Performance Comparison of Optimization Methods for Flat-Top Sector Beamforming in a Cellular Network

Abstract

The flat-top radiation pattern is necessary to form an appropriate beam in a sectored cellular network and to provide users with best quality services. The flat-top pattern offers sufficient power and allows to minimize spillover of signal to adjacent sectors. The flat-top sector beam pattern is relied upon in sectored cellular networks, in multiple-input multiple-output (MIMO) systems and ensures a nearly constant gain in the desired cellular sector. This paper presents a comparison of such optimization techniques as real-coded genetic algorithm (RGA) and particle swarm optimization (PSO), used in cellular networks in order to achieve optimum flat-top sector patterns. The individual parameters of flat-top sector beams, such as cellular coverage, ripples in the flat-top beam, spillover of radiation to the adjacent sectors and side lobe level (SLL) are investigated through optimization performed for 40◦ and 60◦ sectors. These parameters are used to compare the performance of the optimized RGA and PSO algorithms. Overall, PSO outperforms the RGA algorithm.