Optimizing groups of colluding strong attackers in mobile urban communication networks with evolutionary algorithms