The Selberg correlation integrals are averages of the products $\prod_{ s=1}^m\prod_{ l=1}^n (x_s - z_l)^{ \mu_s}$ with respect to the Selberg density. Our interest is in the case $m=1$, $\mu_1 = \mu$, when this corresponds to the $\mu$-th moment of the corresponding characteristic polynomial. We give the explicit form of a $(n+1) \times (n+1)$ matrix linear difference system in the variable $\mu$ which determines the average, and we give the Gauss decomposition of the corresponding $(n+1) \times (n+1)$ matrix. For $\mu$ a positive integer the difference system can be used to efficiently compute the power series defined by this average.
Categories: math-ph math.MP
