A novel hybrid EP-GA method for efficient electromagnetics optimization

We propose a simple and efficient means of combining evolutionary programming (EP) and the genetic algorithm (GA) for a more efficient evolutionary optimization in electromagnetics. The resulting hybrid algorithm takes advantage of the convergence behavior of the GA and EP during different stages of evolution to efficiently march toward the global solution. We have applied the hybrid method to several multi-modal test functions as well as to various unconstrained and constrained antenna and microwave optimization problems. We have found that the proposed method works very well for a given problem, if EP and the GA perform well for that problem individually. Examples presented in this paper include optimization of a twenty dimensional Auckley function and the gain optimization of a six-element Yagi array. Other examples involving antenna arrays and frequency selective surfaces are also given.