Fuzzy AD

Description

Fuzzy arithmetic on fuzzy numbers can be used to introduce a form of uncertainty quantification into numerical computation. We compare different ways of quantifying sensitivities for fuzzy based implementations especially using adjoint methods. We also compare how a fuzzy based uncertainty quantification relates to probabilistic approaches such as Monte Carlo and first order moments estimation using Taylor expansion

• Documentation and first results
• Source code

Contact

Jonathan Hüser (email)