Power generation mix optimization using mean-lower partial moments (LPM) portfolio theory

Abstract

Optimization of power generation technology mix using portfolio theory is related to finding the optimal set of technologies under acceptable level of (price) risk which will provide minimal cost of electricity production for the generation company, or provide maximum profit. On the other hand, a generation company can set the cost of production as a fixed parameter, and then look for optimal set of technologies which would minimize price risk.The classical approach to power generation mix optimization considers renewable energy as a generation technology without price risk, or to a certain extent considers that risk being very small. In this work intermittency of renewable energy sources and accuracy in the day-ahead forecast was taken into account in the evaluation of price risk of those technologies. Energy not delivered because of wrong forecast must be bought on the balancing market and poses a burden on the price of production from those technologies. Portfolio optimization is performed using mean-LPM approach and compared to the results given by mean-variance approach. The results of the optimization show that based on the historical prices mean-variance and mean-LPM optimization give similar results only in case of second order LPM. Other orders of lower partial moments can account for risk aversion of the investor or decision maker.