abstract
- Effective optimization of model variables is essential yet demanding in engineering and industrial processes. Metaheuristics (MHs) offer a proficient approach, but their design and tuning incorporate notable challenges. Automated Algorithm Design (AAD) methodologies provide a solution by enabling automated algorithm construction. This study utilizes a Hyper-Heuristic (HH) framework within AAD, which obtains a tailored MH to optimize a Proportional, Integral, and Derivative (PID) controller in an Automatic Voltage Regulation (AVR) system. We identify a preference for search operators from Spiral Dynamic, Swarm Dynamic, and Differential Mutation families, offering valuable insights for MH algorithm design in complex electrical systems. Our contributions include a novel methodology for distilling search operators for specific problem families and presenting effective search operators for MHs in electrical engineering scenarios. The study highlights the importance of precise controller tuning, demonstrated through the effectiveness of the tailored MH compared to others. © 2024 IEEE.