We give a proof of the Universality Conjecture for orthogonal and symplectic ensembles of random matrices in the scaling limit for a class of weights w(x)=exp(-V(x)) where V is a polynomial, V(x)=kappa_{2m}x^{2m}+..., kappa_{2m}>0. For such weights the associated equilibrium measure is supported on a single interval. The precise statement of our results is given in Theorem 1.1 below. For a proof of the Universality Conjecture for unitary ensembles, for the same class of weights, see [DKMVZ2]. Our starting point is Widom's representation [W] of the orthogonal and symplectic correlation kernels in terms of the kernel arising in the unitary case plus a correction term which is constructed out of derivatives and integrals of orthonormal polynomials (OP's) {p_j(x)}, j=0,1,..., with respect to the weight w(x). The calculations in [W] in turn depend on the earlier work of Tracy and Widom [TW2]. It turns out (see [W] and also Theorems 2.1 and 2.2 below) that only the OP's in the range j=N+O(1), N->infinity, contribute to the correction term. In controlling this correction term, and hence proving Universality for both the orthogonal and symplectic cases, the uniform Plancherel--Rotach type asymptotics for the OP's found in [DKMVZ2] play an important role, but there are significant new analytical difficulties that must be overcome which are not present in the unitary case. We note that we do not use skew orthogonal polynomials.
Read full abstract