Abstract
For infinite Gaussian unitary ensemble random matrices the probability density function Snn(t) for the nearest neighbor eignenvalue spacing (as distinct from the spacing between consecutive eigenvalues) is computed in terms of the solution of a certain nonlinear equation, which generalizes the σ form of the Painlevé V equation. Comparison is made with the empirical value of Snn(t) for the zeros of the Riemann ζ function on the critical line, including data from 106 consecutive zeros near zero number 1020.

This publication has 8 references indexed in Scilit: