Daria Sushnikova

Research Interests

Broadly speaking, my area of interest is fast linear algebra algorithms. My research mainly focuses on rank-structured matrices such as the FMM-like, HSS, H, HODLR matrices, et al. I design, implement, and apply algorithms to perform fast computations with such matrices.

Selected Publications

“FMM-LU: A fast direct solver for multiscale boundary integral equations in
three dimensions,” Sushnikova, D., Greengard, L., O’Neil, M., & Rachh, M.,arXiv preprint arXiv:2201.07325, 2022.

“Simple non-extensive sparsification of the hierarchical matrices,” Sushnikova, D., Oseledets, I., Adv Comput Math, T. 46. C. 52, 2020.

“Compress and eliminate solver for symmetric positive definite sparse matrices,” Sushnikova, D., Oseledets, I., SIAM J. Sci. Comput., T. 40, No. 3, C.A1742-A1762, 2018.

“Application of block low-rank matrices in Gaussian processes for regression,” Sushnikova, D., Computational Methods and Programing, No. 18, P. 214-220 (in Russian), 2017.

“Preconditioners for hierarchical matrices based on their extended sparse form,” Sushnikova, D., Oseledets, I., Russ. J. Numer. Anal. Math. Modelling, No.
1, P. 29-40, 2016.

“Numerical solution of diffraction problems using large matrix compression,” Ryzhakov, G., Mikhalev, A., Sushnikova, D., Oseledets, I., 9th European Conference on Antennas and Propagation (EuCAP), P. 1-3, 2015

Research Interests Keywords

Rank-structured matrices FMM hierarchical block-low rank matrices direct solvers and preconditioners