New optimisation algorithm for planar antenna array synthesis

Deterministic techniques have been extensively employed in antenna array synthesis, using physical parameters of the array to determine an optimal solution. These deterministic algorithms, requiring good initial conditions, can become unstable and require long convergence times. In contrast, optimisation methods using stochastic principles have become popular techniques for solving the synthesis problem. The cross entropy (CE) method is one of those more popular algorithms, such as genetic algorithm (GA) and particle swarm optimisation (PSO), possessing good convergence properties. Then, the CE method is a global stochastic optimisation technique for solving both continuous and discrete systems.This paper describes the planar antenna arrays synthesis using the CE method to produce array responses with minimum peak sidelobe levels (SLL).Good simulation results are obtained with SLL of −22.7 dB for isotropic elements and −32.17 dB for non-isotropic elements. By comparison to more popular algorithms, the CE method can optimise a large number of discrete parameters in a short time but GA requires a long time (large iteration number) to achieve the same performance.