Multiobjective Integrated Optimization Design of Lens Antenna Based on a Competition and Cooperation Model With Neural Network and Evolutionary Algorithm

A competition and cooperation model algorithm (CCMA) is proposed and it was used to design a multiperformance lens antenna. The algorithm makes the genetic algorithm (GA) and the inverse neural network (INN) to compete and cooperate with each other. The cooperation section of the CCMA achieves highly efficient utilization of the full wave simulation data to advance the optimization process. While the compete section of the CCMA enables high-efficiency optimization across different antenna models. A multiobjective dielectric lens antenna is designed by the CCMA, which obtains good results. The aperture efficiency is greater than 62.1% in 66 GHz to 76 GHz (peak efficiency of 69.3% at 70 GHz), the side lobe level at the E- and H-planes are lower than −16 dB, the 1 dB gain bandwidth is 14.1%, and the peak gain is 28 dbi.