On a Marinacci uniqueness theorem for measures. (Q1414202): Difference between revisions

The authors prove some generalizations of the Marinacci theorem [see \textit{M. Marinacci}, Decis. Econ. Finance 23, No.~2, 121--132 (2000; Zbl 0987.28002)] for a D-poset \(L\) [\textit{F. Kôpka} and \textit{F. Chovanek}, Math. Slovaca 44, No.~1, 21--34 (1994; Zbl 0789.03048) and \textit{B. Riečan} and \textit{T. Neubrunn}, ``Integral, measure, and ordering'' (1997; Zbl 0916.28001)]. As an illustration we present one of them. There are given two measures \(\mu,\nu : L \to [0,\infty)\), \(\mu\) is convex-ranged and there exist \(\alpha\in (0,\mu(1))\), \(\beta\in (0,\nu(1))\) such that \(\mu(a) = \alpha\Leftrightarrow \nu(a) = \beta\) and \(\nu(a) = 0\Rightarrow \mu(a) = 0\). If \(L\) is \(\sigma\)-complete and \(\nu\) is \(\sigma\)-additive, then \(\mu = \mu(1)/\nu(1)\nu\). The results play an important role in the study of the modelization of choice behaviours of decision makers that have to face, simultaneously, risk and ambiguity.