Abstract:In this paper, we study the upper tail large deviation for the one-dimensional frog model. In this model, sleeping and active frogs are assigned to vertices on $\mathbb Z$. While sleeping frogs do not move, the active ones move as independent simple random walks and activate any sleeping frogs. The main object of interest in this model is the asymptotic behavior of the first passage time ${\rm T}(0,n)$, which is the time needed to activate the frog at the vertex $n$, assuming there is only one active frog at $0$ at the beginning. While the law of large numbers and central limit theorems have been well established, the intricacies of large deviations remain elusive. Using renewal theory, Bérard and Ramírez have pointed out a slowdown phenomenon where the probability that the first passage time ${\rm T}(0,n)$ is significantly larger than its expectation decays sub-exponentially and lies between $\exp(-n^{1/2+o(1)})$ and $\exp(-n^{1/3+o(1)})$. In this article, using a novel covering process approach, we confirm that $1/2$ is the correct exponent, i.e., the rate of upper large deviations is given by $n^{1/2}$. Moreover, we obtain an explicit rate function that is characterized by properties of Brownian motion and is strictly concave.
What problem does this paper attempt to address?
### The problems the paper attempts to solve
This paper aims to study the Upper Tail Large Deviation in the one - dimensional frog model. In this model, frogs are assigned to the integer lattice points \( \mathbb{Z} \), and are divided into sleeping and active states. Sleeping frogs do not move, while active frogs move according to independent simple random walks and activate any sleeping frogs they come into contact with. The main object of study is the first - passage time \( T(0, n) \) from the origin \( 0 \) to the lattice point \( n \), that is, the time required to activate the frog located at \( n \), assuming that there is only one active frog at \( 0 \) at the initial moment.
Although the research on the law of large numbers and the central limit theorem is quite mature, the large - deviation problem is still complex and not completely solved. Berard and Ramirez [2] used renewal theory to point out a deceleration phenomenon, that is, the probability that the first - passage time \( T(0, n) \) is significantly greater than its expected value decays at a sub - exponential rate, and the decay rate is between \( \exp(-n^{1/2 + o(1)}) \) and \( \exp(-n^{1/3 + o(1)}) \). In this paper, through a novel covering - process method, it is confirmed that \( 1/2 \) is the correct exponent, that is, the rate of the upper - tail large deviation is given by \( n^{1/2} \). In addition, an explicit rate function is also obtained in this paper, which is determined by the properties of Brownian motion and is strictly convex - concave.
### Main results
For the one - dimensional frog model in discrete time, the main results of this paper are as follows:
- Define the dynamics of frogs and precisely define the diffusion speed.
- The first - passage time \( T(0, n) \) represents the diffusion speed from the origin \( 0 \) to the position \( n \).
- Describe the probability of upper - tail large deviation through the energy function \( E(f) \) of Brownian motion.
Specifically, this paper proves the following theorem:
**Theorem 1.2**: Let \( r^* := r(1) \) and \( \mu := \mu(1) \). Then \( r^* \) is positive and finite, and the first - passage time of the frog model on \( \mathbb{Z} \) satisfies the upper - tail large deviation with a rate of \( \sqrt{n} \), and the rate function is \( \xi \mapsto r^* \sqrt{\xi} \) (\( \xi > 0 \)). More specifically, for all \( \xi > 0 \),
\[ \lim_{n \to \infty} \frac{1}{\sqrt{n}} \log P\left( T(0, n) \geq (\mu + \xi) n \right)=-r^* \sqrt{\xi}. \]
### Related work
This paper also reviews the early literature on the frog model, including the study of recurrence/transient problems, sub - linear variances in multi - dimensional cases, and concentration inequalities, etc. These results provide clues for understanding the behavior of the first - passage time, but have not completely solved the problems of the central limit theorem and the large - deviation principle. By introducing a new covering - process method, this paper solves the problem of estimating the upper - tail large deviation in the one - dimensional frog model in discrete time.
### Proof sketch
The proof in this paper is mainly divided into the following steps:
1. **Scale invariance of energy**: Using the scale invariance of Brownian motion, prove the scale invariance of the energy function.
2. **Localization of upper - tail large - deviation events**: Observe that upper - tail large - deviation events are mainly affected by "bad intervals" of length \( \sqrt{n} \).
3. **Upper - tail estimate of the first - passage time on bad intervals**: Through the step - by - step deformation method, deform the function \( f \) from \( C(\xi) \) to the step function \( C_{\text{Step}}(\xi) \) and estimate the upper - tail probability.
4. **Conclusion**: Combining the above steps, finally prove Theorem 1.2.
### Conclusion
By introducing a new covering - process method, this paper successfully solves the discrete - time one - dimensional frog model.