Antenna Design by an Adaptive Variable Differential Artificial Bee Colony Algorithm

Optimal design of antennas is very useful to attain narrow beam and low sidelobe, while keeping a low manufacturing cost. Artificial bee colony (ABC) algorithm has shown good performance in solving electromagnetic inverse problems. However, the ABC algorithm costs too much number of iterations and converges slowly in the antenna design. This paper proposes an adaptive variable differential method for ABC (AVDABC). This method adaptively decides the number of decision variables to be mutated. Existing modified ABC algorithms employ the same search equation in one iteration. The proposed adaptive variable method assigns different search equations to chosen decision variables in one iteration. The proposed AVDABC algorithm is more diverse than the existing ABC algorithms. Through numerical simulation on mathematical functions, AVDABC converges faster than three compared algorithms. Moreover, demonstrated by the design of sparse nonuniform antenna array and Yagi-Uda antenna, the AVDABC algorithm achieves promising performance. Thus, the proposed algorithm is helpful to solve antenna array design problems.