Minkowski valuations and generalized valuations (Q1664346)

A Minkowski valuation on the space \({\mathcal K}^n\) of convex bodies in \({\mathbb R}^n\) is a mapping \(\Phi\) of \({\mathcal K}^n\) into itself which satisfies \(\Phi(K\cup L)+\Phi(K\cap L)=\Phi(K)+\Phi(L)\) if \(K,L,K\cup L\in {\mathcal K}^n\), where \(+\) denotes Minkowski addition. It was a longstanding problem to characterize the Minkowski valuations that are continuous (with respect to the Hausdorff metric), translation invariant and \(\mathrm{SO}(n)\)-equivariant. (Previous attempts required additional assumptions, such as special degrees, smoothness, or evenness). The authors now succeed with a complete characterization. Their main result says that any Minkowski valuation \(\Phi\) with the listed properties is given by \[ h_{\Phi K}=c_0+\sum_{i=1}^{n-2} S_i(K,\cdot) *\mu_i + S_{n-1}(K,\cdot) * f_{n-1}+c_n V_n(K) \] for \(K\in{\mathcal K}^n\). Here \(h\) denotes the support function, \(S_i\) the \(i\)th area measure, \(V_n\) the volume, and \(*\) is a suitably defined convolution. Further, \(c_0,c_n\geq 0\) are constants, \(\mu_i\) is a finite signed Borel measure and \(f_{n-1}\) a continuous function on \({\mathbb S}^{n-1}\), both having their center of mass at the origin and being invariant under \(\mathrm{SO}(n-1)\) (fixing some given point of the sphere). The characterization is made possible by using techniques involving generalized valuations, which were introduced by Semyon Alesker and Dmitry Faifman, but it is still a long way to go. The authors also prove, and use, a classification of generalized spherical valuations. The paper concludes with applications to Crofton and kinematic formulae for Minkowski valuations.