Response Features for Warm-Start Optimization and Rapid Database Acquisition

Tuning of geometry parameters is one of the essential stages of contemporary high-frequency design. It is necessary because available design methods, whether based on theoretical considerations or on engineering experience, are only capable of yielding initial designs that need further adjustment in order to boost the performance parameters as much as possible. Numerical optimization is also imperative for system redesign with respect to different operating conditions (e.g., center frequency, bandwidth) or material parameters (e.g., substrate permittivity/thickness). As the tuning process normally involves full-wave electromagnetic (EM) analysis, it may incur considerable computational expenses. Reducing these costs is highly desirable from the perspective of both academic research and industry. This chapter discusses several surrogate-assisted frameworks for accelerated parameter tuning of high-frequency structures that allows for reusing the previously acquired design data. The major role is played by response features, which are used to identify the operating parameters of the systems under design and to construct forward and inverse surrogate models, later used as reliable predictors for generating reasonable initial designs, as well as further parameter tuning. The involvement of the response feature technology also leads to a more accurate rendition of the system gradients, which speeds up the design refinement as compared to models constructed at the level of original frequency characteristics. The presented methodologies are demonstrated using a number of microwave and antenna components, as well as compared to state-of-the-art parameter tuning procedures.