A Platform for Antenna Optimization with Numerical Electromagnetics Code Incorporated with Genetic Algorithms

Abstract : This thesis investigation presents a unique incorporation of the Method of Moments (MoM) with a Genetic Algorithm (GA). A GA is used in accord with the Numerical Electromagnetics Code, Version 4 (NEC4) to create and optimize typical wire antenna designs, including single elements and arrays. Design parameters for the antenna are defined and encoded into a chromosome composed of a series of numbers. The cost function associated with the specific antenna of interest is what quantifies improvement and, eventually, optimization. This cost function is created and used by the GA to evaluate the performance of a population of antenna designs. The most successful designs of each generation are kept and altered through crossover and mutation. Through the course of generations, convergence upon a best design is attained. The Yagi-Uda and the Log Periodic Dipole Array (LPDA) antennas are the focus of this study. The objectives for each antenna are to maximize the main power gain while minimizing the Voltage Standing Wave Ratio (VSWR) and the antenna's length. Results for the Yagi-Uda exceed previous designs by as much as 40 dB in the main lobe while maintaining respectable length and VSWR values. The improvements made in the LPDA antenna were not as drastic, finding a nominal increase in power gain while truncating original allowance in the length by more than half, along with nominal VSWR values that were close to the ideal value of one. The percentage of bandwidth covered for the frequencies of interest are 8.11% for the Yagi-Uda and 10.7% for the LPDA. GA performance is evaluated and, based on previous results, implemented with real-numbered chromosomes as opposed to the classic binary encoding. This methodology is very robust and is improved upon in this research, all while using a novel approach with an optimization program platform called iSIGHT, developed by Engineous Software.