Percolation on interdependent one-dimensional long-range networks
π arXiv:2610.10097 Β· π₯ PDF Β· 2026-10-07 Β· cond-mat.stat-mech
Authors: Adar Sharabi [arXiv Β· scholar] , Amir Bashan [arXiv Β· scholar] , Sergey V. Buldyrev [arXiv Β· scholar] , Guy Amit [arXiv Β· scholar]
π Abstract
We study the emergence of a mutual giant component in a multiplex network comprised of two one-dimensional long-range networks on a lattice. The probability that two nodes on one lattice, with a distance $r$ between them, are connected with an edge falls algebraically with distance as $Cr^{-(1+Ο)}$, where $0 < Ο< 1$ and $0 < C \leq 1$. A mutual giant component exists if $C > C_c^{\mathrm{mul}}$, where $C_c^{\mathrm{mul}}(Ο)$ is a critical value of $C$ that is dependent on $Ο$. We find rigorous lower bounds on $C_c^{\mathrm{mul}}$, and a tree approximation whose estimate of $C_c^{\mathrm{mul}}$ agrees with simulations within 0.6% for $Ο\leq 0.5$. The multiplex network is significantly less stable than the single network case, as is evident by the larger value of $C_c^{\mathrm{mul}}$, which is up to 2.4 times the single-layer value. Our simulations indicate that the order of the transition depends on $Ο$. For $Ο\leq 0.35$ the mutual giant component collapses abruptly, in a first-order phase transition, while for $Ο\geq 0.4$ the transition is continuous. When the dependency links connect random pairs of nodes, the transition is first order for all $Ο$.