A Rapid Optimization Scheme for the Simultaneous Optimization of S-Parameters and Radiation Patterns

摘要

For solving multiobjective electromagnetic (EM) problems, heuristic algorithms and surrogate-assisted models are widely utilized. However, most of the existing optimization schemes are not efficient and comprehensive when optimizing the S-parameters and radiation patterns of an antenna. This is caused by the high computational cost of EM simulations and the lack of universal antenna parametric models. To address this problem, a rapid optimization scheme based on surrogate-assisted parametric models (SAPMs) is proposed. Different from traditional surrogate-assisted models, the proposed model is utilized for mapping from design parameters to characteristic coefficients of the parametric models. In this way, the EM responses with regard to different design parameters can be predicted. The proposed scheme is implemented by a model-search-valid iteration and further validated on a patch antenna, an SIW cavity-backed slot antenna, and a flat-topped Fabry-Perot cavity antenna. The validation results indicate that S-parameters and radiation patterns of an antenna can be simultaneously optimized with high efficiency. Compared with relevant works, the optimization efficiency and solution quality of the proposed scheme show a marked promotion, manifesting as the total time cost is significantly reduced to less than 30% and even only 6% of the time required in the state-of-the-art work.

关键词