ANN and space mapping for microwave modelling and optimization

Abstract

Artificial neural network (ANN) and space mapping are recognized as two major recent advances in microwave CAD. ANNs can be trained to learn EM and physics behaviour from component data, and trained ANNs can be used in high-level circuit design. Space mapping has proved to be a breakthrough in engineering optimization allowing expensive EM optimization to be performed effectively with the help of “coarse” or surrogate models. Recent advance also led to neuro-space mapping, combining the advantages of ANN and space mapping for efficient modelling of microwave components. This paper presents an overview of the state-of-art of microwave modelling and design with ANN, space mapping and neuro-space mapping.