Vector Monte Carlo algorithms with finite computational cost
-
Ilia N. Medvedev
Abstract
The issues of finite computational cost of some vector weighted Monte Carlo algorithms are studied in the paper relative to estimation of linear functionals of solutions to systems of the 2nd kind integral equations. A universal modification of the weight vector collision estimator with branching of the chain trajectory relative to the elements of matrix weight is constructed. It is proved that the computational cost of the constructed algorithm is finite in the case when the basic functionals are bounded. The results of numerical calculations are presented for the case of use of a modified weight estimator for some problems of the radiation transfer theory with allowance for polarization.
Acknowledgment
The author is grateful to Corresponding Member of the RAS G. A. Mikhailov and to Doctor of Physical and Mathematical Sciences S. A. Ukhinov for useful advice and remarks.
Funding: The work was supported by the Russian Foundation for Basic Research (projects no. 15–01–00894a, 16–01–00530a, and 17–01–00823a) and the Program of fundamental research of the Presidium of the RAS I.33.
References
[1] S. M. Ermakov and G. A. Mikhailov, Statistical Modelling. Nauka, Moscow, 1982 (in Russian).Search in Google Scholar
[2] G. I. Marchuk, G. A. Mikhailov, M. A. Nazaraliev, et al., The Monte Carlo Methods in Atmospheric Optics. Springer-Verlag, Berlin, Heidelberg, 1980.10.1007/978-3-540-35237-2Search in Google Scholar
[3] I. N. Medvedev and G. A. Mikhailov, Probabilistic-algebraic algorithms of Monte Carlo methods. Russ. J. Numer. Anal. Math. Modelling (2011) 26, No. 3, 323–336.10.1515/rjnamm.2011.018Search in Google Scholar
[4] G. A. Mikhailov, Optimization of Weighted Monte Carlo Methods. Springer-Verlag, Berlin, Heidelberg, 1992.10.1007/978-3-642-75981-9Search in Google Scholar
[5] G. A. Mikhailov and I. N. Medvedev, Optimization of Weighted Algorithms of Statistical Modelling. Omega Print, Novosibirsk, 2011 (in Russian).Search in Google Scholar
[6] G. A. Mikhailov and I. N. Medvedev, Improvement of weight computational statistical modelling via the transition to a subcritical Galton–Watson process. Doklady Math. (2009) 79, No. 1, 59–62.10.1134/S1064562409010177Search in Google Scholar
[7] G. A. Mikhailov, S. A. Ukhinov, and A. S. Chimaeva, Variance of a standard vector Monte Carlo estimator in the theory of polarized radiative transfer. Comp. Math. Math. Phys. (2006) 46, No. 11, 2099–2113.Search in Google Scholar
[8] G. A. Mikhailov and S. A. Ukhinov, Dual representation of the mean square of the Monte Carlo vector estimator. Doklady Math. (2011) 83, No. 3, 386–388.10.1134/S1064562411030380Search in Google Scholar
[9] B. A. Sevast’yanov, Branching Processes. Nauka, Moscow, 1971 (in Russian).Search in Google Scholar
[10] S. A. Ukhinov and D. I. Yurkov, Monte Carlo method of calculating the derivatives of polarized radiation. Russ. J. Numer. Anal. Math. Modelling (1998) 13, No. 5, 425–444.10.1515/rnam.1998.13.5.425Search in Google Scholar
© 2017 Walter de Gruyter GmbH Berlin/Boston
Articles in the same Issue
- Frontmatter
- Analysis of a regularized model for the isothermal two-component mixture with the diffuse interface
- Vector Monte Carlo algorithms with finite computational cost
- The method of hypersingular integral equations in the problem of electromagnetic wave diffraction by a dielectric body with a partial perfectly conducting coating
- Simulation of seismic processes in geological exploration of Arctic shelf
Articles in the same Issue
- Frontmatter
- Analysis of a regularized model for the isothermal two-component mixture with the diffuse interface
- Vector Monte Carlo algorithms with finite computational cost
- The method of hypersingular integral equations in the problem of electromagnetic wave diffraction by a dielectric body with a partial perfectly conducting coating
- Simulation of seismic processes in geological exploration of Arctic shelf
