We prove a conjecture raised by the work of Diaconis and Shahshahani (1981) about the mixing time of random walks on the permutation group induced by a given conjugacy class. To do this we exploit a connection with coalescence and fragmentation proce
sses and control the Kantorovitch distance by using a variant of a coupling due to Oded Schramm. Recasting our proof in the language of Ricci curvature, our proof establishes the occurrence of a phase transition, which takes the following form in the case of random transpositions: at time $cn/2$, the curvature is asymptotically zero for $cle 1$ and is strictly positive for $c>1$.
We introduce a Gibbs measure on nearest-neighbour paths of length $t$ in the Euclidean $d$-dimensional lattice, where each path is penalised by a factor proportional to the size of its boundary and an inverse temperature $beta$. We prove that, for al
l $beta>0$, the random walk condensates to a set of diameter $(t/beta)^{1/3}$ in dimension $d=2$, up to a multiplicative constant. In all dimensions $dge 3$, we also prove that the volume is bounded above by $(t/beta)^{d/(d+1)}$ and the diameter is bounded below by $(t/beta)^{1/(d+1)}$. Similar results hold for a random walk conditioned to have local time greater than $beta$ everywhere in its range when $beta$ is larger than some explicit constant, which in dimension two is the logarithm of the connective constant.
We show that an infinite Galton-Watson tree, conditioned on its martingale limit being smaller than $eps$, agrees up to generation $K$ with a regular $mu$-ary tree, where $mu$ is the essential minimum of the offspring distribution and the random vari
able $K$ is strongly concentrated near an explicit deterministic function growing like a multiple of $log(1/eps)$. More precisely, we show that if $muge 2$ then with high probability as $eps downarrow 0$, $K$ takes exactly one or two values. This shows in particular that the conditioned trees converge to the regular $mu$-ary tree, providing an example of entropic repulsion where the limit has vanishing entropy.
