Markov and Time-Homogeneity Properties in Dempster-Shafer Random Walks

We generalize discrete-time finite-horizon random walks in order to deal with ambiguity or mis-specification inside the framework of belief functions. Referring to the product rule of conditioning for belief functions, we propose suitable definitions of Markov and time-homogeneity properties. Moreover, we investigate to what extent such properties can be constrained by means of one-step time-homogeneity.