Response Features for Local Optimization

Response feature technology exploits a particular structure of the system responses and capitalizes on the dependence of characteristic points of the system outputs on the design variables being considerably less nonlinear than of the entire characteristics. Reformulating design optimization tasks in terms of response features allows for a notable reduction of their computational overhead and improves their reliability. In order to employ the response feature technology, the designer needs to tailor it to a design task at hand. In particular, this requires a cautious selection of the characteristic points of the system response (or responses) so that they reflect the design specifications. A proper definition of response features necessitates a thorough scrutiny of the outputs of a structure under design. In high-frequency electronics, the designers encounter a large variety of the types and shapes of system responses, e.g., narrow-band, multi-band, and broadband. The main goal of this chapter is to present how to select and define response features for structures of different kinds of outputs and for various design problems. The employment of the response feature technology for solving design optimization tasks necessitates reformulation of the objective function in terms of response features, which is also an important topic considered here. The notion of response features has been introduced in the Chap. 5; hence, the reader should refer to it while analyzing the examples presented herein. Some of the application case studies considered in this chapter utilize certain acceleration mechanisms that will be a main topic of Chap. 7, which covers expedited feature-based optimization. The rationale behind including them here is to present the reader with a large variety of system outputs, and diverse design tasks.