Efficient Online Multi-Objective Antenna Optimization

This article proposed an efficient multi-objective methodology for antenna optimization, comprising two main components: an offline initialization module(OIM) and an online model optimization module(OMOM). The OIM primarily focuses on data sampling and initializing the database. Specifically, in the OMOM, the transfer function model serves as a proxy model to reduce the time for traditional full-wave simulation calculations, and this module integrates multi-objective optimization(MOO) using genetic algorithm. The validity and efficiency of this proposed model are confirmed with a Fabry-Perot resonator antenna example. Numerical results substantiate the superior performance of the proposed methodology.