조회 수 : 15746
2011.08.19 (22:29:22)


Trading Agent Competition


One characteristics of complex socio-economic systems is the participation of self-interested agents to the system. SESLab investigates the emergent phenomena of the systems and the agents when such self-interested agents consists of a market as a complex system. For instance, a number of self-interested agents compete and negotiates to maximize their profits in a supply market. Such agents should consider the inventory, the competition, and the pricing. At the same time, they should also assume that the other agents may interfere their operations through monopolies and supply pricing changes. Then, the agents build strategies that, firstly, secure a stable supply chain with a competitive price and, secondly, adjust their supply bids and volumes to obstruct the other agents' operations. This emerging competitions between agents is a key to understand various complex socio-economic systems, and SESLab studies such emergence through a model of a hypothetical trading market.


tacscm.jpg

Retrieved from http://www.sics.se/tac/


Trading Agent Competition (TAC) is a well-known model of complex trading markets, and SESLab follows and joins the tournament of strategy-building fights between agents by developing one of the agents. The agents utilize advanced machine learning and game theoretic analysis which is a combination of micro-economics and artificial intelligence. Furthermore, besides of participating TAC, SESLab applies the developed agent behavior to a hypothetical future market, such as a power market with smart grids.


smart-grid market.jpg



Tag List