Engineering solutions to surveillance operations in wide areas are based on the integration of manned aircrafts and Unmanned Aerial Vehicles in complex System-of-Systems (SoS) architectures. The selection of one specific architecture depends on a very large set of parameters and constraints related to the surveillance scenario, target statistics and the availability of limited resources such as electro-magnetic spectrum, crew, airfields or simply budgetary restrictions. This paper covers the use of Genetic Algorithms (GA) to obtain the best SoS architecture for a given surveillance scenario and a given set of limited resources.