Robust Controller Design Based on a Combination of Genetic Algorithms and Competitive Learning

This paper investigates the robustness of power system stabilizer designs based on an evolutionary algorithm called Population-Based Incremental Learning (PBIL). PBIL combines Genetic Algorithms (GAs) and simple competitive learning derived from Artificial Neural Networks (ANN). The controller design issue is formulated as an optimization problem that is solved via PBIL algorithm. The resulting controllers (PBIL-PSSs) are tested over a wide range of operating conditions for robustness. Simulation results show that PBIL-PSSs are able to stabilize the system adequately over the entire range of operating conditions considered. PBIL-PSSs perform comparably to GA-PSSs under small disturbances but outperform GA-PSSs under large disturbances.