Bridging the Gap between Scheduling and Control of Distributed Energy Resources: Dynamic Parametrization of a Metaheuristic Method

Bridging the gap between market-oriented operational scheduling and grid-oriented control of distributed energy resources entails the formulation of a nonlinear optimization problem solvable with metaheuristic methods. Despite the generally advantageous computing time of metaheuristics compared to complex nonlinear problems, metaheuristic methods require a time-consuming parametrization of each new instance of the underlying optimization problem. Considering the time constraints of control strategies for distributed energy resources, the time-consuming parametrization poses a significant challenge. This paper introduces dynamic parametrization approaches based on machine learning, which can increase the performance of the parametrization process.