Systems-of-Systems (SoSs) are evolutionary developed from independent systems to achieve missions through emergent behavior. As concrete systems which will actually participate in an SoS are, in general, not known at design-time, SoSs need to dynamically create coalitions among possible constituent systems at run-time. Due to these uncertainties, the SoS architecture needs to be described abstractly at design-time and then concretely constituted at run-time. An abstract SoS architecture description can thereby specify the correctness properties of coalitions that can be formed among constituent systems at run-time. In particular, it is important to ensure that concrete SoS architectures that may emerge at run-time from such coalitions correctly comply with the abstract SoS architecture description. To automatically check the feasibility of such concrete SoS architectures, in its initial form as well as further evolution, this paper presents an approach that formally expresses the abstract SoS architecture description and the rules governing the organization of SoSs in concrete architectures at run-time. Then, we demonstrate how this approach is used for determining the feasibility of different coalitions in a flood monitoring SoS.