An Efficient Optimization Framework for Material and Conductor Designs of Antennas

Design optimization has become an inevitable task to accommodate for stringent performance requirements of wireless applications. Existing techniques are mainly restricted to parametric design of the size or more recently material of the device. However, to allow for global exploration of novel designs, large-scale design optimization problems need to be solved possibly via efficient reanalysis based on electromagnetic analysis tools. Needless to say, reanalysis is very time consuming, hence is the bottleneck of iterative synthesis problems. In this paper, a simple but effective technique based on Bayesian trained rational functions is introduced to lower the number of reanalysis calls to a full wave analysis tool. It allows for robust interpolation of frequency based response data, i.e. return loss. The technique is also implemented on a design example.