Multicriteria Optimization of Modal Filters Using Evolutionary Algorithms and Random Search Method

The paper proposes a novel approach to multicriteria optimization of strip modal filters (MF) that uses a modified algorithm of evolutionary strategies (ES) and a random search (RS) method. The results of such optimization are compared with the results of the genetic algorithm (GA). As an MF, we considered two structures: a structure based on a 3-conductor microstrip line and a 2-conductor MF with broad-side coupling. The MFs were optimized according to the amplitude criterion and the matching criterion. For all algorithms, we obtained the optimal values of the multicriteria objective function for both MFs. The best convergence of the optimized parameters and the objective function was for ES and GA. The results of RS optimization are close to those of GA and ES. Based on the results of this study, we identified the most preferred algorithms for optimizing these MFs and provide recommendations regarding the application of evolutionary algorithms and the RS method for optimizing structures with modal phenomena.