math0612856/fls.tex
1: %using general template file for the LaTeX package SVJour2
2: % for Springer journals.          Springer Heidelberg 2004/10/15
3: 
4: \documentclass[nospthms]{svjour2m}                    % onecolumn
5: %\documentclass[twocolumn,fleqn]{svjour2}    % twocolumn
6: %\documentclass[onecollarge]{svjour2}       % onecolumn "king-size"
7: %
8: %\smartqed  % flush right qed marks, e.g. at end of proof
9: %
10: %\usepackage{graphicx}
11: %
12: \usepackage{mathptmx}      % use Times fonts if available on your TeX system
13: %
14: % insert here the call for the packages your document requires
15: %\usepackage{latexsym}
16: % etc.
17: \usepackage{amsfonts}
18: \usepackage{amsmath}
19: \usepackage{amssymb}
20: \usepackage{amsthm}
21: \usepackage[usenames]{color}
22: %\usepackage[notcite,notref]{showkeys}
23: %\usepackage{a4wide}
24: %\setlength{\textheight}{19cm} %(laptop)
25: 
26: %
27: % please place your own definitions here and don't use \def but
28: % \newcommand{}{}
29: 
30: \newtheorem{theorem}{Theorem}%[section]
31: \newtheorem{lemma}[theorem]{Lemma}
32: \newtheorem{proposition}[theorem]{Proposition}
33: \newtheorem{corollary}[theorem]{Corollary}
34: \newtheorem{conjecture}[theorem]{Conjecture}
35: \newtheorem{definition}[theorem]{Definition}
36: \newtheorem{remark}[theorem]{Remark}
37: \newtheorem{claim}[theorem]{Claim}
38: 
39: \newcommand{\mc}[1]{{\mathcal #1}}   
40: \newcommand{\mf}[1]{{\mathfrak #1}}
41: \newcommand{\mb}[1]{{\mathbf #1}}
42: \newcommand{\bb}[1]{{\mathbb #1}}
43: 
44: \theoremstyle{definition}
45: \newtheorem*{cor}{Corollary}
46: \newtheorem{case}{Case}
47: 
48: %\renewcommand{\thecondition}{\Alph{con}}
49: 
50: \newcommand\R{{\mathbb R}}
51: \newcommand\Z{{\mathbb Z}}
52: \newcommand\N{{\mathbb N}}
53: \newcommand\B{{\mathfrak B}}
54: \newcommand\E{{\mathfrak E}}
55: \newcommand\D{{\mathfrak D}}
56: \renewcommand\L{{\mathfrak L}}
57: \newcommand\A{{\mathfrak A}}
58: \newcommand\cX{{\mathcal X}}
59: \newcommand\cR{{\mathcal R}}
60: %\newcommand\fO{{\mathfrak O}}
61: %\newcommand\fC{{\mathfrak C}}
62: \newcommand\fC{{\textsf{C}}}
63: \newcommand\fO{{\textsf{O}}}
64: 
65: \newcommand{\tx}[1]{{\mathbf #1}}
66: %\newcommand{\bb}[1]{{\mathbb #1}}
67: 
68: 
69: \newcommand\eps{\varepsilon}
70: \newcommand\kap{\varkappa}
71: \newcommand\la{\langle}
72: \newcommand\ra{\rangle}
73: \newcommand\noth{\varnothing}
74: 
75: %\DeclareMathOperator{\sgn}{sgn}
76: \newcommand{\dop}[1]{\overline{#1}}
77: 
78: 
79: 
80: 
81: %
82: % Insert the name of "your journal" with
83: \journalname{JSP}
84: %
85: \begin{document}
86: 
87: \title{Condensation for a fixed number\\ of independent random variables
88: %\thanks{Grants or other notes
89: %about the article that should go on the front page should be
90: %placed here. General acknowledgments should be placed at the end of the article.}
91: }
92: %\subtitle{Do you have a subtitle?\\ If so, write it here}
93: 
94: %Short form of title}        % if too long for running head
95: %\titlerunning{}
96: 
97: \author{Pablo A.~Ferrari \and Claudio Landim \and Valentin~V.~Sisko}
98: 
99: %\authorrunning{%Short form of author list} % if too long for running head
100:   
101: \institute{Pablo A.~Ferrari  \at
102: IME USP, Caixa Postal 66281, 05315-970 - S\~{a}o Paulo, BRAZIL \\
103: \email{pablo@ime.usp.br} 
104: \and
105: Claudio Landim  \at
