Stationary distributions for a class of generalized Fleming-Viot processes (Q2450253): Difference between revisions

The author considers a class of generalized Fleming-Viot processes [\textit{J. Bertoin} and \textit{J.-F. Le Gall}, Probab. Theory Relat. Fields 126, No. 2, 261--288 (2003; Zbl 1023.92018)] whose stationary distributions can be identified. Such processes could be used to model the one-dimensional Wright-Fisher diffusion with mutation. Two problems are addressed: (1) how to modify both mechanisms of reproduction and mutation to find a jump process whose generator extends the generator of the standard Fleming-Viot process, and whose stationary distribution can be identified and (2) how to establish an analogous generalization for the Fleming-Voit process with parent independent mutation [\textit{T. Shiga}, J. Math. Kyoto Univ. 30, No. 2, 245--279 (1990; Zbl 0751.60044)]. The line of reasoning is based on the relationship to measure-valued branching processes with immigration. The stationary distributions are obtained from normalized stable random measures after a suitable biased transformation followed by mixing random measure with the same parameter measure by the law of a Dirichlet random measure. Moreover, the processes under consideration are proved to be irreversible.