Central limit theorem for the contact process (Q1820517)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Central limit theorem for the contact process |
scientific article |
Statements
Central limit theorem for the contact process (English)
0 references
1986
0 references
This tightly written paper proves a central limit theorem for a cylindrical function f: P(\({\mathbb{Z}})\to {\mathbb{R}}\) and a supercritical contact process \(\xi^ A(t)\) starting on an initial configuration \(| A| =\infty\) in \({\mathbb{Z}}\). Briefly, as \(T\to \infty\), if \(\mu\) is the nontrivial invariant measure, then \[ T^{1/2}[T^{- 1}\int^{T}_{0}f(\xi^ A(t))dt-\int f d\mu] \] approaches a (possibly trivial) normal distribution with 0 mean. The main steps of the proof, motivated by methods developed by \textit{J. T. Cox} and \textit{D. Griffeath} [Ann. Probab. 11, 876-893 (1983; Zbl 0527.60095)] for occupation time limit theorems in the voter process, is as follows. (\(\zeta\) (t),t\(\geq 0)\), the contact process with random initial distribution with measure \(\mu\), is introduced. A similar limit theorem is proved with \(\zeta\) (t) replacing \(\xi^ A(t)\) by a skeleton argument involving integrals related to \(\zeta\) (t). An associated FKG set of variables bounds the correlations over time of the integrals. The bound depends on a decay time for the correlations of f(\(\zeta\) (t)), derived by considering the dual percolation structure for the contact process. The proof is completed by a standard theorem showing that the difference between the limit theorems involving \(\zeta\) (t) and \(\xi^ A(t)\) approaches 0 in probability.
0 references
central limit theorem for a cylindrical function
0 references
supercritical contact process
0 references
occupation time limit theorems in the voter process
0 references
percolation structure
0 references