A multi-agent differential evolution for linear array synthesis

Abstract

This paper describes a multi-agent differential evolution (MADE) for optimizing linear arrays synthesis. In order to find better solution, each individual of MADE as a agent compete or cooperate with their neighbors, then perform crossover, mutation and selection to diffused global knowledge. And it is used for optimization to reduce the peak side lobe level (PSLL) with minimum element spicing constraints, through dynamic computing lower bound and upper bound, constraints can be handled. Contrast with other result, MADE has greater efficiency and robustness. © 2010 IEEE.