Open AccessFast multi‐objective optimisation of pencil beam reflector antenna radiation pattern responses using Kriging
Author(s) -
Villiers Dirk I.L.,
Koziel Slawomir M.
Publication year - 2018
Publication title -
iet microwaves, antennas and propagation
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.555
H-Index - 69
eISSN - 1751-8733
pISSN - 1751-8725
DOI - 10.1049/iet-map.2017.0090
Subject(s) - kriging , reflector (photography) , pencil (optics) , periscope antenna , radiation pattern , antenna (radio) , optics , radiation , beam (structure) , computer science , physics , telecommunications , light source , machine learning
A Kriging interpolation‐based method for the rapid multi‐objective optimisation (MO) of the radiation patterns of pencil beam reflector antenna systems is presented. MO provides a designer with quantitative trade‐off information about conflicting design goals, but typically requires a large number of time‐consuming function evaluations. The smoothly varying response surfaces of pencil beam reflector systems (gain, receiving sensitivity, sidelobe levels, etc.), as functions of the design parameters, are well suited to an interpolation‐based response surface approximation optimisation strategy. By constructing Kriging interpolants of all the design goals and performing the MO on these fast models the optimisation can be significantly accelerated. This study presents an optimisation strategy where an iterative method is used to improve the accuracy of the response surface interpolants along the perceived Pareto front. Several examples are provided to illustrate the algorithmic flow, as well as confirm the accuracy and acceleration of the optimisation strategy achievable for typical reflector problems.