Groups of convex bodies (Q6112059)

The reviewer's polytope algebra [\textit{P. McMullen}, Adv. Math. 78, No. 1, 76--130 (1989; Zbl 0686.52005)] describes the abstract structure that underlies translation invariant valuations on the space of convex polytopes (in a \(d\)-dimensional vector space \(V\) over \(\mathbb{R}\), but any ordered field will serve). The author's aim in this paper is to extend, so far as possible, the properties of the polytope algebra to the family \(\mathcal{K}(V)\) of convex bodies in \(V\), topologized by the Hausdorff metric. If \(G\) is a group of affine transformations of \(V\) that contains all translations, then the group of convex bodies \(\mathcal{CB}(V)\) is the free Hausdorff topological abelian group on the space \(\mathcal{K}(V)\), modulo the closed subgroup generated by the elements of the form \[ [gX] - [X] \text{ and } [B \cup C] - [B] - [C] + [B \cap C] \] for \(g \in G\), \(X \in \mathcal{K}(V)\), and \(B,C \in \mathcal{K}(V)\) with \(B \cup C \in CK(V)\). It is the target of a universal continuous \(G\)-invariant valuation \(\Phi\colon \mathcal{K}(V) \to \mathcal{CB}(V)\). What is shown here is that, if \(\varphi\colon \mathcal{K}(V) \to A\) is a continuous translation-invariant valuation with values in a Hausdorff topological abelian group \(A\), and \(c \in A\) denotes the value of \(\varphi\) on any one-point convex body, then the valuation \(\varphi - c\colon \mathcal{K}(V) \to A\) admits a factorisation \[ \mathcal{K}(V) \stackrel{f}{\to} \mathcal{V} \stackrel{g}{\to} A, \] where \(\mathcal{V}\) is a Hausdorff topological vector space, \(f\) is a continuous translation-invariant valuation, and \(g\) is a continuous homomorphism of topological abelian groups. One needs to subtract a constant, since the constant map \(\mathcal{K}(V ) \to \mathbb{Z}\) with value \(1\) is a continuous translation invariant valuation. This result follows from an analogue of what is called here McMullen polynomiality, namely, a direct sum decomposition of Hausdorff topological abelian groups \[ \mathcal{CB}(V,G) \cong \mathcal{CB}_0(V,G) \oplus \cdots \oplus \mathcal{CB}_d(V,G), \] and a corresponding decomposition \[ \Phi = \Phi_0 \oplus \cdots \oplus \Phi_d \] of \(\Phi\) into continuous \(G\)-invariant valuations \(\Phi_i\colon \mathcal{K}(V) \to \mathcal{CB}_i(V,G)\), with \(\mathcal{CB}_0(V,G)\) a copy of \(\mathbb{Z}\) and \(\mathcal{CB}_i(V,G)\) having the structure of a real vector space for \(1 \leq i \leq d\), with respect to which dilatation by \(\lambda > 0\) satisfies \(\Phi(\lambda X) = \lambda^i\Phi_i(X)\) for \(X \in \mathcal{K}(V)\). A key contribution is the replacement of the simplex decompositions in the polytope case by the fact that, if \(\varphi\) is a polynomial function of degree \(r\), then \(\psi\) defined for fixed \(t\) by \[ \psi(x) = \varphi(x + t) - \varphi(x) \] is polynomial of degree \(r - 1\).