Baronchelli, A. and Loreto, V. (2006). Ring structures and mean first passage time in networks. Physical Review E  Statistical, Nonlinear, and Soft Matter Physics, 73(2), doi: 10.1103/PhysRevE.73.026103
Abstract
In this paper we address the problem of the calculation of the mean first passage time on generic graphs. We focus in particular on the mean first passage time on a node s for a random walker starting from a generic, unknown, node x. We introduce an approximate scheme of calculation which maps the original process in a Markov process in the space of the socalled rings, described by a transition matrix of size O(ln N∕ln ⟨k⟩×ln N∕ln ⟨k⟩), where N is the size of the graph and ⟨k⟩ the average degree in the graph. In this way one has a drastic reduction of degrees of freedom with respect to the size N of the transition matrix of the original process, corresponding to an extremely low computational cost. We first apply the method to the ErdösRenyi random graphs for which the method allows for almost perfect agreement with numerical simulations. Then we extend the approach to the BarabásiAlbert graph, as an example of scalefree graph, for which one obtains excellent results. Finally we test the method with two realworld graphs, Internet and a network of the brain, for which we obtain accurate results.
Publication Type:  Article 

Subjects:  Q Science > QC Physics 
Departments:  School of Mathematics, Computer Science & Engineering > Mathematics 
URI:  http://openaccess.city.ac.uk/id/eprint/2670 

PDF
Download (898kB)  Preview 
Export
Downloads
Downloads per month over past year
Actions (login required)
Admin Login 