A Mixed Integer Stochastic Optimization Model for Settlement Risk in Retail Electric Power Markets

Recent changes in the US electric power markets have contributed to volatility in hourly prices and loads. In this paper we consider the position of the electric power retailer who typically contracts with suppliers and end-users and must provide future load requirements to the suppliers. As part of this energy supply chain, the retailer is faced with great uncertainty in both market prices as well as end-user loads. Based on actual data for the PJM market covering Pennsylvania, New Jersey, and Maryland, we develop a probabilistic optimization model to optimize the net profits for the retailer for a forecast time horizon (typically one or more hours) given the cumulative performance in previous time periods (hours). The resulting model is formulated as a mixed integer linear program with binary variables due to the disjunctive nature of certain forward load estimation "bandwidth" tolerance constraints. In addition, we also provide an existence result to this optimization model. Lastly, we present a numerical example of the optimization model to validate its workings and provide some insight into model sensitivities. [PUBLICATION ABSTRACT]