Second Order Stochastic Dominance Portfolio Optimization for an Electric Energy Company

Abstract

This paper presents a framework of portfolio optimization for energy markets from an electric energy company's perspective. The objective of this research is to determine the best possible investment plan by combining two potentially conflicting portfolio investment goals. First, given the general characteristics of the generating assets and forecast of market variables, the decision maker selects an efficient set of portfolios by optimizing the expected portfolio return. Secondly, an optimal portfolio is chosen based on company's risk profile. This risk is controlled by guaranteeing that the portfolio model has second-order stochastic dominance (SSD) over the cumulative distribution of a minimum tolerable reference distribution. Decision criteria are then applied to obtain an optimal and robust portfolio. The proposed approach is used to determine the amount of optimal market share value that maximizes the expected value of the profit. This is performed by treating risk as a distribution that represents the minimum expected profit acceptable by the energy company. Results show that different risk profile leads to different optimal portfolio. The optimal portfolio which gives the highest expected profit may not have the best robustness. This approach is also applicable to problems characterized by other sources of epistemic uncertainty besides unknown dependencies.