Preprints
Abstract:
Spatial birth and death chains have long been used to
simulate point processes such as the Strauss process.
Kendall and Thonnes showed how to utilize coupling from
the past with these chains in order to generate perfect
samples. This paper presents an analysis of their method,
giving sufficient conditions for the algorithms to run in
polynomial time. In addition, a new chain is proposed for
Strauss processes with an extra type of move: a swap.
This gives a new perfect simulation algorithm that runs in
polynomial time over a larger class of models. While the
behavior of this move is described for the Strauss
process, it should be applicable to a wide range of
processes.