Information Gap Decision Theory and data mining for competitive biddingCheong, Mei-Peng (2004) Information Gap Decision Theory and data mining for competitive bidding. Masters thesis, Iowa State University. Full text available as:
AbstractThe restructuring process associated with the economic deregulation of the electricity supply industry is transforming the electricity market from a vertically integrated industry to a horizontally integrated open market system. This new competitive framework affects the operational planning of generation companies. Traditionally a cost-minimizing process aimed at feasible dispatch of its units, the generation company now needs to develop bidding models to maximize its benefits. The objective of this research is to advance strategic bidding decision-making to not only consider the technical aspects of unit operation such as capacity limits, but also to incorporate information about other market participants and the uncertainty of their behaviors. These additional factors are significant especially in an oligopoly market because they influence the amount of electricity sold and purchased, hence affecting net profit. This thesis presents a decision tree approach to a basic bidding problem and applies information gap decision theory to quantify uncertainty. The information gap decision theory estimates the uncertainty associated with each bid of the generating unit. With a minimum reward level specified using different decision criterion, a feasible range of bids and the maximum price bid can be obtained. Data mining technique is proposed next to infer the behavior of the competitors and improve the bidding strategy. The results show that one can exploit the additional knowledge discovered to outperform other market participants.
Archive Staff Only: edit this record |
