What is it about?
This article shows a way for quantifying and propagating general uncertainty through a black-box computational model, that is not based on assumptions and idealisations, but on expert knowledge and empirical evidence. The approach respects the fundamental differences between randomness and imperfect knowledge during the entire quantification effort. A heuristic approach towards quantifying numerical uncertainty is also presented.
Featured Image
Photo by Katja Anokhina on Unsplash
Why is it important?
Uncertainty is present all around us and its careful quantification is among the top priorities for developing reliable computer models to advance digital engineering. The work presented here lays out a systematic effort in reasoning and working with general uncertainties.
Perspectives
The article is a response to an uncertainty quantification (UQ) challenge, the sort of which are of critical importnace in bridging the persisting divide between academic research and practical application. The article itself is a step towards bringing to the attention of the industry and regulators that diligent UQ is neither impossible to perform, nor is it nullifying tradtional deterministic modelling results, when careful consideration is given to the available knowledge and evidence to efficiently integrate UQ into mainstream aerospace engineering.
Petar Hristov
GATE Institute
Read the Original
This page is a summary of: How to Exploit What We Know About Input and Model: A Trans-probabilistic Approach to the 2022 AIAA UQ Challenge, January 2024, American Institute of Aeronautics and Astronautics (AIAA),
DOI: 10.2514/6.2024-0944.
You can read the full text:
Contributors
The following have contributed to this page
