zbMATH — the first resource for mathematics

On the convergence in distribution of measurable multifunctions (random sets), normal integrands, stochastic processes and stochastic infima. (English) Zbl 0611.60004
In Trans. Am. Math. Soc. 266, 275-289 (1981; Zbl 0501.28005) the authors have given various characterizations for the a.s. convergence and the convergence in probability of sequences of closed-valued measurable multifunctions (random closed sets). In the present paper they study their convergence in distribution (weak*-convergence of the induced probability measures) and the epi-convergence in distribution of normal integrands (random lower semicontinuous functions).
The relation between the epi-convergence in distribution and the convergence in the classical sense of stochastic processes is analyzed. As an application a modified derivation of Donsker’s invariance principle is given. The potential application to the study of the convergence of stochastic infima is also suggested.
Reviewer: V.Mackevičius

60B10 Convergence of probability measures
90C15 Stochastic programming
60D05 Geometric probability and stochastic geometry
Zbl 0501.28005
Full Text: DOI