B-spline quasi-interpolant representations and sampling recovery of functions with mixed smoothness (Q647925): Difference between revisions

For a given set of \(n\) functions defined on a \(d\)-dimensional unit cube, quasi-interpolants are considered to functions \(f\) from Besov-spaces constructed from samples \(f(x^j)\) of the function at prescribed \(n\) sample points \(x^j\). The infimum of all choices of samples and all families of basis functions \(\varphi_j\) in \(L_q\) over the unit cube is taken, and suitable basis functions \(\varphi_j\) are constructed that give asymptotic optimality with respect to the mentioned infimum. Upper bounds for the worst error with respect to the choice of \(f\) are also established.