A space mapping method allowing models with different parameter rank and physical meanings for coarse and fine model

Abstract

-Automated optimization of complex microwave circuits is still a critical task in nowadays. This contribution demonstrates two new approaches in space mapping technology. One substantial benefit is the improvement of calculation efficiency of the TRASM algorithm by enhancing the multiparameter extraction step. Furthermore, the possibility of mapping coarse and fine models of different degrees of freedom, i.e. with different parameter rank, and unequal physical parameter meanings has been introduced to the method. Therefore, arbitrary microwave circuits of high complexness with well-matching coarse models can be efficiently and automatically optimized.

Quasi-lumped suspended stripline filters and diplexers

“…The procedure may be separated into two or three successive steps with different computational accuracy, leading to a considerable reduction of optimization time. Recently, a space-mapping approach was made using the lumped-element circuit for the coarse model and the geometrical structure for the fine model [12], together with a special optimizer applied to the coarse model to find global optima for the involved circuits.…”

Reliable design optimization of microwave structures using multipoint-response-correction space mapping and trust regions

“…Efficient EM‐based optimization of CPU‐intensive structures can be realized using surrogate models [4, 5]. One of the most successful approaches of this kind in microwave engineering is space mapping (SM) [6–10]. SM replaces the optimization of an expensive or “fine” model by means of the iterative optimization and updating of so‐called “coarse” model, which is less accurate but cheaper to evaluate.…”