owner. On the other hand, when the owner of a mobile agent needs to communicate with the mobile agent, he needs to track down the mobile agent in order to send it messages. Since mobile agents move autonomously, it is hard to predict the latest location of a mobile agent. That is why a mobile agent searching mechanism is crucial to the communication mechanism of a mobile agent system.

　　This thesis proposes a mobile agent searching mechanism with generality and efficient. Named Hybrid-Proxy searching mechanism, it takes advantages from both the Home-Proxy and Forward-Proxy searching mechanisms but excludes their major deficiencies. It also uses the cluster concept to help the mobile agent on deciding which kind of information, location update message or forward point, should be left over as movement information when they move. When the users or the other mobile agents need to communicate with a mobile agent, they track down the mobile agent according the movement information it left.

　　This research also develops a simulator to simulate an entire mobile agent working environment. The three mechanism are all evaluated by this simulator to compare their performance and to find out in what conditions the mechanism performs better.

誌謝I
中文摘要II
AbstractIII
目錄IV
圖目錄VII
表目錄IX

第一章 緒論1
1.1研究背景與動機1
1.2研究目的2
1.3研究成果與貢獻3
1.4章節介紹3
第二章 相關研究探討與回顧4
2.1 行動代理者系統相關研究4
2.1.1 行動代理者概論4
2.1.2 行動代理者技術的優點6
2.1.3 行動代理者系統之架構與應用7
2.1.4 程式碼的移動性(Code Mobility)10
2.1.5 行動代理者技術與傳統主從式運算的比較13
2.1.6 行動代理者技術與Java applet的比較16
2.2 Home-Proxy搜尋機制17
2.3 Forward-Proxy搜尋機制19
2.4 鄰近節點搜尋機制20
第三章 混合搜尋機制架構與運作25
3.1 混合搜尋機制設計之動機與概念25
3.2 系統概觀26
3.3 叢集拓樸之建立27
3.3.1 鄰近節點搜尋機制27
3.3.2 使用Beaconing鄰近節點搜尋機制的原因29
3.3.3 使用Beaconing鄰近節點搜尋機制的相關問題29
3.3.4 使用Beaconing鄰近節點搜尋機制可能出錯的情形30
3.4 混合搜尋機制運作模式32
3.4.1 行動代理者的遷移32
3.4.2 搜尋行動代理者35
3.4.3 行動代理者伺服器失敗情形37
3.5 位置訊息與足跡之刪除38
第四章 混合搜尋機制模擬與結果40
4.1 行動代理者移動區域性質(Locality)高低的討論40
4.1.1 行動代理者移動具有中區域性質的情況41
4.1.2 行動代理者移動具有低區域性質的情況42
4.1.3 行動代理者移動具有高區域性質的情況43
4.2 叢集範圍大小的討論45
4.3 使用Beaconing建立叢集成員數量與所需訊息量比較46
4.4 模擬三種搜尋機制的結果與分析48
4.4.1 搜尋行動代理者訊息量比較49
4.5.2 執行搜尋機制所需要的訊息與記憶體比較53
4.5.3 搜尋行動代理者的成功率比較57
4.5 執行三種演算法的整理與比較60
4.5.1 移動區域性質比較60
4.5.2 移動次數比較62
4.5.3 總整理63
4.6 加入縮短足跡路徑的演算法之比較64
4.6.1 縮短足跡路徑演算法64
4.6.2 縮短足跡路徑之模擬結果與比較64
第五章 結論與未來工作68
5.1 研究成果與結論68
5.2 未來工作69
參考文獻71
自 述73

參考文獻
[1]C. Kommareddy, N. Shankar, and B. Bhattacharjee. “ Finding close friends on the Internet”, In Proceedings of ICNP, November 2001.

[2]James D. Guyton and Michael F. Schwartz. “Locating Nearby Copies of Replicated Internet Servers”, In Proceedings of SIGCOMM, pages 288–298, 1995.

[3]D. Lange and M. Oshinma, “Programming and Deploying Java Mobile Agents with Aglets”, Addison Wesly, 1998

[4]D. Katabi and J. Wroclawski. “ A Framework for Scalable Global IP-Anycast (GIA) ”. In Proceedings of ACM SIGCOMM2000, 2000.

[5]D. Katabi and J. Wroclawski. “A framework for global ipanycast (GIA) ”. Internet Draft, Internet Engineering Task Force, June 1999. Work in progress.

[6]Dwight Deugo. “Mobile Agent Messaging Models” , School of Computer Science, Carleton University, 2001 IEEE.

[7]Alberto Medina, Anukool Lakhina, Ibrahim Matta, John Byes. “BRITE：Universal Topology Generation From a User’s Perspective” , Computer Science Department Boston University, April 12,2001.

[8]ObjectSplace Inc. “The ObjectSplace Voyager Universal ORB. 1991.

[9]S. Hotz. “Routing Information Organization To Support Scalable Interdomain Routing with Heterogeneous Path Requirements”. PhD thesis, University of Southern California, 1996.

[10]S. Jamin, C. Jin, Y. Jin, D. Raz, Y. Shavitt, and L. Zhang. “On the placement of internet instrumentation. In Proceedings of Infocom”00, Tel Aviv, Israel, March 2000.

[11]Wang T. I. “A Mobile Agent Carrier Environment with Mobile Computing Facilities”, IIP: International Conference on Intelligent Information Processing, The 16th IFIP World Computer Congress. 21~25/08, 2000, Beijin.

[12]Wang T. I. “A Mobile Agent Carrier Environment for Mobile Information Retrieval”, 11-th International Conference on Database and Expert Systems Applications - DEXA 2000, 05~08/09, 2000, Greenwich, London.

[13]Wang, T.I. “A Mobile Agent Carrier Environment”, ICS2000, 6-8 December, 2000, Chiayi, Taiwan, R.O.C.

[14]Wang, T.I,“以行動代理者支援行動資訊擷取”, 第五屆人工智慧與應用研討, Taipei, 17/Nov. 2000.

[15]周晉立：“Mobile Agent Carrier Environment”，國立成功大學工程科學研究所碩士論文，June 1998.

[16]葉丁源 ：“行動代理者之位置追蹤機制”， 國立成功大學工程科學研究所碩士論文，June 2001.

[17]黃源龍 ：“使用行動代理者之互動式推播技術”， 國立成功大學工程科學研究所碩士論文，June 2002.