Probabilistic Electric Vehicle Charging Optimized With Genetic Algorithms and a Two-Stage Sampling Scheme

Probabilistic power flow studies represent essential challenges in nowadays power system operation and research. Here, especially the incorporation of intermittent supply plants with optimal control of dispatchable demand like electric vehicle charging power shows nondeterministic aspects. Using simulation-based optimization, such probabilistic and dynamic behavior can be fully integrated within the metaheuristic optimization process, yielding into a generic approach suitable for optimization in uncertain environments. A practical problem scenario is demonstrated that computes optimal charging schedules of a given electrified fleet in order to meet both power flow constraints of the distribution grid while satisfying vehicle-owners’ energy demand and considering stochastic supply of wind power plants. Since solution- evaluation through simulation is computational expensive, a new fitness-based sampling scheme will be proposed, that avoids unnecessary evaluations of less-performant solution candidates.