Convex bodies generated by sublinear expectations of random vectors (Q820925): Difference between revisions

Denoting by \(L^p({\mathbb R})\) the family of all \(p\)-integrable random vectors in \({\mathbb R}^d\), \(p\in[1,\infty],\) a sublinear expectation is defined as a function \(e:L^p({\mathbb R})\to(-\infty,\infty]\) which is monotonous, equivariant under translation, sublinear and lower semi-continuous. In the paper under review, the authors use a sublinear expectation \(e\) to associate a convex closed set \({\mathcal E}_e(\xi)\subset{\mathbb R}^d\) to each \(p\)-integrable random vector \(\xi\in{\mathbb R}^d\), and study properties of these convex sets. This association is done by setting \(h(u)=e(\langle\xi,u\rangle)\) as the support function of \({\mathcal E}_e(\xi)\). They also particularize this general construction to the case when \(\xi\) is uniformly distributed on a convex body. Then, among others, the authors show that some well-known transforms in convex geometry (e.g., the centroid body, the convex floating body, or the Ulam floating body) are special cases of this construction.