Abstract:We introduce a new self-interacting random walk on the integers in a dynamic random environment and show that it converges to a pure diffusion in the scaling limit. We also find a lower bound on the diffusion coefficient in some special cases. With minor changes the same argument can be used to prove the scaling limit of the corresponding walk in Z^d.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the behavior and limit properties of a new type of self - interacting random walk in a dynamic random environment. Specifically, the author introduced a continuous - time nearest - neighbor symmetric random walk model on a one - dimensional integer lattice, where each time a particle jumps over a bond connecting two adjacent lattice points, the bond has a certain probability of breaking, and it needs to be repaired before it can be used again.
### Main problems and goals:
1. **Introducing a new model**: The paper introduced a new model of self - interacting random walk in a dynamic random environment, where the environment evolves over time and is non - Markovian.
2. **Proving recurrence**: It is proved that this random walk process is recurrent, that is, the particle will almost surely return to the origin infinitely many times.
3. **Determining the scaling limit**: The scaling limit of this random walk was studied, and it was proved that it converges to a pure diffusion process under the scaling limit, and a lower bound of the diffusion coefficient was given.
### Model description:
- The particle performs a continuous - time nearest - neighbor symmetric random walk on the integer lattice, with the initial position at the origin.
- The time interval between each jump follows an independent exponential distribution with parameter \(\lambda\).
- Each time a particle jumps over a bond, the bond breaks with probability \(0 < p\leq1\), and cannot be used again until it is repaired.
- The repair time of the broken bond also follows an independent exponential distribution with parameter \(\mu\).
- All bonds are intact at the initial moment.
### Main results:
1. **Recurrence**:
- Theorem 1: The process \(\{X(t)\}_{t\geq0}\) is recurrent.
- It was proved by Lemma 2 that at any time \(s\geq0\), there exists a time \(t\geq s\) such that there are no broken bonds at this time.
2. **Scaling limit**:
- Theorem 4: Under appropriate scaling, \(X(nt)/\sqrt{n}\) converges to a standard Brownian motion \(B(t)\), and the diffusion coefficient is \(\beta/\sqrt{\alpha}\).
- Where \(\alpha = E(\tau_1)\) is the expected time of the first return to zero, and \(\beta^{2}=\text{Var}(X(\tau_1))\) is the variance of the displacement when first returning to zero.
3. **Lower bound of the diffusion coefficient**:
- Proposition 6: In some special cases (such as when \(p = 1\) and \(\mu\) is large enough), the diffusion coefficient \(\beta/\sqrt{\alpha}\) is greater than 1, indicating that the effect of bond breaking makes the process move faster under the scaling limit.
### Summary:
This paper introduced a new model of self - interacting random walk in a dynamic random environment, analyzed its recurrence and scaling limit in detail, and gave a lower bound of the diffusion coefficient, revealing the influence of bond breaking on the behavior of random walk.