No label defined (Q4359488)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 1075548
Language Label Description Also known as
default for all languages
No label defined
    English
    No label defined
    scientific article; zbMATH DE number 1075548

      Statements

      0 references
      16 October 1997
      0 references
      necessary and sufficient convexity conditions
      0 references
      ranges of finite-dimensional vector measures
      0 references
      Radon-Nikodým derivative
      0 references
      Lyapunov's theorem
      0 references
      zonoids
      0 references
      According to the celebrated theorem of Lyapunov, the range of an atomless vector measure \(\sigma:{\mathcal S}\to\mathbb{R}^n\) is convex, where \(\mathcal S\) denotes a \(\Sigma\)-field of subsets of a set \(X\). In the present work, this result is generalized to necessary and sufficient convexity conditions for the ranges of all finite-dimensional vector measures. In particular, it is shown that the range of \(\sigma\) is convex if and only if \(\sigma\) is atomless or the measure space \((X,{\mathcal S},\sigma)\) possesses a characteristic atomic structure. The latter condition means concretely that for every \(\sigma\)-atom \(A\), the inverse image \(U_A= f^{-1}\left(\left\{\pm{\sigma(A) \over|\sigma(A) |}\right\}\right)\) of the antipodes \(\left\{\pm{\sigma(A)\over|\sigma(A)|}\right\}\) contains finitely or countably many disjoint \(\sigma\)-atoms \(A_k\) (\(k\in J\subseteq\mathbb{N}\)), where \(f\) denotes the Radon-Nikodým derivative of \(\sigma\) with respect to the total variation \(v(\sigma)\); additionally, the value \(\sigma(A)\) must satisfy the estimation \(|\sigma(A)|\leq v(\sigma)(U_A)- \sum_{j\in J_A}|\sigma(A_j)|\), where the finite index set \(J_A\) is given by \(J_A=\{j\in J\mid|\sigma(A_j)|\geq |\sigma(A)|\}\) and \(|\cdot|\) denotes the Euclidean norm on \(\mathbb{R}^n\) induced by the usual inner product \(\langle\;,\;\rangle\). The converse of Lyapunov's theorem, presented in this paper, can be stated as an immediate consequence of the above criterion, which may, moreover, be supplemented by necessary and sufficient topological convexity conditions. So it is shown in Chapter 4 that the convexity of the range \(\sigma({\mathcal S})\) is equivalent to the star-shapedness of a significant subset consisting of at most countably many compacta which have the topological dimension 0 or 1. Furthermore, in the case \(n\geq 2\), the range of \(\sigma\) is convex if and only if it contains every one-dimensional face of its convex hull. To verify these statements, a short summary of needed notations and results concerning a certain class of convex bodies, called zonoids, is given in Chapter 2. In Chapter 3 the distribution of the one-dimensional faces of a zonoid \(Z_\mu\) is characterized with respect to the generating Borel measure \(\mu\). The result of this investigation is essentially needed in Chapter 4 to show the necessity of the asserted convexity conditions.
      0 references

      Identifiers