On a decomposition of Hunt convolution kernels
Put \(A=\{-1,1,0\}\), \(S_n = A\times A\times\cdot\times A\) \((n\) factors), \(R_{\sigma} = \{x=(x_1,\ldots,x_n)\in\mathbb{R}^n: \text{sgn }x_j=\sigma_j, 1\leq j\leq n\}\cup \{0\}\). Let \(M_+\) denote the set of all positive Radon measures in \(\mathbb{R}^n\). The paper deals with representations of Hunt kernels (in \(\mathbb{R}^n)\) as convolutions of Hunt kernels supported by octants \(R_{\sigma}\). A Hunt kernel is a measure \(N\in M_+\) of the form \(N=\int^{+\infty}_{0}\alpha_t\,dt,\) \(\{\alpha_t\}\) being a weakly continuous semigroup in \(M_+\) \((\alpha_0=\delta\), \(\alpha_{t+s}=\alpha_ t*\alpha_s\), \(t,s\geq 0)\). A measure of the form \(c(\delta +\sum^{\infty}_{n=1}(\sigma)^{*n}),\) where \(c\in (0,+\infty)\), \(\sigma \in M_+\), is called an elementary kernel; it is a Hunt kernel. Theorem. Every elementary kernel is a convolution of a family \(\{N_{\sigma}\}\) \((\sigma \in S_n)\) of elementary kernels with \(N_{\sigma}(\mathbb{R}^n\setminus R_{\sigma})=0\); kernels \(N_{\sigma}\) are uniquely determined by \(N\) (up to a constant factor). An analogous theorem is proved for general Hunt kernels. As an application on a class of Hunt kernels on the real line is explicitly described. [For the entire collection see Zbl 0527.00005.]
- Uniqueness of the decomposition of Hunt convolution kernels
- scientific article; zbMATH DE number 3204069
- scientific article; zbMATH DE number 3003689
- scientific article; zbMATH DE number 3003688
- scientific article; zbMATH DE number 3986277
- On some results of HK-convolution
- scientific article; zbMATH DE number 3013547
- scientific article; zbMATH DE number 4109330
- scientific article; zbMATH DE number 4127858
- scientific article; zbMATH DE number 3852206
This page was built for publication: On a decomposition of Hunt convolution kernels
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3345055)