We prove a variant of the classical Whitney extension theorem, in which the C m -norm of the extending function is controlled up to a given, small percentage error.
IntroductionHere and in [16], we compute the least possible (infimum) C m norm of a function F having prescribed Taylor polynomials at N given points of R n . Moreover, given > 0, we exhibit such an F, whose C m norm is within of the least possible. Our computation consists of an algorithm, to be implemented on an (idealized) digital computer. The algorithm works, thanks to a variant of the classical Whitney extension theorem, in which we control the C m norm of the extending function up to an percentage error. This paper gives the variant of Whitney's theorem, while [16] presents the algorithm and the rest of the mathematics behind it. The number of operations used by our algorithm is C( )N log N, where N is the number of given points, and C( ) grows rapidly as tends to zero.To state our results precisely, we set up notation. Fix m, n ≥ 1. We pick a norm on C m (R n ), subject to restrictions to be spelled out in the next section. We write F C m (R n ) to denote the norm of F. Given x ∈ R n and F ∈ C m (R n ), we write J x (F) to denote the m th order Taylor polynomial of F at x. Thus, J x (F) belongs to P, the vector space of all (real) m th degree polynomials on R n . Let E ⊂ R n . We write #(E) for the number of points in E. (If E is infinite, then #(E) = +∞.) A Whitney field on E is a family P = (P x ) x∈E of polynomials P x ∈ P, indexed by x ∈ E. If P = (P x ) x∈E is a Whitney field and S ⊂ E is a subset, then in an obvious way we can define the restriction P| S of P to S.