The envelope of the error for trigonometric and Chebyshev interpolation

From MaRDI portal
(Redirected from Publication:1176533)





The error in Fourier or Chebyshev interpolation is the product of a rapidly varying factor with a slowly varying modulation (called the envelope fo the error). The author studies here the envelope which controls entirely the magnitude and spatial uniformity of the error. One interesting conclusion is that for simple functions, i.e., those with a single domainant pair of complex conjugate singularities or a single peak, the Fourier/Chebyshev error is anything but uniform. It is observed that the error envelope is sharply peaked about the point where \(f(x)\) is peaked or which is nearest the poles or branch points. It is also concluded that the error in solving a second-order differential equation using the pseudospectral method with \(N\) collocation points is typically smaller than the `residual' function by a factor of \(O(N^ 2)\).











This page was built for publication: The envelope of the error for trigonometric and Chebyshev interpolation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1176533)