Fast automatic differentiation in problems variations four-dimensional data assimilation (4Dvar)статья
Информация о цитировании статьи получена из
Scopus
Статья опубликована в журнале из списка Web of Science и/или Scopus
Дата последнего поиска статьи во внешних источниках: 25 апреля 2019 г.
Аннотация:The experience of using the technology of Fast Automatic Differentiation (FAD) is
described to restore the initial data of a model hydrodynamic flow with a free boundary based
on the results of spatial-temporal monitoring of velocity, density and free surface components.
The solution of the direct problem is realized according to the CABARET scheme, for which
the construction of the conjugate system is a serious independent task. Currently, there are
available open computer codes that allow the implementation of FAD technology in the "black
box" mode, submitting to its input the executable "author's" code for solving a direct problem.
The model problem was solved with the help of the Adept 1.1 package, while using the 3-pass
algorithm, which reduces the requirements for the amount of RAM.