A Machine Learning Enhanced Evolutionary Algorithm for Antenna Design

Antenna optimization based on traditional algorithms is always computationally expensive, because it is inevitable to conduct many full-wave simulations during the iterative process. Machine learning (ML) can reduce the computation budget during the optimization, but it requires a large dataset for training. This paper proposes a novel algorithm that combines ML with evolutionary algorithm to achieve efficient antenna synthesis. A synthesis experiment on a wide-band dielectric resonator antenna is conducted, and results show that the proposed algorithm can greatly improve the optimization efficiency compared with traditional optimization methods. The proposed algorithm can also be applied to other types of antenna synthesis problems. This work demonstrates the potential of the ML-enhanced evolutionary algorithm for efficient antenna designs and optimizations.