Modeling and evaluating electricity options markets with intelligent agents

Under deregulation, the formation of electricity markets is a topic of great interest in the power industry and in financial institutions worldwide. Using derivative financial instruments (including options) becomes important for hedging against uncertainty and managing risk-limiting exposure to adverse market conditions. Black and Scholes' equation is often used to value options, but its validity is questionable due to assumptions that may not hold for electricity, most notably the assumption of log-normally distributed prices for the underlying commodity. In this research, a put options market for electricity is modeled. Adaptive agents trade in this market to maximize profit. They are not forced to use an explicit economic or financial model (e.g., Black-Scholes) in their valuation. A genetic algorithm (GA) is used to find alternate valuations that are used to generate buy and sell signals. The results show that it is possible to evolve profitable valuations for use with buying and selling options in this simple model. Reasons for and implications of this finding (e.g., that Black-Scholes may not be a good method for pricing electricity derivatives) are discussed.