Discontinuous information in the worst case and randomized settings

Aicke Hinrichs, Erich Novak, Henryk Wozniakowski

We believe that discontinuous linear information is never more powerful than continuous linear information for approximating continuous operators. We prove such a result in the wor…