Abstract:
For numerical decision of many applied tasks we have to work with a large data. To solve this problem in the article it is offered to apply data in TT (Tensor Train) format. It effectively compresses multidimensional arrays. The robust algorithm of construction TT format also is known. In work features of application of approximation of multiindex arrays for dynamic problems are described. It is shown on an example of a three-dimensional separated flow problem. In the article it is demonstrated the features of parameterization of array what assign efficiency algorithm. The efficiency of the algorithm using compressing data in the TT format is shown on an example of a problem of flow three-part canopy parachute system.
Keywords:multidimensional arrays, Tensor train format, incompressible liquid, vortex methods.
Fulltext:
PDF file (559 kB)