Markov's theorem revisited (Q1331078): Difference between revisions

Markov's theorem in the theory of orthogonal polynomials gives the asymptotic behaviour of the ratio \(P_{n-1}^{(1)}(z)/P_ n(z)\), where \(P_ n(z)\) are orthogonal polynomials on the real line with respect to a positive measure \(\mu\), and \(P_{n-1}^{(1)}(z)\) are the associated polynomials. Originally Markov gave the result when the support of \(\mu\) is contained in a compact interval \([a,b]\), in which case the limit is the Stieltjes transform \(\int_ a^ b d\mu(x)/(z- x)\) of the measure \(\mu\) and the asymptotic results hold uniformly on compacta avoiding the interval \([a,b]\). The author looks at various extensions of this results, in particular the case when the support of \(\mu\) is unbounded. Then the moment problem associated with the measure \(\mu\) may be determinate or indeterminate. For the indeterminate case the limit turns out to be one particular Nevanlinna extremal measure. The Stieltjes case \(\text{supp}(\mu)\subset [0,\infty)\) receives special attention. Finally, Markov's theorem is applied to the shifted moment problem, obtained by replacing the recurrence coefficients \(a_ n\) and \(b_ n\) for the orthogonal polynomials by \(a_{n+1}\) and \(b_{n+1}\).