Short-time blowup statistics of a Brownian particle in repulsive potentials
Baruch Meerson · 2025
Read Short-time blowup statistics of a Brownian particle in repulsive potentials by Baruch Meerson (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-01252.txt.
Introduction
Baruch Meerson studies a tiny particle jittering under Brownian motion in a repulsive potential, a landscape that pushes it outward ever faster. For steep enough potentials, the particle reaches infinity in a finite time, and that blowup time varies at random from run to run. Meerson focuses on very short blowup times.
Earlier work by Krapivsky and Meerson found the leading shape of that short-time tail, which has a sharp essential singularity at zero. Meerson now computes the large pre-exponential factor that was missing, for all integer steepness values. The method applies a WKB approximation, a wave-based estimate, to a transformed equation, at two orders of accuracy. Odd cases need an added fix in a thin layer near the origin, completing a precise statistical picture.
Catalog record
| CREATOR | Baruch Meerson |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-01252
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Short-time blowup statistics of a Brownian particle in repulsive potentials Baruch Meerson Abstract We study the dynamics of an overdamped Brownian particle in a repulsive scale-invariant potential V ( x ) ∼ − x n + 1 V(x)\sim-x^{n+1} . For n > 1 n>1 , a particle starting at position x x reaches infinity in a finite, randomly distributed time. We focus on the short-time tail T → 0 T\to 0 of the probability distribution P ( T , x , n ) P(T,x,n) of the blowup time T T for integer n > 1 n>1 . Krapivsky and Meerson [Phys. Rev. E 112 , 024128 (2025)] recently evaluated the leading-order asymptotics of this tail, which exhibits an n n -dependent essential singularity at T = 0 T=0 . Here we provide a more accurate description of the T → 0 T\to 0 tail by calculating, for all n = 2 , 3 , … n=2,3,\dots , the previously unknown large pre-exponential factor of the blowup-time probability distribution. To this end, we apply a WKB approximation—at both leading and subleading orders—to the Laplace-transformed backward Fokker–Planck equation governing P ( T , x , n ) P(T,x,n) . For even n n , the WKB solution alone suffices. For odd n n , however, the WKB solution breaks down in a narrow boundary layer around x = 0 x=0 . In this case, it must be supplemented by an “internal” solution and a matching procedure between the two solutions in their common region of validity. I Introduction Finite-time singularities are ubiquitous in nature, and they have attracted much attention from physicists and mathematicians, see e.g. Refs. MM1975 ; CP1993 ; cosmology ; EggersRMP ; EF2000 . When a dynamical quantity reaches infinity in a finite time, we call this singularity a blowup. Probably the simplest model of a blowup is provided by a first-order differential equation \dot{x}(t)=x^{n}(t)\,, (1) where n>1 . Indeed, for any positive initial condition x(t=0)\equiv x_{0}>0 , the solution of this equation, x(t)=\frac{x_{0}}{\left[1-(n-1)x_{0}^{n-1}t\right]^{\frac{1}{n-1}}}, (2) exhibits a blowup at time t=(n-1)^{-1}x_{0}^{1-n} . Equation ( 1 ) describes an overdamped deterministic motion of a particle on the line in a repulsive potential U(x,n)=-\frac{x^{n+1}}{n+1}\,. (3) In the presence of noise, which we will model as an additive white Gaussian noise with zero mean, Eq. ( 1 ) gives way to the Langevin equation \dot{x}(t)=x^{n}(t)+\sqrt{2}\eta(t)\,,n=2,3,\ldots\,, (4) with \langle\eta(t_{1})\eta(t_{2})\rangle=\delta(t_{1}-t_{2}) . Equation ( 4 ) describes an overdamped motion of a Brownian particle in the power-law repulsive potential rescaling . Here the particle reaches infinity almost surely even for negative x_{0} . For odd n>1 , the blowups to plus and minus infinity are both possible. For even n\geq 2 , the blowup is only to +\infty . The blowup time is random, and it is interesting to determine its distribution. Mathematically the blowup problem is a particular case of the first passage problem, and one can adapt to this problem some well-developed mathematical tools Gardiner ; SRbook ; KRBbook . Different aspects of the model, described by Eq. ( 4 ), have been studied in theory and in experiment Ornigotti ; SilerPRL ; Ryabov ; KM2025 . Besides being a rich and interesting model, it appears to be