Comparison of Different Heuristic Optimization Methods for Near-Field Antenna Measurements

A comparison between different modern heuristic optimization methods applied to antenna far-field radiation pattern reconstruction from planar near-field data is presented in this paper. The antenna under test is represented by means of equivalent magnetic currents (EMC) whose components are optimized using several heuristic algorithms such as simulated annealing (SA), genetic algorithms (GA), and particle swarm optimization (PSO), as well as a traditional local optimization method, the Nelder Mead downhill simplex algorithm. Several schemes for GA (classical real-valued and binary encoding, and their hybrid versions) and PSO (global or local topologies with synchronous or asynchronous updates of the swarm) have been considered in the analysis and the pros and cons of each one are reported and discussed. A study of the performance and limitations of each algorithm using as a canonical problem a small size antenna aperture, along with results of near-field to far-field (NF-FF) transformation are also included.