Convergence of metric two-level measure spaces (Q2175330)

A metric two-level measure space, denoted m2m, is a Polish metric space \((X,r)\) equipped with a finite measure \(\nu\) on the set \(\mathcal{M}_f(X)\) on finite Borel measures on \(X\); it is assumed that \(X\subset\mathbb{R}^{\mathbb{N}}\). In order to introduce a topology on \(\mathbb{M}^{(2)}\), the set of equivalence classes on m2m spaces, a set of test functions is defined. A two-level probability measure \(\nu\in\mathcal{M}_1(\mathcal{M}_1(X))\) can then be reconstructed in the sense of Proposition 3.5. Test functions separate points in \(\mathbb{M}^{(2)}\). The initial topology on \(\mathbb{M}^{(2)}\) induced by the test functions is called the two-level Gromov-weak topology; this is implied by weak convergence. A two-level Gromov-Prokhorv metric \(d_{2GP}\) is introduced that makes \(\mathbb{M}^{(2)}\) a Polish metric space. Compact sets, as well as compact nets, in \((\mathbb{M}^{(2)},d_{2GP})\) are characterised. The characterisation of compact sets is used in order to prove that the Gromov-weak and the Gromov-Prokhorov topologies coincide. In the final section, the finite and infinite nested Kingman coalescents are defined and studied.