Loop-erased random walk on finite graphs and the Rayleigh process (Q927256)

The author considers the loop-erased random walk on large finite graphs \((G_{n})_{n\geq 1}\) with \(\lim_{n\to\infty}| G_{n}| =\infty\) and satisfying moreover some technical conditions. The goal is to study how the length \(Y_{t}\) of the loop-erased random walk on \(G_{n}\), after \(t\), steps evolves over time. Set \(Z_{n}(t)=\frac{b_{n}}{\sqrt{| G_{n}| }} Y_\lfloor a_{n}\sqrt{| G_{n}| }t]\rfloor\),\(t\geq0\), where \((a_{n}),(b_{n})\) are such that \(0<\inf a_{n}\leq\sup a_{n}<\infty\) and \(0<\inf b_{n}\leq\sup b_{n}<\infty\). Then \(Z_{n}\to R\), as \(n\to\infty\), in the Skorokhod topology, with \(R\) the Rayleigh process introduced by \textit{S. N. Evans, J. Pitman} and \textit{A. Winter} [Probab. Theory Relat. Fields 134, No. 1, 81--126 (2006; Zbl 1086.60050)]. One splits the proof into two steps. To prove the convergence of finite-dimensional distributions of \(Z_{n}(t)\) the author uses the strategy introduced by \textit{Y. Peres} and \textit{D. Revelle} [Electron. J. Probab. 9, Paper No. 26, 825--845, electronic only (2004; Zbl 1064.60095)] of splitting the random walk in short segments, the coupling between the loop-erased random walk and the Rayleigh process and some bounds for the Rayleigh process. Finally, one proves that the sequence of processes \((Z_{n})\) is relatively compact.