Joint distribution of the first and second eigenvalues at the soft edge of unitary ensembles (Q2839495)

When considering a Gaussian unitary ensemble (a special type of the random matrix model), after some scaling, expressions of the limit joint distribution of the first and second smallest eigenvalues are known, when the dimension of the matrix goes to infinity. It is also known that after a scaling limit one can pass to the largest and second largest eigenvalues precisely from the first and second ones. The authors thus perform this task to derive an expression for the joint density of the largest and second largest eigenvalues. The procedure involves writing the joint densities in terms of two functions that come from something called the associated isomonodromic problem for certain Painlevé equations, and these functions are solutions of a Lax system of differential equations. An important tool in the paper is Okamoto's theory to deal with the scaling limits of the Painlevé equations. At the end, the authors also give a table with the values of the probability density of the spacing between the two largest eigenvalues.