Shape-Preserving Approximation by Real and Complex Polynomials

In many problems arising in engineering and science one requires approxi- tion methods to reproduce physical reality as well as possible. Very schema- cally, if the input data represents a complicated discrete/continuous quantity of information, of “shape” S (S could mean, for example, that we have a “monotone/convex” collection of data), then one desires to represent it by the less-complicated output information, that “approximates well” the input data and, in addition, has the same “shape” S.