Nowadays, Navy ship decks are more and more crowded with antennas. Often, in the design stage, brass scale models are used by the Navies of many countries to evaluate radiators impedances and their layout. This approach may suffer from inaccuracies due to scalability and can be quite time consuming and costly. A possible different approach is the one pursued in this work, where accurate modeling of radiators (wires, loaded wires, loaded wires on ground planes, etc.) has been combined with efficient global optimization techniques. As bees in a swarm looking for the best flower, global optimizers could fly in the antenna garden to find the antenna design that best fits the requirements.