Some of the interpolation methods (Thiele and MTT for instance) fail on the rational test function, when the interpolation support is [-1,1]. Is this expected behavior? Does it depend on the number of interpolation points, and that the function is even? Is it because there is no early termination for these methods (AAA handles it fine)?
Some of the interpolation methods (Thiele and MTT for instance) fail on the rational test function, when the interpolation support is [-1,1]. Is this expected behavior? Does it depend on the number of interpolation points, and that the function is even? Is it because there is no early termination for these methods (AAA handles it fine)?