Mixing markovian laws; with an application to path decompositions†
Paavo Salminen · Stochastics · 1983
Let {P α,α∊I} be a family or Markovian laws over some measurable space, and let μ be a probability measure on I. In this paper we give a simple condition which ensures that the mixed law ∫μ(dα)P α is Markovian. Mixing is then used to derive generalizations to J. Pitman's 2M–X theorem and D. Williams' path decomposition theorem.