A Bayesian Framework for Uncertainty Quantication in the Design of Complex Systems

One of the main challenges of current system design practices is the inability to recognize performance, cost, and schedule risks as they emerge. This paper presents a Bayesian framework for the design of complex systems, in which uncertainty in various parameters and quantities of interest is characterized probabilistically, and updated through successive design iterations as new estimates become available. Incorporated in the proposed model are methods to quantify system complexity and risk, and reduce them through the allocation of resources for redesign and re nement. This approach enables the rigorous quanti cation and management of uncertainty, thereby serving to help mitigate technical and programmatic risk. The Bayesian system design framework is demonstrated on the notional design of a hybrid infantry ghting vehicle for military applications.