Strategic Generation Bidding and Scheduling under Price Uncertainty

Market deregulation, emergence of new technologies, and the rising penetration of variable renewable generators are increasing electricity price variability and the risk faced by generation portfolios in spot markets. In this paper, we propose a method to suggest bidding strategies for price-taking generators in a day-ahead market under price uncertainty. Electricity price scenarios are generated through a seasonal ARIMA model and grouped into clusters using the Partitioning Around Medoids algorithm (PAM) to obtain representative price scenarios. Expected profits under price uncertainty are maximized using a Stochastic Mixed Integer Linear Programming (SMILP) formulation that considers the operational restrictions of the thermal units. The problem is solved in XPRESS-Mosel. The results show that explicitly modeling price variability in the optimization can increase the benefits seen by a price-taking generation portfolio by up to 1.4% with respect to a deterministic approach.