106: IMPA, Estrada Dona Castorina 110, CEP 22460-320 Rio de Janeiro, Brasil\\
107: CNRS UMR 6085, Universit\'e de Rouen, UMR 6085, Avenue de
108: l'Universit\'e, BP.12, Technop\^ole du Madrillet, F76801
109: Saint-\'Etienne-du-Rouvray, France  \\
110: \email{landim@impa.br} 
111: \and
112: Valentin V.~Sisko \at
113: IMPA, Estrada Dona Castorina 110, CEP 22460-320 Rio de Janeiro, Brasil\\
114: \email{valentin@impa.br}
115: } 
116: 
117: 
118: %\date{Received: date / Accepted: date}
119: % The correct dates will be entered by the editor
120: 
121: 
122: \maketitle
123: 
124: \begin{abstract}
125:  A family of $m$ independent identically distributed random variables indexed by
126: a chemical potential $\varphi\in[0,\gamma]$ represents piles of particles. As
127: $\varphi$ increases to $\gamma$, the mean number of particles per site converges
128: to a maximal density $\rho_c<\infty$. The distribution of particles conditioned
129: on the total number of particles equal to $n$ does not depend on $\varphi$
130: (canonical ensemble). 
131: For fixed $m$, as $n$ goes to infinity the canonical ensemble measure
132: behave as follows: removing the site with the maximal number of particles, the
133: distribution of particles in the remaining sites converges to the grand
134: canonical measure with density $\rho_c$; the remaining particles concentrate
135: (condensate) on a single site.
136: \keywords{condensation \and critical
137: density}
138: % \PACS{PACS code1 \and PACS code2 \and more}
139: \subclass{MSC 60K35 \and MSC 82C22}
140: \end{abstract}
141: 
142: \section{Introduction}
143: 
144: Condensation phenomena appears in many physical systems. 
145: From a physical point
146: of view, in Bose-Einstein condensation a large fraction of the atoms
147: collapses
148: into the lowest quantum state, which is possible to observe macroscopically.
149: From a mathematical point of view it can be seen as the spontaneous migration of
150: a macroscopic number of particles to some region of the space.  
151: A physical
152: account of different models and discussion may be found in the recent
153: articles:
154: Evans \cite{E2000}, Jeon and March \cite{JM2000}, Jeon, March and Pittel
155: \cite{JMP2000}, Godr\`eche \cite{G2003}, Grosskinsky, Sch\"utz and Spohn
156: \cite{GSS2003}, Evans, Majumdar and Zia \cite{EMSZ2004,EMSZ2006}, Evans and
157: Hanney \cite{EH2005}, Majumdar, Evans and Zia \cite{MEZ2005} and Godr\`eche and
158: Luck \cite{GL2005}.
159: 
160: These references consider the following mathematical model. Fix a natural number
161: $m$. The state space consists of particle configurations $\xi\in\N^m$; $\xi_i$
162: represents the number of particles at site $i\in\{1,\dots,m\}$. 
163: Consider a family $\mu_\varphi$ of product measures on $\N^m$ indexed
164: by a chemical potential $\varphi$ with range in $[0,\gamma]$ for some
165: $\gamma>0$.  The density $\rho = R(\varphi)$ (mean number of particles
166: per site) is an increasing function of $\varphi$. We assume that the
167: maximal density is finite: $\rho_c = R(\gamma)<\infty$.
168: 
169: For an integer $n\ge 0$, denote by $\nu_n$ the canonical measure
170: associated to the family $\mu_\varphi$. 
171: This is the measure
172: $\mu_\varphi$ conditioned on the hyperplane of configurations with $n$
173: particles: $\nu_n (\xi) = \mu_\varphi(\xi \mid \sum_i \xi_i = n)$.
