Using a real chromosome in a genetic algorithm for wire antenna optimization

A genetic algorithm (GA) is composed of several elements: the chromosome representation of a design, the mating and mutation methods that generate new designs from existing populations, the cost function, selection and replacement schemes, and runtime parameters like population size. The first two of these elements, which determine in large part the effectiveness of the algorithm, are the main subject of this paper. The author applied the mating and mutation methods to wire antenna problems and shows how it has performed compared to traditional binary GAs for two very different wire antenna optimizations. The first example involves the Yagi antenna, which consists of a series of parallel wires, one of which is excited by the driving source. The design requires a value for every spacing and the length of every wire. The second example involves what has been called the crooked-wire genetic antenna. This antenna consists of seven wires connected in series over a ground plane, with endpoints designated by the GA.