Supply Curves in Electricity Markets: A Framework For Dynamic Modeling and Monte Carlo Forecasting

During my MSc research, we developed a model for supply curves in short-term electricity markets. This model captured the contributions of various resources to the aggregate supply curve and accounted for the influence of external factors on both pricing and supply levels. To enhance its capabilities, we integrated a unified Monte Carlo approach for tracking latent variables, forecasting, and hyperparameter estimation. Specifically, we introduced a sequential Markov chain Monte Carlo (S-MCMC) algorithm to track latent variables, enabling day-ahead supply curve forecasting. For hyperparameter estimation, we incorporated S-MCMC into the expectation step of two variants of the expectation-maximization (EM) algorithm. We applied our framework to the Turkish electricity market and evaluated its performance using real market data, demonstrating that our method is competitive with state-of-the-art prediction models.

Recommended citation: S. Yıldırım, M. Khalafi, T. Güzel, H. Satık and M. Yılmaz, "Supply Curves in Electricity Markets: A Framework for Dynamic Modeling and Monte Carlo Forecasting," in IEEE Transactions on Power Systems, vol. 38, no. 4, pp. 3056-3069, July 2023, doi: 10.1109/TPWRS.2022.3208765