Convergence of extended Lagrange interpolation in weighted \(L_ p\) norm (Q1901527)

Let \(\phi: [- 1, 1]\to (0, \infty)\) such that \(\phi\in \text{Lip}(\lambda), \lambda\in (0, 1]\) and \(\phi, \phi^{-1}\in L^\infty\). If \(n\in \mathbb{N}\), \(\alpha, \beta> - 1\) \(\gamma_j> -1\) and \(|y_j|< 1\) for \(j\in 1,\dots, n\) then the function \[ w(x):= \phi(x)(1- x)^\alpha(1+ x)^\beta \prod^n_{j= 1} |x- y_j|^{\gamma_j} \] is a smooth generalized Jacobi weight (\(w\in GSJ\)). If \(w\in GSJ\) and \(w_1(x):= (1- x)w(x)\), \(w_2(x):= (1+ x)w(x)\), then \(w_1\) and \(w_2\) are in \(GSJ\) too. Let \(p_m(w)\), \(p_m(w_1)\), \(p_m(w_2)\), \(m\in \mathbb{N}\), the corresponding systems of orthonormal polynomials with positive leading coefficient. If \(t_1,\dots, t_{3m}\) are the simple zeros of the polynomial \(q_{3m}:= p_m(w) p_m(w_1) p_m(w_2)\), the author considers the additional nodes \((\sigma)\) and \((\tau)\) defined as \(\sigma_i:= - 1+ s^{- 1}(i- 1)(t_1+ 1)\), \(i\in 1,\dots, s\), and \(\tau_j:= t_{3m}+ r^{- 1} j(- t_{3m}+ 1)\), \(j\in 1,\dots, r\). For \(f\), which is a continuous function on \([- 1, 1]\), \(L_{3m,r,s}(w; f)\) is the Lagrange polynomial interpolating \(f\) at the knots \((\sigma)\cup(t)\cup (\tau)\). The main purpose of this paper is to establish necessary and sufficient conditions (Theorem 2.1) for the weighted \(L_p\) convergence of the interpolation process. Finally, Theorem 2.2 gives sufficient conditions for the simultaneous interpolation.