quite useful. This is because potentials which grow with the distance faster than quadratically successfully model different aspects of a host of stochastic processes in physics and biology. For example, Eq. ( 4 ) with n=2 describes the dynamics at the saddle-node bifurcation Horsthemke ; Broeck94 which is encountered in the context of optical bistability in lasers Colet ; laser91 , firing of neurons neuron-1 ; neuron-2 , Brownian ratchets Broeck01 ; Dean17 ; Arzola , and nonlinear maps Hirsch . As of present, the blowup time probability distribution \mathcal{P}(T,x_{0},n) – or, more precisely, its generating function, or the Laplace transform T\to s – has been calculated explicitly only for the quartic potential, n=3 KM2025 . Even in that particular case the exact solution – the inverse Laplace transform of an expression which includes a bi-confluent Heun function, see below – can only be evaluated numerically. To get a useful insight into this problem it is desirable to obtain analytical asymptotics for the large- T and small- T tails of the distribution \mathcal{P}(T,x_{0},n) . The large- T tail of \mathcal{P}(T,x_{0},n) is purely exponential KM2025 . It is determined by the closest to s=0 pole of the generating function or, in the spectral language, by the ground state of the direct Fokker-Planck operator corresponding to the Langevin equation ( 4 ) KM2025 . The small- T tail of \mathcal{P}(T,x_{0},n) is quite different and nontrivial. Here infinitely many terms in the pole decomposition of the generating function (or, equivalently, infinitely many eigenstates of the Fokker-Planck operator) contribute KM2025 , calling for a different approach. The authors of Ref. KM2025 evaluated this tail by employing the optimal fluctuation method (OFM) which involves a leading-order saddle-point calculation of the path integral corresponding to the Langevin equation ( 4 ). The OFM is a variant of time-dependent WKB approximation. The OFM calculation boils down to determining the optimal (that is, most likely) path x_{*}(t) – a deterministic trajectory which minimizes the action functional of the path integral conditioned on starting at a specified point x_{0} and reaching infinity at a specified time T KM2025 . This calculation yields a leading-order asymptotic of the logarithm of \mathcal{P}(T\to 0,x_{0},n) . In particular, for x_{0}=0 (a convenient benchmark which we will also adopt in the following) they obtained KM2025 \mathcal{P}(T\to 0,n)\sim\exp\left(-\alpha_{n}\,T^{-\tfrac{n+1}{n-1}}\right)\,,\quad n=2,3,\dots (5) where \alpha_{n}=\frac{\Gamma\left(\frac{3}{2}-\frac{1}{2n}\right)\Gamma\!\left(\frac{1}{2n}\right)\left[\Gamma\left(1+\frac{1}{2n}\right)\Gamma\!\left(\frac{n-1}{2n}\right)\right]^{\frac{n+1}{n-1}}}{4(n+1)\pi^{\frac{n}{n-1}}}>0\,, (6) and \Gamma(\dots) is the gamma function. The n -dependent essential singularity of the T\to 0 tail is clearly seen from Eq. ( 5 ). Equation ( 5 ), however, misses a pre-exponential factor which can be important because of its possible dependence on the small parameter T . As we will see shortly, the prefactor does depend on T in this problem, and it is large. The determination of this prefactor demands going beyond the leading-order OFM/WKB calculation. In the remainder of this paper we provide an accurate description of the T\to 0 tail of the blowup time distribution \mathcal{P}(T,x_{0},n) by explicitly calculating, for all n=2,3,\dots , this previously unknown pre-exponential factor. To achieve this goal we employ a different version of the WKB method, in the leading and subleading orders, by applying it to the Laplace-transformed backward Fokker–Planck equation, which describes the generating function of \mathcal{P}(T,x_{0},n) . This variant of WKB theory is advantageous because – by virtue of the Laplace transform and in contrast to the time-dependent OFM – it deals with a time-independent problem. As we will see, for even n the (leading and subleading) WKB solution alone suffices for calculating the prefactor. For odd n , however, the WKB solution breaks down in a small boundary layer around x=0 . Here we proceed by finding a separate, “internal” solution, which then can be matched to the WKB solution in their common region of validity. Here is a plan of the remainder of the paper. In Sec. II we recap the exact formulation of the problem. The WKB solution, the internal solution and their matching are presented in Sec. III .