Evolutionary optimization under uncertainty in energy management systems

Abstract

To support the utilization of renewable energies, an optimized operation of energy systems is important. In recent years, many different optimization methods have been used in this field, including exact solvers and metaheuristics. Quite often, evolutionary algorithms yield good optimization results and allow for a flexible formulation of the optimization problem. Nevertheless, most approaches do not respect the dynamic nature of energy systems with time-dependent properties and stochastic variations. In this work, typical uncertainties are categorized and appropriate measures that help handling uncertainties in energy systems are presented and evaluated using an implementation of a building energy management system that may be used in simulation and practical application.