Differential Evolution Optimization Algorithm for Antenna Designs with Mixed Discrete-Continuous Variables

Differential evolution algorithm has been widely used in the optimization of electromagnetic problems due to its simple structure and excellent performance. The optimization of mixed discrete-continuous variables is necessary for most of the electromagnetic problems, but it remains a difficult problem. To maintain the simplicity in structure, an adaptive discrete variable selection approach is proposed for the differential evolution algorithm to optimize the design problems with mixed variables effectively. For demonstration, the adaptive discrete variable selection based on differential evolution algorithm (ADVSDE) was adopted to optimize a stacked wideband dielectric resonator antenna (DRA) design with two discrete variables. The results show that the proposed algorithm can effectively optimize the design problems with mixed variables, which cannot be handled by the existing optimization algorithms applied to electromagnetic problems. Good agreement between simulation and measurement demonstrates the validation of the proposed algorithm.