System-of-Systems Architecture Selection: A Survey of Issues, Methods, and Opportunities

System-of-systems (SoSs) architecture concerns the structure of components, their relationships, and their principles and guidelines governing their design and evolution over time. Well-designed architecture is key to manage the SoS complexity, and thus plays a considerable role in the SoS development and evolution. Architecture selection that aims to identify the desired architectures from a large alternative design space is an initial but critical step toward successful architecture design. Decision-makers need scientific methods to identify those better, more informed but perhaps unexpected architecture solutions. However, the unique SoS characteristics such as interdependency and autonomy bring significant challenges to the architecture selection studies. This article identifies seven critical perspectives in SoS architecture selection, including the conceptual framework, evaluation criteria, interdependency, uncertainty, autonomy, dynamic evolution, and computational methods. The available methods, tools, and processes in the SoS and associated domains are reviewed thoroughly regarding each perspective. This review provides a comprehensive understanding of the current research status of SoS architecture selection. It builds a holistic picture that brings the fragmented pieces of the puzzle together. Towards the end, this article presents a series of research directions for future SoS architecture selection studies.