Building optimal offer curves for an electricity spot market: a decomposition solution methodology

Abstract

We present a Mixed Integer Programming (MIP) approach to optimize the hydro-production offers submitted by a power generation company to an electricity spot market in which offer curves are required to be stepwise functions. Our method takes an initial set of offer curves and proposes changes that increase the expected profit of the generation company while complying with a variety of constraints such as preserving the stepwise shape of the curves or accomplishing guidelines proposed by other longer time-scope models. As well as optimal offer curves, solution yields a set of shadow prices to be contrasted with the imposed guidelines. Market uncertainty is taken into account by means of scenarios of residual demand curves. Modeling each scenario of hourly residual demand curves requires a large number of binary variables which dramatically increases the size of the problem, making it unaffordable for commercial optimizers. Since the model is expected to be used in a realistic case, no simplifications are introduced into the model and two different solution approaches are adopted to solve this problem: Lagrangean Relaxation decomposition and nested Benders decomposition. On the one hand, the Lagrangean Relaxation approach provides a valuable insight about medium-term guidelines, although the feasibility of the solution is not guaranteed. On the other hand, an algorithm based on the nested Benders decomposition for integer programming is implemented in order to achieve feasibility.