MBSE in Architecture Design Space Exploration

Abstract

Architecting decisions play a large role in the final system performance. However, the architecting process suffers from a combinatorial explosion of alternatives, often requiring judgment based on expert experience to reduce the design space size. Systematic design of complex system architectures enables a more thorough exploration of the complete design space and less exposure to expert bias, subjectivity, and conservatism. Two aspects of system architecture design are treated in this chapter: the automated generation of architecture candidates by capturing architecting decisions and quantitative evaluation strategies. An overview of different design space modeling techniques and their applicability to design space exploration is given at first. A graph-based method that models architecture decisions related to function assignment, component characterization, and component structure is presented in more depth, along with how a formalized optimization problem can be constructed from this design space model. Thereafter, the main features of architecture evaluation are discussed. Applications of the introduced method to the Apollo mission architecting case and a hybrid-electric propulsion system architecture problem are presented.