Summary
It has been found recently that an increase in phase space dimension by including simulated auxiliary random variables in the number of phase coordinates can be effective for the construction of weight modifications. In this paper the effectiveness of “value” and partial “value” modelling is considered. These types of modelling are related to the construction of simulated distribution for some auxiliary random variable by multiplying the initial density by the “value” function which is usually corresponds to the solution of adjoint integral equation of the second kind. It is proved that the weight estimator variance in case of the partial value modelling is finite. On the basis of this fact a new criterion based on the use of majorant adjoint equation was proposed for finiteness of the weight estimator variance. Using this criterion the classical “exponential transformation” method is studied for the free path simulation in one and three dimensional modifications.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
B. Davidson. Neutron Transport Theory. Atomizdat, Moscow, 1960 (in Russian).
I. Dimov. Minimization of the probable error for some Monte Carlo methods. In Proc. of the Summer School on Mathematical Modelling and Scientific Computations, 23–28.09.1990, Albena, Bulgaria, Sofia, Publishing House of the Bulgarian Academy of Sciences, pp. 159–170, 1991.
S. Ermakov and G. Mikhailov. Statistical Simulation. Nauka, Moscow, 1982 (in Russian).
I. Medvedev. The variance of weight estimate in the case of partial “value” modelling. In Proceedings of the 5th St. Petersburg Workshop on Simulations. NII Chemistry St.Petersburg University Publishers, pp. 465–470, 2005.
G. Mikhailov. Optimization of Monte-Carlo Weighted Methods. Springer-Verlag, 1992.
G. Mikhailov. Construction of Weighted Monte Carlo Methods via an Increase in the Phase Space Dimension. Dokl. Ross. Akad. Nauk., Vol. 389, No. 4, pp. 461–464, 2003.
G. Mikhailov and I. Medvedev. Optimization of weighted Monte-Carlo methods with respect to auxiliary variables. Siberian Mathematical Journal, Vol. 45, No. 2, pp. 331–340, 2004.
I. Medvedev and G. Mikhailov. Optimization of weighted methods for part of variables. Russ. J. Numer. Anal. Math. Modelling. Vol. 18, No. 5, pp. 397–412, 2003.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2008 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Medvedev, I., Mikhailov, G. (2008). A New Criterion for Finiteness of Weight Estimator Variance in Statistical Simulation. In: Keller, A., Heinrich, S., Niederreiter, H. (eds) Monte Carlo and Quasi-Monte Carlo Methods 2006. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-74496-2_33
Download citation
DOI: https://doi.org/10.1007/978-3-540-74496-2_33
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-74495-5
Online ISBN: 978-3-540-74496-2
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)