В.Д. Сулимов, П.М. Шкапов, А.В. Сулимов
64
ISSN 1812-3368. Вестник МГТУ им. Н.Э. Баумана. Сер. Естественные науки. 2016. № 5
OPTIMIZATION OF SINGULAR VALUES OF PARAMETER
DEPENDENT MATRICES USING HYBRID ALGORITHMS
V.D. Sulimov
1
P.M. Shkapov
1
spm@bmstu.ruA.V. Sulimov
2
1
Bauman Moscow State Technical University, Moscow, Russian Federation
2
Lomonosov Moscow State University, Branch, Sevastopol,
Russian Federation
Abstract
Keywords
This article considers extremal problems for singular spec-
trum components of real-valued matrices depending on
parameters. Criterion functions are supposed to be conti-
nuous, Lipschitzian, multiextremal and not necessarily dif-
ferentiable. While searching for global solutions we used new
hybrid algorithms that combine a stochastic algorithm for
scanning a search space and deterministic methods for local
search. The first hybrid algorithm determines local solutions
using the linearization method with smoothing approxima-
tions. The second algorithm does it by using the modified
space-filling curve method. Our work provides numerical
examples
Singular value, criterion function,
Lipschitzian constant, smoothing
approximation, Peano curve,
global optimization, Metropolis
algorithm, hybrid algorithm
REFERENCES
[1] Nicoud F., Toda H.B., Cabrit O., Bose S., Lee J
.
Using singular values to build a subgrid-
scale model for large eddy simulations.
Physics of Fluids
, 2011, vol. 23, no. 8, pp. 085106-1–
085106-12.
[2] Vaidya P.G., Anand S.P.S., Nagaraj N. A nonlinear generalization of singular value decom-
position and its applications to mathematical modelling and chaotic cryptanalysis.
Acta Appli-
candae Mathematicae
, 2010, vol. 112, no. 2, pp. 205–221.
[3] Danforth C.M., Kalnay E. Using singular value decomposition to parametrize state-
dependent model errors.
Journal of the Atmospheric Sciences
, 2008, vol. 65, no. 4, pp. 1467–
1478.
[4] Dax A. From eigenvalues to singular values: a review.
Advances in Pure Mathematics
, 2013,
vol. 3, pp. 8–24.
[5] Dieci L., Elia C
.
The singular value decomposition to approximate spectra of dynamical
systems. Theoretical aspects.
Journal of Differential Equations
, 2006, vol. 230, no. 2,
pp. 502–531.
[6] Gu M. Subspace iteration randomization and singular value problems.
SIAM Journal
on Scientific Computing
, 2015, vol. 37, no. 3, pp. A1139–A1173.
[7] Polyakova A
.
Reconstruction of potential part of 3D vector field by using singular value
decomposition.
Journal of Physics: Conference Series
, 2013, vol. 410, p. 012015.
DOI: 10.1088/1742-6596/410/1/012015