Comparative Analysis of Numerical Methods for Parameter Determination
Abstract
We made a comparative analysis of numerical methods for multidimensional optimization. The main parameter is a number of computations of the test function to reach necessary accuracy, as it is computationally "slow". For complex functions, analytic differentiation by many parameters can cause problems associated with a significant complication of the program and thus slowing its operation. For comparison, we used the methods: "brute force" (or minimization on a regular grid), Monte Carlo, steepest descent, conjugate gradients, Brent's method (golden section search), parabolic interpolation etc. The Monte-Carlo method was applied to the eclipsing binary system AM Leo.
Cite
@article{arxiv.1310.1967,
title = {Comparative Analysis of Numerical Methods for Parameter Determination},
author = {Ivan L. Andronov and Maria G. Tkachenko},
journal= {arXiv preprint arXiv:1310.1967},
year = {2013}
}
Comments
8 pages, 6 figures, to be published in: Cz{\ke}stochowski Kalendarz Astronomiczny 2014, ed. Bogdan Wszo{\l}ek