Potential Analysis ( IF 1.1 ) Pub Date : 2020-06-26 , DOI: 10.1007/s11118-020-09858-0 Masatoshi Fukushima , Yoichi Oshima
For a Dirichlet form \((\mathcal {E},\mathcal {F})\) on L2(E;m), let \(\mathbb {G}(\mathcal {E})=\{X_{u};u\in \mathcal {F}_{e}\}\) be the Gaussian field indexed by the extended Dirichlet space \(\mathcal {F}_{e}\). We first solve the equilibrium problem for a regular recurrent Dirichlet form \(\mathcal {E}\) of finding for a closed set B a probability measure μB concentrated on B whose recurrent potential \(R\mu ^{B}\in \mathcal {F}_{e}\) is constant q.e. on B (called a Robin constant). We next assume that E is the complex plane \(\mathbb {C}\) and \(\mathcal {E}\) is a regular recurrent strongly local Dirichlet form. For the closed disk \(\bar B(\textbf {x},r)=\{\textbf {z}\in \mathbb {C}:|\textbf {z}-\textbf {x}|\le r\}\), let μx, r and f(x, r) be its equilibrium measure and Robin constant. Denote the Gaussian random variable \(X_{R\mu ^{\textbf {x}.r}}\in \mathbb {G}(\mathcal {E})\) by Yx, r and let, for a given constant γ > 0, \(\mu _{r}(A,\omega )={\int \limits }_{A} \exp (\gamma Y^{\textbf {x},r}-(1/2)\gamma ^{2} f(\textbf {x},r))d\textbf {x}.\) Under a certain condition on the growth rate of f(x, r), we prove the convergence in probability of μr(A, ω) to a random measure \(\overline {\mu }(A,\omega )\) as r↓ 0. The possible range of γ to admit a non-trivial limit will then be examined in the cases that \((\mathcal {E}.\mathcal {F})\) equals \((\frac 12{\textbf {D}}_{\mathbb {C}},H^{1}(\mathbb {C}))\) and \((\textbf {a},H^{1}(\mathbb {C}))\), where a corresponds to the uniformly elliptic partial differential operator of divergence form.
中文翻译:
Dirichlet形式的高斯场,平衡势和可乘混沌
对于L 2(E ; m)上的Dirichlet形式\((\ mathcal {E},\ mathcal {F})\),令\(\ mathbb {G}(\ mathcal {E})= \ {X_ { \ mathcal {F} _ {e} \} \)中的u}; u \是由扩展Dirichlet空间\(\ mathcal {F} _ {e} \)索引的高斯字段。我们首先解决常规复发狄利克雷形式的平衡问题\(\ mathcal {E} \)找到一个闭集的乙一个概率测度μ乙集中在乙其复发潜在\(R \亩^ {B} \在\ mathcal {F} _ {e} \)在B上为常数qe(称为Robin常数)。接下来,我们假设E是复平面\(\ mathbb {C} \),而\(\ mathcal {E} \)是规则的递归强局部Dirichlet形式。对于封闭的磁盘\(\ bar B(\ textbf {x},r)= \ {\ textbf {z} \ in \ mathbb {C}:| \ textbf {z}-\ textbf {x} | \ le r \} \) ,让μ X,- [R和˚F(X,- [R )是它的平衡的措施和Robin恒定。表示高斯随机变量\(X_ {R \亩^ {\ textbf {X} .R}} \在\ mathbb {G}(\ mathcal {E})\)由Ý X, - [R和让,对于给定常数γ > 0,\(\ mu _ {r}(A,\ omega)= {\ int \ limits} _ {A} \ exp(\ gamma Y ^ {\ textbf {x},r}-(1/2)\ gamma ^ {2} F(\ textbf {X},R))d \ textbf {X}。\)根据上的生长速度在一定条件˚F(X,- [R )中,我们证明了收敛的概率μ - [R (甲,ω),以一个随机测度\(\划线{\亩}(A,\欧米加)\)作为- [R ↓ 0的可能范围的γ承认非平凡限制将然后在箱子待检查\( (\ mathcal {E}。\ mathcal {F})\)等于\((\ frac 12 {\ textbf {D}} _ {\ mathbb {C}},H ^ {1}(\ mathbb {C}) )\)和\((\ textbf {a},H ^ {1}(\ mathbb {C}))\),其中a对应于散度形式的均匀椭圆偏微分算子。