Geodesic
Quasi-Monte Carlo Methods for some Linear Algebra Problems. Convergence and Complexity
Serdica Journal of Computing, Tome 4 (2010) no. 1, pp. 57-72

Voir la notice de l'article provenant de la source Bulgarian Digital Mathematics Library

We present quasi-Monte Carlo analogs of Monte Carlo methods for some linear algebra problems: solving systems of linear equations, computing extreme eigenvalues, and matrix inversion. Reformulating the problems as solving integral equations with a special kernels and domains permits us to analyze the quasi-Monte Carlo methods with bounds from numerical integration. Standard Monte Carlo methods for integration provide a convergence rate of O(N^(−1/2)) using N samples. Quasi-Monte Carlo methods use quasirandom sequences with the resulting convergence rate for numerical integration as good as O((logN)^k)N^(−1)). We have shown theoretically and through numerical tests that the use of quasirandom sequences improves both the magnitude of the error and the convergence rate of the considered Monte Carlo methods. We also analyze the complexity of considered quasi-Monte Carlo algorithms and compare them to the complexity of the analogous Monte Carlo and deterministic algorithms.
Keywords: Quasi-Monte Carlo Methods, Matrix Computations, Markov Chains, Quasirandom Sequences 
@article{SJC_2010_4_1_a5,
     author = {Karaivanova, Aneta},
     title = {Quasi-Monte {Carlo} {Methods} for some {Linear} {Algebra} {Problems.} {Convergence} and {Complexity}},
     journal = {Serdica Journal of Computing},
     pages = {57--72},
     publisher = {mathdoc},
     volume = {4},
     number = {1},
     year = {2010},
     language = {en},
     url = {http://geodesic.mathdoc.fr/item/SJC_2010_4_1_a5/}
}
TY  - JOUR
AU  - Karaivanova, Aneta
TI  - Quasi-Monte Carlo Methods for some Linear Algebra Problems. Convergence and Complexity
JO  - Serdica Journal of Computing
PY  - 2010
SP  - 57
EP  - 72
VL  - 4
IS  - 1
PB  - mathdoc
UR  - http://geodesic.mathdoc.fr/item/SJC_2010_4_1_a5/
LA  - en
ID  - SJC_2010_4_1_a5
ER  - 
%0 Journal Article
%A Karaivanova, Aneta
%T Quasi-Monte Carlo Methods for some Linear Algebra Problems. Convergence and Complexity
%J Serdica Journal of Computing
%D 2010
%P 57-72
%V 4
%N 1
%I mathdoc
%U http://geodesic.mathdoc.fr/item/SJC_2010_4_1_a5/
%G en
%F SJC_2010_4_1_a5
Karaivanova, Aneta. Quasi-Monte Carlo Methods for some Linear Algebra Problems. Convergence and Complexity. Serdica Journal of Computing, Tome 4 (2010) no. 1, pp. 57-72. http://geodesic.mathdoc.fr/item/SJC_2010_4_1_a5/