Evidence-Based Resilience Optimisation with Evidence Network Models: a Space Systems Example
The presentation focuses on the concept of Resilience Engineering as an extension of both Robustness and Reliability Based Design Optimisation. It will then introduce a way to model complex engineering systems made of a number of interconnected components whose performance and reliability are affected by some form of epistemic uncertainty. This model, called Evidence Network Model, uses belief functions to quantify the epistemic uncertainty in some of the parameters defining each component. The robustness and reliability of the system are then computed with a decomposition technique that exploits the relationship among uncertain quantities and their relative importance. This approach is applied to an example of space systems engineering as part of a recent project (Uncertainty Treatment and Optimisation in Aerospace Engineering) with ESTECO, supported by the European Space Agency.