Saved in:
| Main Author: | |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2406.15974 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866907954612011008 |
|---|---|
| author | Takeda, Masayoshi |
| author_facet | Takeda, Masayoshi |
| contents | Keller and Lenz \cite{KL} define a concept of {\it stochastic completeness at infinity} (SCI) for a regular symmetric Dirichlet form $(\cE,\cF)$. We show that (SCI) can be characterized probabilistically by using the predictable part $ζ^p$ of the life time $ζ$ of the symmetric Markov process $X=({\bf P}_x,X_t)$ generated by $(\cE,\cF)$, that is, (SCI) is equivalent to $\bfP_x(ζ=ζ^p<\infty)=0$. We define a concept, {\it explosion by killing} (EK), by $\bfP_x(ζ=ζ^i<\infty)=1$. Here $ζ^i$ is the totally inaccessible part of $ζ$. We see that (EK) is equivalent to (SCI) and $\bfP_x(ζ=\infty)=1$. Let $X^{\rm res}$ be the {\it resurrected process} generated by the {\it resurrected form}, a regular Dirichlet form constructed by removing the killing part from $(\cE, \cF)$. Extending work of Masamune and Schmidt (\cite{MS}), we show that (EK) is also equivalent to the ordinary conservation property of time changed process of $X^{\rm res}$ by $A^k_t$, where the $A^k_t$ is the positive continuous additive functional in the Revuz correspondence to the killing measure $k$ in the Beurling-Deny formula (Theorem \ref{ma-sh}). We consider the maximum principle for Schrödinger-type operator $\cL^μ=\cL-μ$. Here $\cL$ is the self-adjoint operator associated with $(\cE,\cF)$ %with non-local part and $μ$ is a Green-tight Kato measure. Let $λ(μ)$ be the principal eigenvalue of the trace of $(\cE,\cF)$ relative to $μ$. We prove that if (EK) holds, then $λ(μ)>1$ implies a Liouville property that every bounded solution to $\cL^μu=0$ is zero quasi-everywhere and that the {\it refined maximum principle} in the sense of Berestycki-Nirenberg-Varadhan \cite{BNV} holds for $\cL^μ$ if and only if $λ(μ)>1$ (Theorem \ref{RMP}). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_15974 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Explosion by Killing and Maximum Principle in Symmetric Markov Processes Takeda, Masayoshi Probability Keller and Lenz \cite{KL} define a concept of {\it stochastic completeness at infinity} (SCI) for a regular symmetric Dirichlet form $(\cE,\cF)$. We show that (SCI) can be characterized probabilistically by using the predictable part $ζ^p$ of the life time $ζ$ of the symmetric Markov process $X=({\bf P}_x,X_t)$ generated by $(\cE,\cF)$, that is, (SCI) is equivalent to $\bfP_x(ζ=ζ^p<\infty)=0$. We define a concept, {\it explosion by killing} (EK), by $\bfP_x(ζ=ζ^i<\infty)=1$. Here $ζ^i$ is the totally inaccessible part of $ζ$. We see that (EK) is equivalent to (SCI) and $\bfP_x(ζ=\infty)=1$. Let $X^{\rm res}$ be the {\it resurrected process} generated by the {\it resurrected form}, a regular Dirichlet form constructed by removing the killing part from $(\cE, \cF)$. Extending work of Masamune and Schmidt (\cite{MS}), we show that (EK) is also equivalent to the ordinary conservation property of time changed process of $X^{\rm res}$ by $A^k_t$, where the $A^k_t$ is the positive continuous additive functional in the Revuz correspondence to the killing measure $k$ in the Beurling-Deny formula (Theorem \ref{ma-sh}). We consider the maximum principle for Schrödinger-type operator $\cL^μ=\cL-μ$. Here $\cL$ is the self-adjoint operator associated with $(\cE,\cF)$ %with non-local part and $μ$ is a Green-tight Kato measure. Let $λ(μ)$ be the principal eigenvalue of the trace of $(\cE,\cF)$ relative to $μ$. We prove that if (EK) holds, then $λ(μ)>1$ implies a Liouville property that every bounded solution to $\cL^μu=0$ is zero quasi-everywhere and that the {\it refined maximum principle} in the sense of Berestycki-Nirenberg-Varadhan \cite{BNV} holds for $\cL^μ$ if and only if $λ(μ)>1$ (Theorem \ref{RMP}). |
| title | Explosion by Killing and Maximum Principle in Symmetric Markov Processes |
| topic | Probability |
| url | https://arxiv.org/abs/2406.15974 |