Stavrogiannis, Lampros C.
Development of an Autonomous Double Auction Market for its Effective Operation in Competitive Environments
Aristotle University of Thessaloniki, Dept. of Electrical and Computer Engineering,
- Version of Record
The objective of this thesis is the development of an autonomous Double Auction (DA) market capable of operating efficiently in an isolated as well as in competitive environments. A well known example of such an institution is the stock market. Besides their multi-annual study and operation, DAs continue constituting a theoretical paradox. They succeed to exhibit an increased efficiency with the implementation of very simple rules, even if there still has not been achieved a satisfactory theoretical model from the field of Mechanism Design. In order to overcome this shortcoming, researchers have turned their attention to experimental techniques for the study and implementation of new, innovative rules, a field known as Automated Mechanism Design. However, all studies up to now deal with markets that operate free of charge in an isolated environment, something which does not correspond well in today’s global economy, where each country’s stock markets compete with each other as well as with the remainder stock exchanges worldwide in order to achieve high profits and market-share. TAC Market Design (or CAT) is an attempt to study this kind of institutions and made its appearance in 2007. In CAT, entrants represent stock markets that compete with each other while being evaluated in a number of realistic criteria. This thesis presents the game of CAT and all the strategies that were implemented by our agent, Mertacor, with which we participated these two years that the competition is being conducted and which was placed 8th and 5th respectively.
||Double Auctions, Mechanism Design, Trading Agent Competition (TAC), Market Based Control (MBC), Market Design
||Agents, Interactions & Complexity, Electronics & Computer Science
||30 May 2011 12:22
||17 Apr 2017 17:45
|Further Information:||Google Scholar|
Actions (login required)