Inverse polynomial mappings and interpolation on several intervals (Q2634179)

The theme of this paper is the construction of optimal data points for Lagrange interpolation on the real line. To this end, so called inverse polynomial images are employed to find such data points that minimise Lebesgue constants. This is well understood for single compact intervals on the real line. The authors now consider consider collections of disjoint intervals on which the relevant polynomials are defined, and for the same (inverse polynomial image) method to work, so called T-polynomials (replacing Chebyshev polynomials) are used if they exist. The authors no only do that but show as well that the existence of T-polynomials is necessary for the method to work and to provide small Lebesgue numbers.