Optimal Bidding Strategy in Day-Ahead Electricity Market for Large Consumers

Electricity price forecasting has become an essential task for electricity buyers and sellers participating in competitive power markets. Compared to the existing point forecasting methods, probabilistic forecasting approaches provide more information about the future forecasts of electricity prices. This paper uses the double exponential smoothing (DES) and triple exponential smoothing (TES) methods to compute volatility forecasts of day-ahead electricity prices. Moreover, we use the elastic net regularization to compute regularized forecasts for volatility. Sample sign correlation of standardized electricity prices (standardized by volatility forecasts) is used to identify the conditional distribution of the electricity price series. Validation of the regularized volatility forecasts is demonstrated using the publicly available hourly Ontario electricity prices. Our data analysis shows that TES forecasts of volatility outperform DES forecasts. Besides, elastic net regularization decreases the mean absolute error of the TES day-ahead volatility forecasts from 11.98 to 11.07. Energy procurement of a large consumer is modelled as an optimization problem to find the optimal bidding strategy. First, a gradient boosting regression (GBR) method computes the optimum electricity prices to be placed in the day-ahead market. Then, a linear programming method is used to obtain the optimum strategy by computing the quantities that need to be bought from the market for the upcoming day. Our simulation results indicate that using probabilistic forecasts of electricity prices leads to a more flexible and efficient bidding strategy than using the point forecasts.