Miniaturization of Stochastic Linear Phased Arrays via Orthogonal Design Initialization and a Hybrid Particle Swarm Optimizer

Abstract

This paper introduces a powerful technique that uses a hybrid PSO for the design optimization of aperiodic linear arrays of tightly packed miniature stochastic dipole elements. The self impedance of the stochastic dipole elements can be widely varied by changing their shape. A hybrid PSO is used to determine the optimal shape of each stochastic dipole array element as well as to perturb the spacing between these elements in order to control the mutual coupling. This method is capable of evolving compact array configurations with stochastic dipole elements that have well-behaved driving point impedances and low side lobe levels over a prescribed scan range. Orthogonal design with quantization (OD/Q) is used for this real-valued optimization problem to generate relatively fit particles at the initialization stage