Quasi-definiteness of generalized Uvarov transforms of moment functionals (Q5955004): Difference between revisions

The authors consider a point masses perturbation of a quasi-definite moment functional and give necessary and sufficient conditions for the perturbed functional also to be quasi-definite. The perturbed functional in question is supposed to be of the form \[ \tau= \sigma+ \lambda \sum^m_{\ell=1} \sum^{m_\ell}_{k=0} {(-1)^k\over k!} u_{\ell k}\delta^{(k)} (x-c_\ell), \] where \(\sigma\) denotes the original quasi-definite functional while \(\lambda\), \(u_{ \ell k}\) are given complex numbers. The main result is expressed in terms of the coefficients \(u_{\ell k}\), as well as the kernel polynomials \(K_n(c_t, c_\ell)\) defined through \[ K_n(x,y)= \sum^n_{j=0} {P_j(x)P_j(y) \over\langle \sigma,P^2_j \rangle},\;n=0,1, \dots. \] Here \(\{P_j\}^\infty_0\) stands for the monic orthogonal polynomial system relative to \(\sigma\). Some properties (the recurrence relation, number of the zeros in the interval of orthogonality etc.) pertinent to the monic orthogonal polynomial system relative to \(\tau\) are also given.