Finding improved wire-antenna geometries with genetic algorithms

A remote monitoring system's performance is measured largely by its associated antennas' efficacy. Improving antenna performance is not a straightforward process, as many (sometimes conflicting) design trade-offs must be considered. An alternative approach specifies desired antenna properties and then searches for antenna geometries producing those properties. We implement this approach of designing a wire antenna having optimal performance under user-defined adverse conditions by using a genetic algorithm, where fitness of promising designs is analysed using the Numerical Electromagnetics Code (NEC) Version 4.1. Two versions of an integrated GA-NEC code are implemented; both produce geometries showing improvement over current designs. Results for both versions are presented, compared, and discussed.