傅立成臺灣大學：資訊工程學研究所蔡明育Tsai, Ming-YuMing-YuTsai2007-11-262018-07-052007-11-262018-07-052004http://ntur.lib.ntu.edu.tw//handle/246246/53988隨著知識經濟時代的來臨，釵h企業組織面臨挑戰與瓶頸，他們的經營環境愈來愈趨於自由化、全球化及資訊化，另一方面，顧客關係管理也逐漸受到相當程度的重視，因此，製造商必須同時掌握品質、交貨日期與成本等競爭因素，所以，諸多企業不斷地提升本身製程能力，並且謹慎地進行供應商管理，才能確保最終產品與服務的品質，讓顧客滿足其需求。過去，公司內部如行銷、生產、採購等各部門往往各自為政，缺乏良好的溝通，公司之間缺乏良好的溝通橋樑，但這種方式在資訊驟變的環境下，公司將會失去其競爭優勢，故企業之間的整合和溝通日趨重要。 根據上述問題的描述，我們設計了一個媒介代理人的架構(Agent-Mediated E-mart Framework)，將企業間商業流程整合起來，在這個架構下，公司不但可透過合作結盟的方式，來增加本身的能力，而且，買賣雙方可利用中間代理人，來完成合適的配對，除此之外，一旦雙方交易關係確定，本研究採取自動化協商的方式，讓使用者得到最大的利益。最後，會以實驗數據來說明協商的結果和參與者的滿意程度。With the rising of Internet many online bargains successfully replace transactions in traditional market. Users can search proper goods/services through the Internet at any moment, and purchase them conveniently by e-cash or credit card everywhere. Therefore, we try to develop an advanced mechanism to integrate a series of processes of E-procurement, to facilitate every participant to perform bargains conveniently, and to obtain the largest benefits everywhere. Based on the above perspectives, an Agent-Mediated E-mart Framework is proposed in this these. It is a superior solution that integrates the processes of business transactions, and maximizes benefits for buyers and sellers. We accomplish several commerce related tasks in E-marketplace, such as partnership alliance, merchant brokering, and automatic negotiations in this framework. Especially, two automatic negotiations are applied to reach an agreement according to users’ choices. These achievements mark the most important stages of E-procurement in E-marketplace. In addition to the development of this framework, this thesis also provides several experiments that show the satisfactory outcomes of automatic negotiations. By these simulated experiments, we demonstrate that this framework not only integrates the processes of the E-procurement but also fulfills the automatic negotiation through intelligent agents.Chapter 1 Introduction 1 1.1 Motivations 1 1.2 Related Works 2 1.3 Contributions 3 1.4 Thesis Organization 5 Chapter 2 Preliminaries 6 2.1 Overview of E-marketplace 6 2.1.1 E-Distributors 7 2.1.2 E-Procurement 7 2.1.3 Exchanges 7 2.1.4 Consortia 8 2.2 Mediated Agents on E-commerce 9 2.3 Automatic Negotiation Agent System 11 2.3.1 BargainFinder 11 2.3.2 Jango 12 2.3.3 Kasbah and Tete-a-Tete (MIT Media Laboratory) 12 2.3.4 eMediator 13 2.3.5 Trading Agent Competition (TAC) 14 2.3.6 Multi AGent Negotiation Testbed (MAGNET) 14 2.4 Web services 16 2.4.1 SOAP (Simple Object Access Protocol) 17 2.4.2 WSDL (Web Service Description Language) 18 2.4.3 UDDI (Universal Description, Discovery and Integration) 19 2.5 Genetic Algorithm 20 Chapter 3 Agent-Mediated E-mart Framework 23 3.1 Overview of Mediated Agents in E-marketplace 23 3.2 Agent-Mediated E-mart Framework 24 3.2.1 Entry to E-marketplace 26 3.2.2 Coordinator Agent 28 3.2.3 Broker Agent 29 3.2.4 Facilitator Agent 31 3.2.5 Adapter Agent 32 3.3 Process-Oriented Virtual Enterprise Framework 33 3.3.1 Partnership Formation Agent 33 3.3.2 Overview of Virtual Enterprise Framework 35 3.3.3 Agent-Based Architecture for Enterprise 35 3.3.4 Agent-Based Architecture for Facilitator Agent 37 3.4 Hierarchical Broker Framework 38 3.4.1 Product/Merchant Brokering Agent 38 3.4.2 Overview of Hierarchical Broker Framework 42 3.4.3 Broker Agent architecture 44 3.4.4 Adapter Agent Architecture 46 Chapter 4 Automatic Negotiations 48 4.1 Overview of Automatic Negotiation 48 4.2 Assumptions of Negotiation 50 4.3 Negotiation Processes 52 4.3.1 Mutual Negotiation 53 4.3.2 Third Party Negotiation 55 4.4 Terms of Genetic Algorithm 57 4.4.1 Definition of Fitness Function 57 4.4.2 Population of Bargain Combination. 59 4.4.3 Genetic Operator 60 4.4.4 Conditions of Striking a Bargain 61 4.5 Negotiation Processes under Different Market Relations 62 4.5.1 Single Seller and Single Buyer 63 4.5.2 Single buyer and Multiple Sellers 63 4.5.3 Multiple Buyers and Single Seller 64 4.5.4 Multiple Buyers and Multiple Sellers 65 Chapter 5 Experiment 67 5.1 Experiment Design 67 5.2 Experiment Parameters 68 5.3 Experiment Result 71 5.3.1 Single Seller vs. Single Buyer 71 5.3.2 Single Seller vs. Multiple Buyers 74 5.3.3 Multiple Buyers vs. Single Seller 77 5.3.4 Multiple Buyers vs. Multiple Sellers 79 5.4 Experiment Summary 80 Chapter 6 Conclusion 82 Reference 841646580 bytesapplication/pdfen-US自動化協商媒合代理人遺傳基因演算法電子市場GeneBrokerAutomatic NegotiationE-marketplacethesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/53988/1/ntu-93-R91922081-1.pdf