Abstract
In this thesis, the problems of forecasting the market clearing price (MCP) in California electricity markets and the optimizing bidding strategies of a generator owner are studied. The MCP forecasting helps producers and consumers of electric power to prepare their bidding strategies, and, consequently to maximize their profits. MCP prediction is a difficult task since bidding strategies used by market participants are complicated and there are many uncertainties that interact with one another. Nevertheless, bidding strategies can be optimized based on the forecasted MCP. In this thesis, the autoregressive with exogenous variable (ARX) technique is used to forecast the hourly MCPs. The MCP forecasts are modeled using current and lagged values of hourly electricity consumption, current and lagged values of the hourly MCP, hourly natural gas prices, and hourly energy produced from renewable sources. Optimal lag lengths are chosen based on the lowest AIC, the highest adjusted R2, and the lowest mean absolute percentage error (MAPE). The forecasted MCPs are then used tooptimize bidding strategies of a generator owner to maximize profits. The methods and techniques used in this thesis are able to forecast future electricity prices in California as well as those obtained by other researchers, but use only using publicly available, and not proprietary, data sources.