Method of Antenna Optimization Based on Gaussian Regression and Genetic Algorithm

In this paper, the gaussian regression algorithm is used to establish the mapping relationship between the physical model and the antenna electrical parameters, and the genetic algorithm (GA) is used to find the optimal physical parameters that meet the target value. Using these two algorithms, the design of the dual-band antenna is completed. Two physical parameters are used for optimization and 10 antenna samples are used to train the gaussian regression model. The gaussian regression model takes only 34 ms to calculate once, while the full-wave electromagnetic simulation takes 82 s to calculate once. The GA takes only 2 s to calculate once. This greatly reduces the time required for calculations. The two resonant points of the dual-band antenna are 1.72 GHz and 2.54 GHz, respectively. The genetic algorithm completes the optimal design of the antenna after only 10 iterations. This greatly improves the efficiency of antenna optimization.