Source identification and correlation between near field-far field tolerances when applying a genetic algorithm

Abstract

Genetic Algorithm is applied to find an optimal electric and magnetic dipole set for a known source. The algorithm reaches good repeatability by manipulating the processes of selection, crossover and mutation. The correlation between fitness values for near field matching and far field prediction is analysed. The evolution of dipole parameters and fitness correlation is illustrated, and the far field pattern prediction is compared with the reference data.