The Levin-Golovin theorem for Sobolev spaces (Q5930785): Difference between revisions

Let \(\{\lambda_n\}\) be a set of zeros of an entire function \(F\) of one complex variable. The Levin-Golovin theorem describes conditions for function \(F\) that imply that the family of exponentials \(\{e^{i\lambda_n}\}\) is a Riesz basis in \(L_2(0,a)R\). The authors regard the similar problem for Sobolev spaces \(H^s_2(0,a)\). They proved that if \(\{\lambda_n\}\) are simple zeros lying in the strip, \(|\text{Im }z|< \text{const}\), \(|F(x+ ih)|\sim(1+|x|)^s\) for some fixed \(h\) and \(e^{-iaz/2} F(z)\) is of exponential type \(a/2\) in the upper and lower half-plane, then \(\{e^{i\lambda_nt}/(1+ |\lambda_n|)^s\}\) is the Riesz basis in \(H^s_2(0,a)\), provided that \(s\) is not a negative half-integer. If \(s\) is a negative half-integer than the result holds for the space of distributions from \(H^s_2(\mathbb{R})\) supported in \([0,a]\).