We consider the optimization problem of a generating company that operates in a competitive electricity market and wants to select energy offering strategies for its generation units, with the aim of maximizing the profit while considering the uncertainty of market price. We focus on the case of a price-taker company, i.e. the company cannot influence the market price through its generation decisions. Such optimization problem can be modeled as a quadratic mixed integer program. Price uncertainty represents a major issue for price-taker companies: the hourly prices of the market are not known in advance and companies must thus take their offering decisions by making some kind of assumptions about the future unknown prices. Over the years, many methods have been proposed to tackle price-uncertain energy offering problems. Among these methods, Robust Optimization (RO) has recently known a big success and has been widely applied in power system optimization also in the context of energy offering. In this work, we review central references available in literature about the use of RO in energy offering, pointing out their limits, which severely reduce their applicability in practice, exposing to the risk of suboptimal and even infeasible offering. We then propose a new RO method for energy offering that overcomes all the limits of other RO models. We show the effectiveness of our new method on instances provided by our industrial partners and considering real prices from the Italian energy market, getting very high increases in profit. Our method is based on Multiband Robustness, an RO model that has been proposed by Büsing and D’Andreagiovanni in 2012 (e.g., [2]) to generalize and refine the representation of the uncertainty of the classical Gamma-Robustness model by Bertsimas and Sim, while maintaining its computational tractability and accessibility. Multiband Robustness is essentially based on adopting a histogram-like cardinality-constrained uncertainty set that results particularly suitable to represent asymmetric empirical distributions that are commonly available in real-world optimization problems subject to data uncertainty [1].
Part of the results of this talk are presented in [3].
Essential references
[1] T. Bauschert, C. Büsing, F. D’Andreagiovanni, A.M.C.A. Koster, M. Kutschka, U.Steglich. Network Planning under Demand Uncertainty with Robust Optimization. IEEE Communications Magazine, 52(2), 178-185, 2014
[2] C. Büsing, F. D’Andreagiovanni. New Results about Multi-band Uncertainty in Robust Optimization. Experimental Algorithms - SEA 2012, LNCS 7276, 63-74, Springer, 2012.
[3] F. D’Andreagiovanni, G. Felici, F. Lacalandra. Revisiting the use of Robust Optimization for energy offering. Preprint available at: http://arxiv.org/abs/1601.01728, 2016