Monday, May 20, 2013

1305.4154 (Krzysztof Bozek et al.)

Markov Chain Mote Carlo solution of BK equation through
Newton-Kantorovich method
   [PDF]

Krzysztof Bozek, Krzysztof Kutak, Wieslaw Placzek
We propose a new method for Monte Carlo solution of non-linear integral equations by combining the Newton-Kantorovich method for solving non-linear equations with the Markov Chain Monte Carlo (MCMC) method for solving linear equations. The Newton-Kantorovich method allows to express the non-linear equation as a system of the linear equations which then can be treated by the MCMC (random walk) algorithm. We apply this method to the Balitsky-Kovchegov (BK) equation describing evolution of gluon density at low x. Results of numerical computations show that the MCMC method is both precise and efficient. The presented algorithm may be particularly suited for solving more complicated and higher-dimensional non-linear integral equation, for which traditional methods become unfeasible.
View original: http://arxiv.org/abs/1305.4154

No comments:

Post a Comment