Local density of states distribution and multifractal eigenvectors of weighted random networks via the cavity approach

Abstract

We study the local density of states (LDoS) distribution of a general class of weighted Erdős-Rényi graphs. Using the cavity method, we obtain a good approximation to the full LDoS distribution and compact expressions for its power-law tails, which we show to have exponent $3$ in the extended phase. We deduce that the eigenvectors in the continuous part of the spectrum are extended but (weakly) multifractal, and we extract expressions for the associated fractal dimensions and the singularity spectrum. We also demonstrate that the inverse participation ratio in this multifractal phase exhibits an unusual logarithmic scaling with system size, which is neither fully-extended nor localised by the usual definitions. Finally, we verify that some symmetry properties (derived from the non-linear sigma model), which have been shown to hold for many systems exhibiting multifractality, also hold in our case, both for the LDoS distribution and the singularity spectrum.