Computer Aided Design of a Layout of Planar Circuits by Means of Evolutionary Algorithms

Abstract

In this paper we focus on a real-life application of Evolutionary Algorithms (EA) to the design of Radio and Microwave Frequency (RF& M) circuits. The task is to find a shape of a circuit that meets certain electrical and technological requirements. Additionally, manufacturing cost should also influence the evaluation of the design. Several methods, those based on the circuit theory as well as on numerical optimization, work quite effectively for many typical applications. However, with complicated tasks improved hybrid algorithms are needed that combine better topology search (for shape finding) with continuous parameter optimization. In this contribution, we overview Evolutionary Algorithms that have been used for the task, and propose a new approach. These approaches differ in encoding method and genetic operators. We provide experimental verification of the proposed approach, and discuss several example designs yielded by the method.