In this study, we analyzed urn models by solving the discrete-time master equation using an expansion in moments. This approach is a viable alternative to conventional methods, such as system-size expansion, allowing for the determination of analytical expressions for the mean and variance in an exact form and thus valid for any system size. In particular, this approach was used to study Bernoulli-Laplace and Ehrenfest urns, for which analytic expressions describing their evolution were found. This approach and the results will contribute to a more comprehensive understanding of stochastic systems and statistical physics for small-sized systems, where the thermodynamic limit cannot be assumed.
Large deviations of a long-time average in the Ehrenfest Urn Model
自1907年成立以来,EUM一直是统计力学关键概念的测试平台。在这里,我们采用这个模型来研究时间可加量的大偏差。我们考虑两个连续时间版本的EUM与$K$瓮和$N$球:没有和在同一瓮中的球之间的相互作用。我们评估概率分布$\mathcal{P}_T(\bar{n}= a N)$,即一个瓮中的球的平均数量随着时间的推移$T$,$\bar{n}$,取任何指定值$aN$,其中$0\leq a\leq 1$。对于长观测时间,$T\to \infty$,即Donsker-Varadhan大偏差原理成立:$-\ln \mathcal{P}_T(\bar{n}= a N) \simeq T I(a,N,K,\dots)$,其中$\dots$表示模型的附加参数。我们计算率函数$I(a,N,K, \dots)$准确地由两种不同的方法,由于Donsker和Varadhan和比较的确切结果与WKB近似的变体(后Wentzel,Kramers和布里渊)。在不存在相互作用的情况下,$I(a,N,K, \dots)$的WKB预测对于任何$N$都是精确的。在存在相互作用的情况下,WKB方法给出了$N\gg 1$的渐近精确结果。WKB方法还揭示了系统的(非常简单的)时间历程,该时间历程主导了不同时间历程对$\mathcal{P}_T(\bar{n}= a N)$的贡献。
Vertex-cover in random graphs with small connectivity: an exact solution
本文已被作者撤回,因为其中的主要结果已在[1]中得到,对任意c < e,也可参见[2]和[3]。此外,在[4]中已经导出了树的最小顶点覆盖的公式(见摘要)。感谢M。Bauer,O. Golinelli,F. Ricci-Tersenghi,G. Semerjian和M.我很感激你引起了我的注意[1]和M.B.和O.G.为[4]。[1]M. Bauer和O. Golinelli,Eur.物理学杂志B 24,339-352(2001); [2] R. M. Karp和M. Sipser,Proc. 22nd IEEE Symposium on Foundations of Computing,(1981),364-375; J. Aronson,A. Frieze和B.G. Pittel,Random Structures and Algorithms 12(1998)111-177; [4] M. Bauer,O. Golinelli,Journal of DNA Sequences,2000年,第3期。