A Strong Central Limit Theorem for a Class of Random Surfaces
A Strong Central Limit Theorem for a Class of Random Surfaces
复制标题
一类随机曲面的强中心极限定理
DOI:
--
复制
发表时间:
2011
期刊:
影响因子:
--
通讯作者:
T. Spencer
中科院分区:
文献类型:
--
作者:
J. Conlon;T. Spencer
This paper is concerned with d = 2 dimensional lattice field models with action $${V(\nabla\phi(\cdot))}$$V(∇ϕ(·)), where $${V : \mathbf{R}^d \rightarrow \mathbf{R}}$$V:Rd→R is a uniformly convex function. The fluctuations of the variable $${\phi(0) - \phi(x)}$$ϕ(0)-ϕ(x) are studied for large |x| via the generating function given by $${g(x, \mu) = \ln \langle e^{\mu(\phi(0) - \phi(x))}\rangle_{A}}$$g(x,μ)=ln〈eμ(ϕ(0)-ϕ(x))〉A. In two dimensions $${g'' (x, \mu) = \partial^2g(x, \mu)/\partial\mu^2}$$g′′(x,μ)=∂2g(x,μ)/∂μ2 is proportional to $${\ln\vert x\vert}$$ln|x|. The main result of this paper is a bound on $${g''' (x, \mu) = \partial^3 g(x, \mu)/\partial \mu^3}$$g′′′(x,μ)=∂3g(x,μ)/∂μ3 which is uniform in $${\vert x \vert}$$|x| for a class of convex V. The proof uses integration by parts following Helffer–Sjöstrand and Witten, and relies on estimates of singular integral operators on weighted Hilbert spaces.