scientific article; zbMATH DE number 3932080

From MaRDI portal
Publication:3704658





This paper presents a multivalued version of \textit{N. Etemadi}'s [Z. Wahrscheinlichkeitstheor. Verw. Geb. 55, 119-122 (1981; Zbl 0438.60027)] and Kolmogorov's strong law of large numbers (SLLN). The values of the random sets are assumed to be closed in a separable Banach space and the convergence is obtained in the \textit{U. Mosco} sense [Adv. Math. 3, 510- 585 (1969; Zbl 0192.491)]. The proofs are based on the real and vector versions of the SLLN and also on a previous result of the author, concerning the distribution of random sets with closed values in a separable metric space [C. R. Acad. Sci., Paris, Sér. I 296, 883-886 (1983; Zbl 0539.60010)]. A converse to the SLLN is also given for random sets of which values are closed convex, locally weakly compact and contain no whole line. Since the values of the random sets are not assumed to be bounded, the foregoing results can be applied to epigraphic multifunctions associated to normal integrands; this may be useful, for instance, in stochastic optimization.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3704658)