Convex cores of measures on R d (Q4780290)

Given a finite Borel measure \(Q\) on the Euclidean space \(R^d\), the intersection of all closed convex sets of full measure is termed the convex support of \(Q\). The authors consider and investigate the intersection of all convex sets of full measure, call it the convex core of \(Q\) and denote it by cc\((Q)\). Among other results they prove: A convex set \(C\subseteq R^d\) is the convex core of some finite Borel measure if and only if \(C\) has at most a countable number of faces. The convex core of a finite Borel measure \(Q\) equals \(\{\int_{R^d}xP(dx):P\ll Q\}\), where \(P\) ranks over all finite Borel measures absolutely continuous w.r.t. \(Q\). cc\((Q_1*Q_2)=\text{cc}(Q_1)+\text{cc}(Q_2)\) for any two finite Borel measures \(Q_1\) and \(Q_2\). Here \(*\) denotes convolution and \(+\) the Minkowski sum.