Abstract

The efficiency of the multicanonical procedure can be significantly improved by applying an additional bias to the numerically generated sample space. However, results obtained by biasing in different sampling regions cannot in general be accurately combined, since their relative normalization coefficient is not known precisely. We demonstrate that for overlapping biasing regions a simple iterative procedure can be employed to determine the required coefficients.

© 2007 Optical Society of America

Improved multicanonical algorithms

David Yevick and Tao Lu
J. Opt. Soc. Am. A 23(11) 2912-2918 (2006)

Comparison of transition-matrix sampling procedures

David Yevick, Michael Reimer, and Bjarne Tromborg
J. Opt. Soc. Am. A 26(1) 184-187 (2009)

Multicanonical comparison of polarization-mode dispersion compensator performance

Tao Lu, Weihong Huang, David Yevick, Maurice O’Sullivan, and Michael Reimer
J. Opt. Soc. Am. A 22(12) 2804-2809 (2005)

Use of multicanonical Monte Carlo simulations to obtain accurate bit error rates in optical communications systems

Ronald Holzlööhner and Curtis R. Menyuk
Opt. Lett. 28(20) 1894-1896 (2003)

Simplified transition matrix analysis of the hinge model

David Yevick, Michael Reimer, and Maurice O'Sullivan
J. Opt. Soc. Am. A 26(3) 710-714 (2009)