Uniform convergence of derivatives of Lagrange interpolation (Q1195721)

The Lebesgue constant in the Lagrange interpolation process taking as nodes the zeros of the Chebyshev polynomials of the first kind, has minima order. The theory of orthogonal polynomials presents further examples of sets of nodes with minimal order Lebesgue constants. On the other hand, several sets of nodes give rise to Lebesgue constant with higher order. The Lebesgue constants play an important role in deriving uniform convergence estimates for numerical methods based on Lagrange interpolation processes. In this paper, the author shows that by adding some nodes, they can transform any set of Jacob \(i\) zeros, into sets of points with minimal order Lebesgue constants. Then they prove that there exist infinitely many sets of nodes such that the associated Lagrange interpolation polynomials can be used to simultaneously approximate a function and its derivatives in the uniform norm. They also derive pointwise and uniform error bounds, which are in a certain sense optimal, for the corresponding errors. The case of multiple knots is almost considered.