174: In this context, condensation means that all but a few particles
175: concentrate on the same site. We prove in this article such a
176: statement under some conditions on the product measure $\mu_\varphi$.
177: More precisely, we show that if we remove the site with the largest
178: number of particles, the projection of the canonical measure on the
179: remaining sites converges to the product measure with maximal density
180: $\rho_c$, when the total number of sites $m$ is
181: fixed and the total number $n$ of particles increases to infinity.
182: 
183: 
184: \section{Notation and result}
185: \label{s:mtp}
186: 
187: Denote by $\bb N$ the nonnegative integers and let $f : \bb N \to \bb
188: R_+$ be a positive real function such that
189: \begin{eqnarray}
190: \label{a2}
191: \lim_{n\to\infty} \frac{f(n)}{f(n+1)} \;=\; \gamma \;, \quad
192: \sum_{n\ge 0} \gamma^n f(n) \;<\; \infty
193: \end{eqnarray}
194: for some $\gamma >0$. Assume also that $\gamma ^n f(n)$ is decreasing
195: and that there exists a function $C : \bb N\to\bb R$ such that
196: \begin{equation}
197: \label{a1}
198: \gamma ^{n/m} f(n/m) \;\le\; C (m) \gamma^{n} f(n)
199: \end{equation}
200: for all $n$. $f(n) = n^{-\alpha}$ with $\alpha > 1$, for instance,
201: satisfies \eqref{a2}, \eqref{a1}. 
202: 
203: We consider a family of product measures in the state space $\bb N^m$.
204: Configurations in this space are denoted by the Greek letter $\xi = (\xi_1 ,
205: \dots , \xi_m)$, with $\xi_j$ in $\bb N$ for $1\le j\le m$.  Let $Z:\bb R_+ \to
206: \bb R_+$ the partition function defined by
207: \begin{equation*}
208: Z(\varphi) \;=\; \sum_{n\ge 0} \varphi^n f(n)\;.
209: \end{equation*}
210: It follows from \eqref{a2} that the radius of convergence of $Z$ is
211: equal to $\gamma$: $Z(\varphi)<\infty$ if and only if $\varphi\le
212: \gamma$. For $0\le \varphi\le \gamma$, denote by $\mu_{m,\varphi}$ the
213: grand canonical measure on $\bb N^m$ given by
214: \begin{equation*}
215: \mu_{m,\varphi} (\xi) \;=\; \frac{1}{Z(\varphi)^m}
216: \prod_{j=1}^m \varphi^{\xi_j} f(\xi_j)\;.
217: \end{equation*}
218: 
219: Denote by $|\xi|$ the total number of particles of the configuration:
220: $|\xi| = \xi_1 + \cdots + \xi_m$ and by $\Sigma_{m,n}$, $n\ge 1$, the
221: subspace of configurations of $\bb N^m$ with $n$ particles:
222: \begin{equation}
223: \label{a3}
224: \Sigma_{m,n} \;=\; \{\xi\in\N^m \,:\, |\xi|=n\}\;.
225: \end{equation}
226: For $n\ge 0$, let $\nu_{m,n}$ be the canonical measure concentrated on
227: configurations with $n$ particles:
228: \begin{equation*}
229: \nu_{m,n}(\xi)\;=\; \mu_{m,\varphi} (\xi\,|\,|\xi|=n) \;.
230: \end{equation*}
231: Notice that the right hand side does not depend on the parameter
232: $\varphi$. 
233: 
234: Denote by $\mc X_m$ the space of monotone configurations:
235: \begin{equation*}
236: \cX_m \;=\; \{\zeta \in \N^m:\zeta(1)\le\dots\le\zeta(m)\}\;.
