A parallel implementation of seeker optimization algorithm for designing circular and concentric circular antenna arrays

• We propose a parallel version of seeker optimization algorithm (SOA) for designing circular and concentric circular antenna arrays with the low sidelobe levels at a fixed beamwidth (BW).
• We locate the subpopulations of SOA are physically in different computers named as the slaves to improve the algorithm performance.
• We realize the parallelization with MPICH2 and employ 40 PCs (slaves) and 1 laptop (master) for the simulations.
• From maximum sidelobe levels and BW points of view, the performances of the patterns obtained by parallel SOA are very good under the constraint of physical size of antenna and dynamic range ratio.

In this paper, a parallel version of seeker optimization algorithm (SOA) is proposed for designing circular and concentric circular antenna arrays with the low sidelobe levels at a fixed beamwidth. The SOA is a relatively new evolutionary optimization algorithm based on the concept of simulating the act of humans’ intelligent search with their memory, experience, and uncertainty reasoning. In this work, The SOA has been parallelized by benefiting from its dividable population form. The numerical results show that the design of circular and concentric circular antenna arrays using the parallel SOA provides good sidelobe levels with a fixed beamwidth. The quality of results obtained by the parallel SOA is checked by comparing with those of several evolutionary algorithms in the literature.

Figure optionsDownload as PowerPoint slide