arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1811.03345cond-mat.stat-mech
收藏

受约束布朗游程下面积的大涨落

Large fluctuations of the area under a constrained Brownian excursion

Baruch Meerson

摘要:

我们研究了大波动的区域$\mathcal{A}$下的布朗漂移$x(t)$的时间间隔$|t|\leq T$,限制远离移动壁$x_0(t)$,使$x_0(-T)=x_0(T)=0$和$x_0(|t|<T)>0$。我们专注于由广义抛物线族$x_0(t)=T^γ [1-(t/T)^{2k}]$描述的壁函数,其中$k\geq 1$。利用最优涨落法(OFM),计算了长时间内面积的大偏差函数(LDF)。OFM提供了一个简单的描述,在布朗运动的最佳路径,或射线的面积波动。我们表明,LDF有一个跳跃的三阶导数相对于$\mathcal{A}$在一个临界值$\mathcal{A}$。这种奇异性是由最佳路径的质的变化引起的,可以解释为三阶动力学相变。虽然OFM是不适用于典型的(小)面积波动,我们认为,它正确地捕捉到他们的幂律缩放的$\mathcal{A}$与$T$的指数,连续依赖于$γ$和$k$。我们还考虑余弦壁$x_0(t)=T^γ \cos[πt/(2T)]$,以说明不同的可能行为的最佳路径和缩放的典型波动。对于某些壁函数,由多种OFM解决方案共存引起的额外相变应该是可能的。

英文摘要:

We study large fluctuations of the area $\mathcal{A}$ under a Brownian excursion $x(t)$ on the time interval $|t|\leq T$, constrained to stay away from a moving wall $x_0(t)$ such that $x_0(-T)=x_0(T)=0$ and $x_0(|t|<T)>0$. We focus on wall functions described by a family of generalized parabolas $x_0(t)=T^γ [1-(t/T)^{2k}]$, where $k\geq 1$. Using the optimal fluctuation method (OFM), we calculate the large deviation function (LDF) of the area at long times. The OFM provides a simple description of the area fluctuations in terms of optimal paths, or rays, of the Brownian motion. We show that the LDF has a jump in the third derivative with respect to $\mathcal{A}$ at a critical value of $\mathcal{A}$. This singularity results from a qualitative change of the optimal path, and it can be interpreted as a third-order dynamical phase transition. Although the OFM is not applicable for typical (small) area fluctuations, we argue that it correctly captures their power-law scaling of $\mathcal{A}$ with $T$ with an exponent that depends continuously on $γ$ and on $k$. We also consider the cosine wall $x_0(t)=T^γ \cos[πt/(2T)]$ to illustrate a different possible behavior of the optimal path and of the scaling of typical fluctuations. For some wall functions additional phase transitions, which result from a coexistence of multiple OFM solutions, should be possible.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