On the average distance property and certain energy integrals (Q1127035): Difference between revisions

A metric space \((X,d)\) has the average distance property provided that there is a unique real number \(r(X,d)\) satisfying the following condition: for every positive integer \(n\) and every finite subset \(\{x_1,\dots,x_n\}\) of \(X\) there exists an \(x \in X\) such that \[ {1 \over n}\sum_{i=1}^nd(x_i,x)=r(X,d). \] The number \(r(X,d)\) is called the rendezvous number of \((X,d)\). In 1964 O. Gross proved that every compact connected metric space has the average distance property. Further, G. Elton generalized this result: he proved that for every compact connected metric space \((X,d)\) the rendezvous number \(r(X,d)\) is the unique number such that for every regular Borel probability measure \(\mu\) on \(X\) there is an \(x \in X\) with \[ \int_X d(x,y) mu(y)=r(X,d). \] Let \(M(X)\) be the set of all regular Borel probability measures on \(X\) and let \[ M(X,d):=\sup_{\mu \in M(X)} \int_X \int_X d(x,y) d\mu(x) d\mu(y). \] For some metric spaces, for instance for \(X=S^{n-1}\) and \(d\) the Euclidean metric, there exists a \(\mu_0 \in M(X)\) such that \[ r(X,d)=\int_X d(x,y) d\mu_0(y) \] for every \(x \in X\). In the paper under review the author studies the connection between the existence of such a measure \(\mu_0\) on the one hand and relationships between \(r(X,d)\) and \(M(X,d)\) on the other hand. Theorem 1. Let \((X,d)\) be a compact connected metric space. Then (1) \( \;r(X,d) \leq M(X,d)\); (2) If \(\;r(X,d)=M(X,d)\), then there exists a \(\mu_0 \in M(X)\) such that \(\int_X d(x,y) d \mu_0 (y)\) is independent of \(x\). The author proves that the implication converse to (2) in general fails (Prop.2). A number of other interesting results is also presented.