English

Geometry and Dynamics for Hierarchical Regular Networks

Disordered Systems and Neural Networks 2008-07-23 v2

Abstract

The recently introduced hierarchical regular networks HN3 and HN4 are analyzed in detail. We use renormalization group arguments to show that HN3, a 3-regular planar graph, has a diameter growing as \sqrt{N} with the system size, and random walks on HN3 exhibit super-diffusion with an anomalous exponent d_w = 2 - \log_2\phi = 1.306..., where \phi = (\sqrt{5} + 1)/2 = 1.618... is the "golden ratio." In contrast, HN4, a non-planar 4-regular graph, has a diameter that grows slower than any power of N, yet, fast than any power of \ln N . In an annealed approximation we can show that diffusive transport on HN4 occurs ballistically (d_w = 1). Walkers on both graphs possess a first- return probability with a power law tail characterized by an exponent \mu = 2 -1/d_w . It is shown explicitly that recurrence properties on HN3 depend on the starting site.

Cite

@article{arxiv.0805.3013,
  title  = {Geometry and Dynamics for Hierarchical Regular Networks},
  author = {S. Boettcher and B. Goncalves and J. Azaret},
  journal= {arXiv preprint arXiv:0805.3013},
  year   = {2008}
}

Comments

15 pages, revtex; published version; find related material at http://www.physics.emory.edu/faculty/boettcher/

R2 v1 2026-06-21T10:42:22.161Z