The significance of genetic representation in genetic antenna design

Abstract

Genetic algorithms have been shown to be effective in the design of small antennas. However, their effectiveness depends on both the mechanisms of the genetic algorithm and the genetic representation of the problem. We present three genetic chromosome representations for a small antenna design problem (lowest VSWR for a specified h//spl lambda/ cube size) and compare the results of the three optimization processes. Our work shows that different chromosome representations lead to solutions in different, overlapping subclasses, and that an important consideration in genetic antenna design is finding a simple chromosome representation which is capable of representing the subclass containing the "optimal" solution. Our initial results show that the chromosome design itself is an important factor in successful genetic antenna design.