237: \end{equation*}
238: Configurations of $\mc X_m$ are denoted by the Greek letter $\zeta$.  Define the
239: order operator $\fO : \bb N^m \to \mc X_m$ which takes a configuration of $\bb
240: N^m$ to the monotone configuration by reordering appropriately the coordinates:
241: \begin{eqnarray*}
242: \!\!\!\!\!\!\!\!\!\!\!\!\!\!\! &&
243: (\fO \xi)_i \;\le\; (\fO \xi)_j \quad \text{for $1\le i\le j\le m$} \\
244: \!\!\!\!\!\!\!\!\!\!\!\!\!\!\! && \quad 
245: \text{and } (\fO \xi)_i = \xi_{\sigma(i)} \quad\text{for 
246: a permutation $\sigma$ of the set $\{1,\dots ,m\}$} \; .
247: \end{eqnarray*}
248: Define the cut operator $\fC:\bb N^m \to \bb N^{m-1}$ which eliminates
249: the last coordinate:
250: \[
251: (\fC \xi)_i  = \xi_i \quad \text{for $1\le i\le m-1$.}
252: \]
253: 
254: These operators induce two measures: $\hat \mu_{m, \varphi}$ defined on
255: $\cX_{m}$ and $\hat\nu_{m-1,n}$ defined on $\cX_{m-1}$:
256: \begin{equation*}
257: \hat \mu_{m,\varphi} \;=\; \mu_{m,\varphi}  \fO^{-1} \; , \qquad  
258: \hat \nu_{m-1,n} = \nu_{m,n} (\fC\circ\fO)^{-1} \;.
259: \end{equation*}
260: To obtain a sample of $\hat \mu_{m, \varphi}$, sample a configuration with the
261: product measure $\mu_{m,\varphi}$ and order it.  To obtain a sample of
262: $\hat\nu_{m-1,n}$, sample a configuration with the product measure
263: $\mu_{m,\varphi}$ conditioned to have $n$ particles (i.e.  $\nu_{m,n}$), order
264: it and cut the last coordinate.
265: 
266: Next we state our result. It asserts that if the total number of sites $m$ is
267: fixed and the total number $n$ of particles increases to infinity, the canonical
268: measure concentrated on configurations with $n$ particles, from which we removed
269: the site with the largest number of particles, converges to the ordered
270: \emph{maximal} grand canonical measure. Maximal means that we set $\varphi$ to
271: be $\gamma$.
272: 
273: By symmetry the position of the maximum site is uniformly distributed in
274: $\{1,\dots,m\}$. Sampling from the ordered measure $\hat\nu_{m-1,n}$ and
275: reordering randomly and uniformly the labels of the particles we obtain a
276: measure which approachs the maximal product measure.
277: 
278: \begin{theorem}
279: \label{t1}
280: Assume that f is a nonincreasing function satisfying hypotheses \eqref{a2} and
281: \eqref{a1}. For each $m\ge 2$, the sequence of probability measures
282: $\hat\nu_{m-1,n}$ converges weakly to $\hat \mu_{m-1, \gamma}$ as $n$ tends to
283: infinity.
284: \end{theorem}
285: 
286: \begin{remark}
287:   \label{t2} {\rm A similar result has been proven in various of the quoted
288:     references in the thermodynamical limit. See for instance
289:     \cite{GSS2003}, where the the number of particles $n$ divided the
290:     number of sites $m=m(n)$ converges to a constant $\rho>\rho_c$ as
291:     $n$ goes to infinity.  Theorem \ref{t1} shows that this phenomenon
292:     is just a combinatorial fact that can be observed already without
293:     making the number of sites grow to infinity.  }
294: \end{remark}
295: 
296: \begin{remark}
297: \label{t3}
298: {\rm The condition that $f$ is decreasing can be relaxed. We can replace
299: this condition and \eqref{a1} by the hypothesis that for every $m\ge 1$,
300: there exists a finite constant $C_1 (m)$ such that 
301: \begin{equation}
302: \label{a6}
303: \frac{\max_{n/m \le k \le n} \gamma ^k f (k)}
304: {\min_{n/m \le k \le n} \gamma ^k f (k)} \;\le\; C_1(m)
305: \end{equation}
306: for all $n\ge 1$.}
307: \end{remark}
308: 
309: \begin{remark}
310:   \label{t4} 
311:   
312:   {\rm Relation with the zero range process. Fix $m\ge 1$, a
313:     irreducible symmetric transition probability $p(x,y)$ on $\{1,
314:     \dots, m\}$ and a positive function $g: \bb N\to\bb R_+$. The
315:     zero-range process associated to $(p,g)$ can be informally
316:     described as follows. Particles evolve on $\{1, \dots, m\}$. If
317:     there are $k$ particles at some site $x$ of $\{1, \dots, m\}$ one
318:     of them jumps to site $y$ at rate $g(k) p(x,y)$, independently
319:     from what happens at the other sites.
