Application of the Hybrid Multiobjective Optimization Methods on the Capacitive Voltage Divider


Abstract

This paper presents application of multiobjective optimization methods on a capacitive divider finite element model. Modeling is based on the differential evolution algorithm (DE), including the concept of the Pareto nondominance. Through calculations, this concept allows differential evolution algorithm variation in order to achieve uniformly distributed results on the Pareto front, including the lowest population size as possible. Selection step, as a part of the differential evolution method, is properly modified to improve the convergence of the algorithm. The entire research leads to a hybrid algorithm development.