A multi-objective approach is applied to the design of linear arrays of antennas. The adopted procedure is based on a standard implementation of the Controlled Elitist Non-Dominated Sorting Genetic Algorithm-II. Non-uniform and symmetrical broadside arrays are considered, with both uniform and non-uniform spacing. The work focuses on the problem of optimizing simultaneously the side lobe level and the main beam width, investigating how and in which cases solutions obtained by the genetic algorithm can approximate, or outperform, the optimal Pareto front for the uniform spacing case computable by the Dolph-Chebyshev design method.
