Best copositive approximation (Q1804008): Difference between revisions

Let \(M\) be an \(n\)-dimensional Chebyshev subspace of \(C[a,b]\). For \(f\in C[a,b]\) define \(M_ f\) as the set of functions \(p\in M\) such that \(p(x)f(x)\geq 0\) for all \(x\in [a,b]\). Best copositive approximation to \(f\) is the approximation by elements of \(M_ f\). The author develops a theory for best copositive approximation which is quite analogous to the theory of Chebyshev approximation. In particular, he obtains characterization theorems by ``zero in the convex hull'' and by alternation, and a strong uniqueness theorem. He proposes also a computational algorithm based on a transformation to a problem of restricted range approximation.