Efficient application of nonlinear stationary operators in adaptive wavelet methods -- the isotropic case (Q2391825)

This paper deals with adaptive wavelet methods, where a nonlinear local operator is applied to an \(n\)-variate function represented in wavelet coordinates in terms of local polynomials. The authors show that a tree structured set of wavelet indices can be represented using local polynomials on a partition of standard hypercubes of the underlying domain. A nonlinear local operator is piecewise applied on this set of disjoint cells and from this the wavelet coefficients of the target tree are reconstructed. Both transformations and the operator are applied in linear time with respect to the number of wavelet coefficients. Numerical examples with \(n\)-variate wavelets \((n=1,\ldots,4)\) demonstrate the linear complexity.