An Efficient Scattering Characteristics Similarity-Based Tuning-Driven Optimization Method for Microwave Filters

In this article, a scattering characteristics similarity-based tuning-driven optimization (SCSTDO) method is proposed for microwave filter tuning applications. The most critical part of the proposed SCSTDO method is constructing a reliable tuning-driven optimization model-matching library (TOML). The TOML is constructed by the tuning elements’ position matrices (TEPMs) and the scattering characteristics of the optimization model samples. In order to obtain high-representative samples and achieve low-cost full-wave electromagnetic (EM) simulation, the TEPMs of the optimization model samples are generated by a Gaussian-like distribution sampling (GLDS) method. The corresponding scattering characteristics are obtained by using full-wave EM simulators. In the SCSTDO method, the detuned filter is matched with the sample with maximum similarity in the TOML according to the similarity of scattering characteristics, and it is tuned using the sample with maximum similarity as a guide. The proposed method is experimentally verified through two examples, including a second-order cavity filter with five tuning elements and a third-order cavity filter with seven tuning elements. The results show that the SCSTDO method can achieve efficient and fast optimization.