A small network of nodes is built as a symmetric double star: two hub nodes are…

PHYSICS · Multi-part · Fermi University Physics · Thermal, Statistical & Complex Systems · Complex and collective systems · Complex and collective systems

Many Parts, One Behaviour, Ch Epidemics on networks

Problem

A small network of N=18N=18 nodes is built as a symmetric double star: two hub nodes are joined to each other by an edge, and each hub is also joined to 88 further leaf nodes that have no other connections. Every leaf therefore has degree 11 and each hub has degree 99. ((a)) Using the degree sequence and the giant-component argument of the random-graph chapter, applied here as the tree-like bond-percolation threshold, compute Tc=⟨k⟩/(⟨k2⟩−⟨k⟩)T_{\rm c} = \langle k\rangle/(\langle k^2\rangle - \langle k\rangle). This is the SIR threshold of a large configuration-model network with the same proportions of degrees, a comparison value rather than a threshold of the pictured 1818-node graph. ((b)) Using heterogeneous mean-field theory (annealed degree classes), compute the SIS threshold λc=⟨k⟩/⟨k2⟩\lambda_{\rm c} = \langle k\rangle/\langle k^2\rangle. ((c)) Using the symmetric eigenvalue problem for the adjacency matrix A\mathbf{A} (exploiting the network's symmetry to reduce the eigenvector equation to leaf and hub components), find the leading eigenvalue κ1\kappa_1, and hence give the quenched SIS threshold λc=1/κ1\lambda_{\rm c} = 1/\kappa_1. ((d)) State, in one sentence, which of the two SIS thresholds in (b) and (c) is larger and why degree-class theory, which knows only the degrees, misjudges this network.

Browse the problem library