Long-term behaviour conjecture for interacting reflected OU processes on graphs

About 1 year old · traced to

Let G=(V,E)G=(V,E) be a graph with ∣V∣≥2|V|\geq 2, let XX be the limiting reflected diffusion from Theorem~, and let ν(G)\nu(G) denote the principal eigenvalue of GG. The parameters are α\alpha and β\beta.

Long-term behaviour conjecture. 1) If α<0\alpha<0 and α+βν(G)<0\alpha+\beta\nu(G)<0, then the process XX is positive recurrent with the stationary distribution μ\mu defined in Theorem~. 2) If α=0\alpha=0, β≤0\beta\leq 0 and ∣V∣=2|V|=2, then the process XX is null recurrent. 3) If either α>0\alpha>0, or α<0\alpha<0 and α+ν(G)β≥0\alpha+\nu(G)\beta\geq 0, then the process XX is transient.

This conjecture concerns the long-term behaviour of the limiting process for a graph with at least two vertices. For a single vertex, the limit is a reflected one-dimensional OU process, whose recurrence properties are known; when β=0\beta=0, the multidimensional process consists of independent reflected OU processes. The interacting case is presented as the problem of determining the collective long-term behaviour.

References

Primary source

Anatolii Puhalskii and Vadim Shcherbakov, “A diffusion limit for Markov chains with log-linear interaction on a graph”, arXiv:2501.09323 (2025).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.