320:     
321:     Let $f(k)^{-1} = g(1) \cdots g(k)$ and assume that $f$ satisfies
322:     assumption \eqref{a2} for some $\gamma >0$.  It is well known
323:     \cite{KL1999} that the product measures $\{\mu_{m,\varphi} : 0\le
324:     \varphi\le \gamma\}$ defined above are invariant for the symmetric
325:     zero-range process.
326:     
327:     Let $R^* = \sum_{n\ge 0} n \gamma^n f(n)$ and assume that
328:     $R^*<\infty$. The function $R:[0,\gamma] \to [0, R^*]$ defined by
329:     $R(\varphi) = \mu_{m,\varphi} [\xi_1]$ gives the density of
330:     particles under the invariant measure $\mu_{m,\varphi}$. A simple
331:     computation of $R'$ shows that $R$ is bijective. Since in the
332:     symmetric case any product invariant measure belongs to the set
333:     $\{\mu_{m,\varphi} : 0\le \varphi\le \gamma\}$, there are no
334:     product invariant measures with density above $R^*$.
335: 
336:     Since the process is ergodic, for every $n\ge 1$, there exists a unique
337:     stationary measure, denoted above by $\nu_{m,n}$. Order the configuration
338:     and remove the site with the largest number of particles.  Under the
339:     additional assumption \eqref{a6}, Theorem \ref{t1} states that this measure
340:     converges, as $n\uparrow\infty$, to the ordered maximal product state.  In
341:     particular, all but a finite number of particles tend to accumulate in one
342:     site.}
343: \end{remark}
344: 
345: 
346: 
347: 
348: \section{Proof of Theorem \ref{t1}}
349: 
350: For an ordered configuration $\zeta$ belonging to $\cX_{m}$ (resp.
351: $\eta$ belonging to
352: $\cX_{m-1}$), let $K_m(\zeta)$ (resp.  $K_{m-1,n}(\eta)$) be the
353: number of configurations in $\bb N^m$ whose ordering (resp. ordering
354: and cutting) gives $\zeta$ (resp. $\eta$):
355: \begin{eqnarray*}
356: \!\!\!\!\!\!\!\!\!\!\!\!\!\! &&
357: K_m(\zeta) \;=\; \sharp\{\xi \in \N^{m}:\fO(\xi)=\zeta\}\; , \\
358: \!\!\!\!\!\!\!\!\!\!\!\!\!\! && \quad
359: K_{m-1,n}(\eta) \;=\; \sharp\{\xi \in \Sigma_{m,n}:
360: \fC\circ\fO(\xi)=\eta\}\;. 
361: \end{eqnarray*}
362: Note that the number of elements of an empty set is zero.  Since
363: $\mu_{m,\varphi}$, $\nu_{m,n}$ are invariant by permutation, for any
364: $\zeta$ in $\cX_{m}$, $\eta$ in $\cX_{m-1}$,
365: \begin{eqnarray}
366: \label{a4}
367: \!\!\!\!\!\!\!\!\!\!\!\!\!\! &&
368: \hat \mu_{m, \varphi} (\zeta) \;=\; K_m(\zeta) \, \mu_{m, \varphi}
369: (\zeta)\; , \\
370: \!\!\!\!\!\!\!\!\!\!\!\!\!\! &&
371: \quad \hat \nu_{m-1,n}(\eta) \;=\;  K_{m-1,n}(\eta) \,
372: \nu_{m,n}(\eta_1, \dots, \eta_{m-1}, n - |\eta|)\;.
