Abstract:In this paper, we construct scaling limits of some branching random walks in random environment whose off-spring distributions have infinite variance. The Laplace functional of the obtained random measure is given by a non-linear PAM, whose existence and uniqueness are also proved as an intermediate step. We also give a martingale characterization of above super-process and show that it possesses the compact support property and exponential persisitency.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the limiting behavior of branching random walks with infinite - variance offspring distributions in a random environment. Specifically, the author explores the large - scale behavior of branching random walks on the two - dimensional lattice \( \mathbb{Z}^2 \), where the branching mechanism and branching rate depend on the random environment at each spatial point. The main objective of the paper is to generalize the results of classical branching random walks, especially when the offspring distribution has an infinite variance.
### Main Problems
1. **Limiting Behavior of Branching Random Walks with Infinite - Variance Offspring Distributions**:
- Classical results of branching random walks usually assume that the offspring distribution has a finite variance. This paper aims to study what the limiting process of branching random walks is when the offspring distribution has an infinite variance.
2. **Branching Random Walks in a Random Environment**:
- Consider a random environment \( V(x)=\xi(x) \), where \( \{\xi(x)\}_{x\in\mathbb{Z}^2} \) are independent and identically distributed random variables, and \( E[\Phi]=0 \) and \( E[\Phi^2]=1 \). What is the limiting process of branching random walks in this environment?
3. **Existence and Uniqueness of Solutions to Nonlinear Partial Differential Equations**:
- The paper studies the existence and uniqueness of solutions to the following nonlinear partial differential equation:
\[
\partial_t\phi = (\Delta+\xi)\phi-\phi^{1 + \beta},
\]
where \( \Delta \) is the Laplace operator, \( \xi \) is white noise, and \( \beta \) is a parameter with \( 0 < \beta < 1 \).
### Main Contributions
1. **Definition and Properties of the Limiting Process**:
- The author defines a new limiting process, called \( \beta \)-rough super - Brownian motion, and gives the form of its Laplace functional.
2. **Martingale Characterization and Compact Support Property**:
- Proves the martingale characterization of \( \beta \)-rough super - Brownian motion and shows that it has the compact support property.
3. **Existence and Uniqueness of Solutions**:
- By introducing the concepts of weighted Besov spaces and para - controlled distributions, proves the existence and uniqueness of solutions to the above - mentioned nonlinear partial differential equation.
### Technical Difficulties
1. **Irregularity of the Anderson Hamiltonian**:
- Due to the irregularity of \( \xi \), it is impossible to directly define \( \xi\phi \). The author uses the theory of regularity structures and para - controlled distributions to deal with this problem.
2. **Difficulties in Weight Operations**:
- When solving the corresponding linear equation \( \partial_t\phi = H\phi \), weight operations bring additional challenges. The author overcomes this difficulty by introducing an exponentially growing weight function.
### Structural Overview
- **Section 3**: Introduces the model and the definition of \( \beta \)-rough super - Brownian motion, and states the main results.
- **Section 4**: Studies variants of PAM and nonlinear equations, and proves the existence and uniqueness of solutions.
- **Section 5**: Gives a detailed proof that the limiting process of branching random walks converges to \( \beta \)-rough super - Brownian motion.
- **Section 6**: Gives the proof of the martingale problem characterization and the compact support property of \( \beta \)-rough super - Brownian motion.
Through these works, the paper successfully extends the classical branching random walk theory and solves the problem of the limiting behavior of branching random walks with infinite - variance offspring distributions in a random environment.