Development of a decision support framework for systems architecting in aerospace applications

Abstract

The exploration of the architectural solution space tends to be iterative, where multiple system architectures are explored over several cycles before a final solution is selected. However the time and cost required to conduct this activity can be significant in the presence of multiple system architectures. This thesis presents a decision support framework that assists the system architect in generating, analysing, and identifying optimal system architectures. The framework achieves this by using a formal modelling approach that represents the architectural decision-making process as a Constraint Optimisation Problem (COP), resulting in a graph representation of interconnected architectural decision variables. The graph-based approach provides computational tools that enables the system architect to automatically synthesise viable architectures based on the constraints defined, and calculate high-impact decision variables within the network. This capability is enabled by synthesising concepts from decision theory, multi-objective optimisation, and centrality measures from network analysis to provide a visual representation of high-impact decision variables. In applying this framework to the design of a low-cost Unmanned Aircraft System (UAS), we identify the choice of design alternatives relating to the implementation of the payload sensor system to have a high-impact on system properties and network connectivity. Suggesting that the exploration of the solution space should be focused towards payload implementation.

More information

Identifiers

Catalogue record

Export record