Abstract:We consider the range $R^{(n)}$, the tree made up of visited vertices by a diffusive null-recurrent randomly biased walk $\mathbb{X}$ on a Galton-Watson tree $\mathbb{T}$ up to the $n$-th return time to its root and we consider the following genealogy problem: pick two vertices uniformly at random in a generation of order $n$ in the tree $R^{(n)}$. Where does the coalescence occur? it turns out that the coalescence happens either in the recent past or in the remote past.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the genealogy problem of diffusive random walk diffusing in a random environment on a tree structure. Specifically, the researchers are concerned with where the most recent common ancestor (coalescence) of two randomly selected vertices in the \(n\) -th generation of a tree \(R(n)\) visited by a randomly biased random walk appears. The paper points out that this coalescence phenomenon occurs either in the recent past or in the remote past.
### Background and Problem Description
- **Research Object**: The paper considers the behavior of a randomly biased random walk in a random environment on a Galton - Watson tree \(T\). \(R(n)\) represents the subtree composed of vertices visited by the random walk within the time of returning to the root node for the \(n\) -th time.
- **Core Problem**: Randomly select two vertices in the \(n\) -th generation of \(R(n)\) and study where their most recent common ancestor appears.
### Main Findings
- **Coalescence Phenomenon**: The main finding of the paper is that the most recent common ancestor of these two vertices occurs either in the recent past (i.e., close to the \(n\) -th generation) or in the remote past (i.e., close to the root of the tree).
- **Mathematical Model**: The paper derives this conclusion by analyzing the properties of the randomly biased random walk, especially the behavior of its local time and multi - type additive martingale.
### Mathematical Tools
- **Randomly Biased Random Walk**: In the paper, a randomly biased random walk \(X\) on a Galton - Watson tree is defined, and its transition probability depends on the environmental variable \(V(x)\).
- **Local Time**: \(L(n)_k\) represents the number of visits of the random walk at the \(k\) -th level within the time of returning to the root node for the \(n\) -th time.
- **Multi - type Additive Martingale**: \(Z(n)_k\) represents the total number of visits of the random walk at the \(k\) -th level within the time of returning to the root node for the \(n\) -th time.
### Conclusion
Through rigorous mathematical analysis, the paper proves that for two randomly selected vertices in the \(n\) -th generation of \(R(n)\), their most recent common ancestor occurs either in the recent past or in the remote past. This finding is of great significance for understanding the genealogy of diffusion processes in random environments.
### Formula Display
- **Relationship between Local Time and Multi - type Additive Martingale**:
\[
L(n)_k=Z(n)_k + Z(n)_{k + 1}
\]
- **Limit Form of Coalescence Probability**:
\[
\lim_{n\rightarrow\infty}P\left(\left|X(1,n)_{\lfloor bn\rfloor}\wedge X(2,n)_{\lfloor bn\rfloor}\right|\geq\lfloor an\rfloor\right)=E\left[\frac{\sum_{j = 1}^{N_{a,b}}(\xi_j)^2}{\left(\sum_{j = 1}^{N_{a,b}}\xi_j\right)^2}\mid N_{a,b}>0\right]
\]
where \(N_{a,b}=\sum_{i = 1}^{N_b}G_{i,a,b}\), \(N_b\) is a Poisson random variable with parameter \(W_\infty/(bc_0)\), \(G_{i,a,b}\) is a geometric random variable with success probability \(1 - a/b\), and \(\xi_j\) is an exponential random variable with mean 1.
Hope this information is helpful for you to understand the content of this paper! If you have any further questions or need more detailed explanations, please feel free to let me know.