Effect of Approximations of the Discrete Adjoint on Gradient-Based Optimization

135 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Improving variable-fidelity surrogate modeling via gradient-enhanced kriging and a generalized hybrid bridge function (2013, unknown): “[18] [19] [20] [33] [34] [37]”
Numerical sensitivity analysis for aerodynamic optimization: A survey of approaches (2010, background): “After frozen eddy-viscosity, perhaps the second most common approximation is that of neglecting the derivatives of artificial diffusion coefficients in convective terms when using a centered convective flux approach [98,43,85], which has almost no effect on the gradient while allowing a considerable”
Review and Unification of Methods for Computing Derivatives of Multidisciplinary Computational Models (2013, background): “the continuous approach in this paper, but ample literature can be found on the subject [3, 27, 37, 38, 50, 87], including discussions comparing the two approaches [23, 65, 72].”
Effective adjoint approaches for computational fluid dynamics (2019, methodology): “Furthermore, by using the NK method to solve the primal flow equations, the GMRES method provides a similar duality preserving behavior as the fixed-point iteration approach because the eigenvalues for Jacobians (primal) and transposed-Jacobians (adjoint) are the same, which ensures consistent conve”