373: \nonumber
374: \end{eqnarray}
375: 
376: \begin{lemma}
377: \label{lm:g}
378: For any  $\eta$ in $\cX_{m-1}$ and any $n\ge 1$, $K_{m-1,n} (\eta) \le  m
379: K_{m-1} (\eta)$. Moreover, for any $\eta$ in $\cX_{m-1}$,
380: \begin{equation*}
381: \lim_{n\to\infty} K_{m-1,n} (\eta) \;=\;  m K_{m-1} (\eta)\;.
382: \end{equation*}
383: \end{lemma}
384:                           
385: The proof is elementary. A fixed ordered configuration $(\eta_1 ,
386: \dots , \eta_{m-1})$ belonging to $\cX_{m-1}$ arises from a permutation of the
387: coordinates $\eta_j$. To compute $K_{m-1,n}$, we add the coordinate
388: $n - |\eta|$.  This extra coordinate accounts for the factor $m$ which
389: corresponds to all its possible positions. If $n$ is large this last
390: coordinate is different from all others and provides $m$ distinct
391: vectors.
392: \medskip
393:                                                                         
394: We are now in position to prove Theorem \ref{t1}. Fix $\eta$ in
395: $\cX_{m-1}$. In view of \eqref{a4},
396: \begin{equation*}
397: \hat \nu_{m-1,n} (\eta) \;=\; 
398: \frac{K_{m-1,n} (\eta) f (\eta_1 ) \cdots  f (\eta_{m-1} ) f (n - |\eta|)}
399: {\sum_{\eta' \in \cX_{m-1}} K_{m-1,n} (\eta') f (\eta'_1 ) \cdots  f
400:   (\eta'_{m-1} ) f (n - |\eta'|) }\;\cdot
401: \end{equation*}
402: By \eqref{a4} and Lemma \ref{lm:g}, to prove Theorem \ref{t1}, we only
403: need to show that
404: \begin{eqnarray}
405: \label{a5}
406: \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\! &&
407: \lim_{n\to\infty} \sum_{\eta' \in \cX_{m-1}} K_{m-1,n} (\eta') 
408: f (\eta'_1 ) \cdots  f  (\eta'_{m-1} ) \frac{f (n - |\eta'|)}
409: {f (n - |\eta|)} \\
410: \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\! && \quad
411: =\; \sum_{\eta' \in \cX_{m-1}} m K_{m-1} (\eta') 
412: \gamma^{|\eta'| - |\eta|} f (\eta'_1 ) \cdots  f  (\eta'_{m-1} )\;.
413: \nonumber
414: \end{eqnarray}
415: 
416: Fix a positive constant $M > 1$. By assumption \eqref{a2} and by Lemma
417: \ref{lm:g}, for configurations such that $|\eta| \le M$ , $\lim_n
418: K_{m-1,n} (\eta) = m K_{m-1} (\eta)$ and 
419: \[
420: \lim_{n \to \infty} f (n - |\eta|)/f (n -
421: |\eta'|) = \gamma^{|\eta'| - |\eta|}.
422: \]
423: In particular, for every $M\ge 1$,
424: \begin{eqnarray*}
425: \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\! &&
426: \lim_{n\to\infty} 
427: \sum_{\substack{\eta' \in \cX_{m-1} \\ |\eta'|\le M}} 
428: K_{m-1,n} (\eta') f (\eta'_1 ) \cdots  f  (\eta'_{m-1} ) 
429: \frac{f (n - |\eta'|)} {f (n - |\eta|)} \\
430: \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\! && \quad
431: =\; \sum_{\substack{\eta' \in \cX_{m-1} \\ |\eta'|\le M}} 
432: m K_{m-1} (\eta') \gamma^{|\eta'| - |\eta|} f (\eta'_1 ) 
433: \cdots  f (\eta'_{m-1} )\;.
434: \end{eqnarray*}
435: 
436: To estimate the sum $|\eta'| > M$, recall from Lemma \ref{lm:g} that 
437: \[
438: K_{m-1,n} (\eta') \le m K_{m-1} (\eta').
