Criteria for the Best Approximation by Simple Partial Fractions on Semi-Axis and Axis

Abstract

We study uniform approximation of real-valued functions f, f(∞) = 0, on ℝ⁺ and ℝ by real-valued simple partial fractions (the logarithmic derivatives of polynomials). We obtain a criterion for the best approximation on ℝ⁺ and ℝ in terms of the Chebyshev alternance. This criterion is similar to the known criterion on finite segments. For the problem of approximating odd functions on ℝ we construct an alternance criterion with a weakened condition on the poles of fractions. We present a criterion for the best approximation by simple partial fractions on ℝ⁺ and ℝ in terms of Kolmogorov. We prove analogs of the de la Vallee-Poussin alternation theorem.