| Summary: | The main purpose of this paper is to present a novel procedure for accelerating a multi-objective optimization method of designing a 5G antenna. The optimization method was chosen after comparing four learning optimization algorithms. The Kriging algorithm was found to be superior to the Artificial Neural Network (ANN), Support Vector Machine (SVM), and Rational algorithms. Our methodology is creatively correlated to exploit some cost functions of height, the Dielectric constant of the substrate, and meta-material design variables, with a view to reducing the return loss and increasing the gain in learning from the Kriging model builder techniques. This was fully achieved in the present study by comparing the results of analyzing and optimizing two effective fundamental characteristics of the antenna with EM simulation software and prototype antenna measurements.
|
|---|