Discrete l 1 Approximation by Rational Functions
George A. Watson · IMA Journal of Numerical Analysis · 1984
The problem of finding a best approximation by a rational function to discrete data, using the l1 norm, is considered. An algorithm is developed which is frequently convergent in a finite number of steps, and failing this usually has a second-order convergence rate. Details are given of the application of the algorithm to a number of rational approximation problems.