A design optimization methodology for multiband stochastic antennas

A versatile design optimization approach for miniature multiband loaded fractal dipole antennas has been developed in Werner et al. (2001). This technique combines the robustness of genetic algorithms (GAs) (Haupt and Haupt 1997; Rahmat-Samii and Michielssen 1999) with the flexibility of iterated function systems (IFSs) (Peitgen et al. 1992) to evolve optimal self-affine fractal antenna shapes as well as determine the load locations and corresponding component values required to achieve the specified multiband operation. In this paper, an alternative design methodology for achieving the same objectives is introduced. This new methodology employs a GA to evolve a class of multiband antennas, called stochastic antennas, which offer optimal performance characteristics. This method is more general than the approach outlined in Werner et al. since it is not restricted to fractal geometries and there are no loads required to achieve the desired multiband performance. The only disadvantage of the new method is the fact that the optimization procedure is not as efficient as the GA-IFS technique. Several examples of optimized stochastic antennas are presented and discussed.