View article

Evolutionary algorithms are powerful optimization and design tools, and the design of fitness function is their crucial phase. The ruggedness of a fitness function landscape has a profound effect on optimization speed. The choice of objectives and penalty coefficients are important problems for evolutionary antenna design, because these parameters influence the ruggedness of the fitness function landscape. In this paper, we analyze several fitness function landscapes for evolutionary design of dipole antennas using the method of basin evaluation. Our approach makes it possible to compare different fitness functions and thus define proper fitness functions for evolutionary antenna design.