Geodesic
Empirical improvement of elementary gradient methods
Matematičeskoe modelirovanie, Tome 17 (2005) no. 6, pp. 43-57.

Voir la notice de l'article provenant de la source Math-Net.Ru

Very simple modified algorithms of quickest descent and minimal discrepancies are offered. These methods allow essentially increasing the convergence speed. The method is universal and does not require adjustment on a spectrum of a matrix. The efficiency of offered methods is confirmed by numerous computing experiments. 
@article{MM_2005_17_6_a4,
     author = {E. A. Alshina and A. A. Boltnev and O. A. Kacher},
     title = {Empirical improvement of elementary gradient methods},
     journal = {Matemati\v{c}eskoe modelirovanie},
     pages = {43--57},
     publisher = {mathdoc},
     volume = {17},
     number = {6},
     year = {2005},
     language = {ru},
     url = {http://geodesic.mathdoc.fr/item/MM_2005_17_6_a4/}
}
TY  - JOUR
AU  - E. A. Alshina
AU  - A. A. Boltnev
AU  - O. A. Kacher
TI  - Empirical improvement of elementary gradient methods
JO  - Matematičeskoe modelirovanie
PY  - 2005
SP  - 43
EP  - 57
VL  - 17
IS  - 6
PB  - mathdoc
UR  - http://geodesic.mathdoc.fr/item/MM_2005_17_6_a4/
LA  - ru
ID  - MM_2005_17_6_a4
ER  - 
%0 Journal Article
%A E. A. Alshina
%A A. A. Boltnev
%A O. A. Kacher
%T Empirical improvement of elementary gradient methods
%J Matematičeskoe modelirovanie
%D 2005
%P 43-57
%V 17
%N 6
%I mathdoc
%U http://geodesic.mathdoc.fr/item/MM_2005_17_6_a4/
%G ru
%F MM_2005_17_6_a4
E. A. Alshina; A. A. Boltnev; O. A. Kacher. Empirical improvement of elementary gradient methods. Matematičeskoe modelirovanie, Tome 17 (2005) no. 6, pp. 43-57. http://geodesic.mathdoc.fr/item/MM_2005_17_6_a4/