Fast evaluation of nonlinear functionals of tensor product wavelet expansions (Q652260)

This paper is concerned with the fast evaluation of nonlinear functionals of tensor product wavelet expansions. First the authors introduce interpolets (interpolating wavelets), and define parent-child relations on the index set of these functions. For a nonlinear functional \(f\) and for \(u\) from the span of a set of interpolates, where the index set forms a tree, the authors give an algorithm to compute the interpolant of \(f(u)\) from the span of this set of interpolets in linear complexity. For the fast evaluation of nonlinear functionals, interpolets were formerly used by \textit{W. Dahmen}, \textit{W. Schneider}, and \textit{Y. Xu} [Numer. Math. 86, No. 1, 49--101 (2000; Zbl 0962.65118)]. Then the authors extend this fast algorithm to tensor products of interpolets with index sets that form multiple trees. Applications are possible in (adaptive) tensor product solution methods for semilinear operator equations by collocation or Galerkin methods.