1995 | OriginalPaper | Buchkapitel
Automatic Differentiation for Gradients, Jacobians, and Hessians
verfasst von : Prof. Dr. Ulrich Kulisch, Dr. Rolf Hammer, Dr. Matthias Hocks, Dr. Dietmar Ratz
Erschienen in: C++ Toolbox for Verified Computing I
Verlag: Springer Berlin Heidelberg
Enthalten in: Professional Book Archive
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
In Chapter 5, we considered automatic differentiation in one variable, but there are also many applications of numerical and scientific computing where it is necessary to compute derivatives of multi-dimensional functions. In this chapter, we extend the concept of automatic differentiation to the multi-dimensional case as given by Rail [72] and many others. We apply well-known differentiation rules for gradients, Jacobians, or Hessians with the computation of numerical values, combining the advantages of symbolic and numerical differentiation. Only the algorithm or formula for the function is required. No explicit formulas for the gradient, Jacobian, or Hessian have to be derived and coded.