439: \]
440: On the other hand, since
441: \[
442: \eta_m = n - |\eta|\ge \max_{1\le j \le m-1} \eta_j, \quad
443: \eta'_m = n - |\eta'|\ge \max_{1\le j \le m-1} \eta'_j,
444: \]
445:  we have that $n/m\le
446: \eta_m$ , $\eta'_m\le n$. Thus, by assumption \eqref{a6}, 
447: \begin{equation*}
448:   \frac{f (n - |\eta'|)}{f (n - |\eta |)} \;\le\; C_1(m) 
449: \gamma^{|\eta'| - |\eta|}.
450: \end{equation*}
451: Therefore,
452: \begin{eqnarray*}
453: \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\! &&
454: \sum_{\substack{\eta' \in \cX_{m-1} \\ |\eta'|> M}} 
455: K_{m-1,n} (\eta') f (\eta'_1 ) \cdots  f  (\eta'_{m-1} ) 
456: \frac{f (n - |\eta'|)} {f (n - |\eta|)} \\
457: \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\! && \quad
458: \le \; C(m) \sum_{\substack{\eta' \in \cX_{m-1} \\ |\eta'|> M}} 
459: K_{m-1} (\eta') \gamma^{|\eta'| - |\eta|} f (\eta'_1 ) 
460: \cdots  f (\eta'_{m-1} )\;.
461: \end{eqnarray*}
462: This expression vanishes as $M\uparrow\infty$ because the sum without
463: the constraint $|\eta'|> M$ is finite, being equal to
464: $\gamma^{-|\eta|} Z(\gamma)^{m-1}$, where $Z(\gamma)$ is the
465: partition function introduced in Section \ref{s:mtp}.  This concludes
466: the proof of the theorem.
467: 
468: 
469: 
470: % [2] M. R. Evans, Braz. J. Phys. 30, 42 (2000).
471: % [3] M. R. Evans, Europhys. Lett. 36, 13 (1996).
472: % [4] P. Bialas, Z. Burda, and D. Johnston, Nucl. Phys. B493,
473: % 505 (1997).
474: % [5] O. J. O\u2019Loan, M. R. Evans, and M. E. Cates, Phys. Rev. E
475: % 58, 1404 (1998).
476: % [6] S. N. Majumdar, S. Krishnamurthy, and M. Barma, Phys.
477: % Rev. Lett. 81, 3691 (1998).
478: % [7] D. Chowdhury, L. Santen, and A. Schadschneider, Phys.
479: % Rep. 329, 199 (2000).
480: % [8] S. N. Coppersmith, C.-h. Liu, S. Majumdar, O. Narayan,
481: % and T. A. Witten, Phys. Rev. E 53, 4673 (1996).
482: % [9] D. van der Meer et al., J. Stat. Mech. Theor. Exp. 04
483: % (2004) P04004; J. Torok, cond-mat/0407567.
484: % 10] S. N. Dorogovtsev and J. F. F. Mendes, Evolution of
485: % Networks (OUP, Oxford, 2003).
486: % 11] 
487: % 12] 
488: % 13]
489: % 37, L275 (2004);  L10001.
490: % 14] J. Krug and J. Garcia, J. Stat. Phys. 99, 31 (2000); R.
491: % Rajesh and S. N. Majumdar, J. Stat. Phys. 99, 943 (2000).
492: % 15] M. R. Evans, S. N. Majumdar, and R. K. P. Zia (unpublished).
493: % 16] M. R. Evans, D. P. Foster, C. Godreche, and D. Mukamel,
494: % Phys. Rev. Lett. 74, 208 (1995).
495: % 17] F. Zielen and A. Schadschneider, Phys. Rev. Lett. 89,
496: % 090601 (2002).
497: % 18] Y. Kafri, E. Levine, D. Mukamel, G. M. Schu¨tz, and J.
498: % To¨ro¨k, Phys. Rev. Lett. 89, 035702 (2002).
