Nash Genetic Algorithm based optimal design of hysteresis inverters for active power filtering applications

In this paper optimal design of a current regulated-voltage source inverter (VSI) to be used for active filtering applications using game theory based multi-objective optimization methods is approached. The Inverter hysteresis band and the DC bus voltage are considered as design variables, while the average switching frequency and the tracking error are considered as two distinct objective functions to be minimized. For the first time, the Nash Genetic Algorithm (Nash-GA), a co-evolutionary optimization approach based on the Nash Equilibrium concept is introduced and used to reach the optimal values for the DC bus voltage as well as the hysteresis band. Both parameters affect the loss and tracking performance of the Inverter. The simulation results obtained in MATLAB/SIMULINK environment prove the superiority of the novel approach presented.