Solving some array synthesis problems by means of an effective hybrid approach

Most global optimization based synthesis procedures act in essentially the same way on all the involved variables. In this paper, it is theoretically discussed and numerically emphasized how explicit exploitation of the convexity of the problem with respect to a part of the degrees of freedom (when available) allows to achieve increased performances with respect to previous results, and to develop new interesting design solutions. Examples are presented for the synthesis of both sum and difference patterns by using nonuniformly spaced arrays and for the so called "optimal compromise amongst sum and difference patterns." Finally, the proposed approach is exploited in developing effective design solutions (based on the concept of interleaved arrays) to the problem of achieving very flexible antennas while using simple feeding networks. As an example, array antennas capable to radiate two different beams with individually steerable patterns, or to perform jammer rejection at the physical layer or even to realize a "flat-top" beam pattern are synthesized and presented.