499: % 4
500: 
501: 
502: 
503: \begin{acknowledgements}
504: The authors are grateful to CNPq, FAPERJ, FAPESP, and PRONEX for financial support. 
505: %If you'd like to thank anyone, place your comments here
506: %and remove the percent signs.
507: \end{acknowledgements}
508: 
509: %\bibliographystyle{spmpsci}      % mathematics and physical sciences
510: %\bibliography{fls}   % name your BibTeX data base
511: 
512: \begin{thebibliography}{10}
513: \providecommand{\url}[1]{{#1}}
514: \providecommand{\urlprefix}{URL }
515: \expandafter\ifx\csname urlstyle\endcsname\relax
516:   \providecommand{\doi}[1]{DOI~\discretionary{}{}{}#1}\else
517:   \providecommand{\doi}{DOI~\discretionary{}{}{}\begingroup
518:   \urlstyle{rm}\Url}\fi
519: 
520: \bibitem{E2000}
521: Evans, M.R.: Phase transitions in one-dimensional nonequilibrium systems.
522: \newblock Braz. J. Phys. \textbf{30}(1), 42--47 (2000)
523: 
524: \bibitem{EH2005}
525: Evans, M.R., Hanney, T.: Nonequilibrium statistical mechanics of the zero-range
526:   process and related models.
527: \newblock J. Phys. A \textbf{38}(19), R195--R240 (2005)
528: 
529: \bibitem{EMSZ2004}
530: Evans, M.R., Majumdar, S.N., Zia, R.K.P.: Factorized steady states in mass
531:   transport models.
532: \newblock J. Phys. A \textbf{37}(25), L275--L280 (2004)
533: 
534: \bibitem{EMSZ2006}
535: Evans, M.R., Majumdar, S.N., Zia, R.K.P.: Canonical analysis of condensation in
536:   factorised steady states.
537: \newblock J. Stat. Phys. \textbf{123}(2), 357--390 (2006)
538: 
539: \bibitem{G2003}
540: Godr{\`e}che, C.: Dynamics of condensation in zero-range processes.
541: \newblock J. Phys. A \textbf{36}(23), 6313--6328 (2003)
542: 
543: \bibitem{GL2005}
544: Godr{\`e}che, C., Luck, J.M.: Dynamics of the condensate in zero-range
545:   processes.
546: \newblock J. Phys. A \textbf{38}(33), 7215--7237 (2005)
547: 
548: \bibitem{GSS2003}
549: Gro{\ss}kinsky, S., Sch{\"u}tz, G.M., Spohn, H.: Condensation in the zero range
550:   process: stationary and dynamical properties.
551: \newblock J. Statist. Phys. \textbf{113}(3-4), 389--410 (2003)
552: 
553: \bibitem{JM2000}
554: Jeon, I., March, P.: Condensation transition for zero range invariant measures.
555: \newblock In: Stochastic models (Ottawa, ON, 1998), \emph{CMS Conf. Proc.},
556:   vol.~26, pp. 233--244. Amer. Math. Soc., Providence, RI (2000)
557: 
558: \bibitem{JMP2000}
559: Jeon, I., March, P., Pittel, B.: Size of the largest cluster under zero-range
560:   invariant measures.
561: \newblock Ann. Probab. \textbf{28}(3), 1162--1194 (2000)
562: 
563: \bibitem{KL1999}
564: Kipnis, C., Landim, C.: Scaling limits of interacting particle systems,
565:   \emph{Grundlehren der Mathematischen Wissenschaften [Fundamental Principles
566:   of Mathematical Sciences]}, vol. 320.
567: \newblock Springer-Verlag, Berlin (1999)
568: 
569: \bibitem{MEZ2005}
570: {Majumdar}, S.N., {Evans}, M.R., {Zia}, R.K.: Nature of the condensate in mass
571:   transport models.
572: \newblock Phys. Rev. Lett. \textbf{94}(18), 180,601--+ (2005).
573: \newblock \doi{10.1103/PhysRevLett.94.180601}
574: 
575: \end{thebibliography}
576: 
577: \end{document}
578: % end of file template.tex
579: 
580: