UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Automating Internet Auctions with Adaptable Mobile Agents

Seymour, Mark; (1999) Automating Internet Auctions with Adaptable Mobile Agents. Doctoral thesis (Ph.D), UCL (University College London). Green open access

[thumbnail of Automating internet auctions with adaptable mobile agents.pdf] Text
Automating internet auctions with adaptable mobile agents.pdf

Download (10MB)

Abstract

This thesis investigates the automation of different auction methods using computer technology. There are many auction methods. However, four methods are considered for automation. They are First Price Sealed-bid (FPSB), Vickrey, English (e.g. antique and cattle auctions) and Dutch (e.g. flower and tobacco auctions). Since these auctions have well defined bidding and selling rules, it is feasible that the whole process of bidding and selling could be automated. In particular, this thesis investigates automating these auctions using software agents and the Internet. The main aim of this thesis was to create an Internet Auction System (IAS) where users can auction items on their own computer and find and bid in remote auctions held on other users' computers. Adaptable (Seller) agents are used to automate selling and Adaptable Mobile (Bidder) Agents (AMA) are used to find these remote auctions and to automate the bidding process once they arrive there. The actual items auctioned are not important, though it is intended that agents can only sell or bid for one item at a time. Developing the IAS is important for two reasons. Firstly, by demonstrating that AMA(s) can automate decentralised auctions illustrates how they could make effective tools to automate other electronic commerce and trading systems. Secondly, by effectively embedding intelligence into agents, users would have greater confidence in allowing agents to operate in these markets on their behalf. There are four parts to this thesis: (i) the investigation of existing agent-based auction systems; (ii) the development of AMA(s); (iii) the utilisation of these agents to build an IAS and (iv) an evaluation of the IAS using AMA(s). Researching existing agent-based auction systems highlighted the lack of fully automated auction systems. Most Internet or Web-based auctions do not use agents. They require users to submit bids via e-mail or web page forms. Fewer still give users any control over where or when the auction is conducted or what auction method is used. Therefore, in this thesis agent scripts were implemented as finite automata that were both mobile over the Internet and self-modifiable. Within each script, the agent's intelligence was encoded using fuzzy-genetic strategies. This enabled users to evolve their agents in an auction simulator that used Fuzzy-Genetic Algorithms (FGA). These tools and concepts were then used to design the IAS. The IAS was assessed by how well the AMA(s) automated various auction methods and by how well the auction simulator evolved rational bidding and selling strategies. Three specific research contributions were made to agent-based auction systems. Firstly, novel AMA(s) were developed that used self-modifying SGML-based scripts with embedded fuzzy-genetic adaptable strategies. Secondly, AMA(s) were successfully used to automate various auction methods over the Internet. Thirdly, an auction simulator was developed that enabled users to evolve profitable (and in many cases rational) bidding and selling strategies for their respective Bidder and Seller agents to use in the IAS auctions.

Type: Thesis (Doctoral)
Qualification: Ph.D
Title: Automating Internet Auctions with Adaptable Mobile Agents
Open access status: An open access version is available from UCL Discovery
Language: English
Additional information: Thesis digitised by ProQuest
Keywords: Applied sciences; Internet auctions
URI: https://discovery.ucl.ac.uk/id/eprint/10099413
Downloads since deposit
26Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item