A STUDY ON THE DILEMMA OF LEARNING IN THE LINEAR SUPPLY FUNCTION BIDDING IN POWER MARKETS

摘要

This paper addresses the issue of the market evolution with repeated linear supply function bidding. Under this circumstance, it is expected that power suppliers will learn from the market for making optimal decisions. However, it is unclear whether the market can converge; and if so, what is the feature of the equilibrium point. With detailed mathematical formulations, it is shown in this paper that supply function equilibrium is a Nash equilibrium; and that in certain condition the overall learning behaviors will reduce the MCP which in turn increases consumers' surplus and decreases suppliers' profits; while in some other conditions the results are just the contrary, which is indeed a dilemma of linear supply function bidding. The economic mechanism on why and how such dilemma will happen is discussed. Numerical results support the analytical conclusions very well.