Readme file

SERIES B 
Statistical Methodology

Extensions of Fill's algorithm for perfect simulation, by J. Møller and K.Schladitz,
Journal of the Royal Statistical Society, Series B, Statistical Methodology

The datasets contain the code for perfect simulation of two discrete Markov random fields -- the

AUTO POISSON MODEL and the HARDCORE MODEL

on a square lattice using 4 different algorithms (Fill's algorithm and Coupling from the past with anti-monotone Gibbs sampler or sandwiching).

Fill's algorithm with anti-monotone Gibbs sampler: Algorithm 2,chapter 3

Fill's algorithm with sandwiching: Algorithm 3, chapter 4

Files: AutoPois.S contains Splus functions for the auto-Poisson model

autopois.c contains C-subroutines called by the functions in

AutoPois.S, to be compiled using !Splus COMPILE

Hardcore.S contains Splus functions for the hardcore model

hardcore.c contains C-subroutines called by the functions in

Hardcore.S, to be compiled using !Splus COMPILE

Contact address:

Institut fuer Techno- und Wirtschaftsmathematik
Aussenstelle PRE-Park:
Luxemburger Straße 1-3
D-67657 Kaiserslautern

email: schladitz@itwm.uni-kl.de

The code may be used in any manner without permission from the author. It is provided with no warranties.

Journals

SERIES A
Statistics in Society

SERIES B
Statistical Methodology

SERIES C
Applied Statistics

SERIES D
The Statistician