Reconstruction and collocation of a class of non-periodic functions by sampling along tent-transformed rank-1 lattices (Q5964235)

The cosine series of one variable and their \(d\)-dimensional tensor products are taken as the starting point for Fourier analysis of non-periodic functions. Approximation and collocation methods are developed by restricting the frequency support of the functions to weighted hyperbolic crosses. Rank-1 lattice point sets \(\Lambda(z,N)\subset \mathbb{R}^d\), for any \(N\in \mathbb{N}\) and \(z \in \mathbb{Z}_N^d\) of the form \[ \Lambda(z,N)=\left\{\frac{nz}{N}\mod 1 \,\big|\,n =0,1,\dots,N-1 \right\}, \] are componentwise transformed by a tent-transform \(\varphi(x)=1-|2x-1|\). The resulting tent-transformed rank-1 lattice point multiset \(\Lambda_\varphi(z,N)\) contains \(\left \lfloor{N/2+1}\right \rfloor\) points. It is shown that the approximation degree of the set \(\Lambda_\varphi(z,N)\) with respect to the cosine series coincides with the approximation degree of the original lattice \(\Lambda(z,N)\) with respect to the Fourier exponential series. An algorithm for constructing the point sets \(\Lambda(z,N)\) and \(\Lambda_\varphi(z,N)\) is formulated, fast reconstruction and evaluation methods are developed. The convergence rate of approximations of solutions of the Poisson equation, subjected to the Dirichlet and von Neumann boundary conditions, is numerically tested.