The mean-field zero-range process with unbounded monotone rates: mixing time, cutoff, and Poincaré constant (Q6104011)

A \textit{mean-field zero-range process} is an irreducible continuous-time Markov chain which describes the evolution of \(m\) \textit{particles} jumping randomly across \(n\) \textit{sites}, such that the movements of these particles are governed by a potential function \(r:\{1,2,\ldots \}\to(0,\infty)\). Thus, \(r(k)\) indicates the rate at which a site with \(k\) particles expels a particle that will choose its new destination uniformly among all sites. The present paper highlights some refined asymptotic results in the regime \(n\to\infty\), under the condition that \(m/n\) is bounded. The novelty of this study is ensured by the consideration of a potential function \(r\) that increases to infinity with a sub-linear speed. The speed of convergence of the process, starting from an initial distribution, to the stationary distribution is quantified by the \(\varepsilon\)-\textit{mixing time} (\(\varepsilon >0\)), which is defined in terms of the total variation distance between the two distributions. The main result of the paper expresses an asymptotic upper bound for the mixing times, which involves the function \(\displaystyle R(k)=\sum_{i=1}^k\frac{1}{r(i)},\ k\ge 1.\) The cutoff phenomenon is then deduced at time \(R(m)\). The author also shows that the associated Poincaré constant is bounded. The \textit{path-coupling method} is used for some proofs. \par The work is an important contribution to the study of the behavior of a zero-range process when the number of its sites tends to infinity.