A fixed point theorem for measurable selection valued correspondences induced by upper Caratheodory correspondences (Q2105903)

From MaRDI portal
scientific article
Language Label Description Also known as
English
A fixed point theorem for measurable selection valued correspondences induced by upper Caratheodory correspondences
scientific article

    Statements

    A fixed point theorem for measurable selection valued correspondences induced by upper Caratheodory correspondences (English)
    0 references
    0 references
    0 references
    8 December 2022
    0 references
    Let \((\Omega, B_\Omega, \mu)\) be a probability space; \(Y := Y_1 \times \dots\times Y_m = [-M,M]^m \subset \mathbb{R}^m, M > 0;\) \(X := X_1 \times \dots\times X_m\), where \(Y_d\), \(d = 1,\dots,m\) is a convex, compact metrizable subset of a locally convex topological vector space. Let \(\mathcal{L}^\infty_Y :=\mathcal{L}^\infty_{Y_1} \times \dots\times \mathcal{L}^\infty_{Y_m}\), where for each \(d = 1,\dots,m\), \(\mathcal{L}^\infty_{Y_d}\) is a convex, weak star compact metrizable subset of \(\mathcal{L}^\infty_{\mathbb{R}}\). Let \(P_f(X)\) denote the collection of all nonempty closed subsets of \(X\) and \(\mathcal{N} \colon \Omega \times \mathcal{L}^\infty_{Y} \to P_f(X)\) be a \(uC\) correspondence, i.e., it is jointly measurable in \((\omega, v)\) and upper semicontinuous in \(v\) for each \(\omega\). It is supposed that a function \(u \colon \Omega \times \mathcal{L}^\infty_{Y} \times X \to Y\), \(u(\omega, v, x) := (u_1(\omega, v, x),\dots, u_m(\omega, v, x))\) is measurable in \(\omega\) and continuous in \((v,x)\). The main result is as follows: Let \[ (\omega, v) \to \mathcal{P}(\omega, v) := u(\omega, v, \mathcal{N}(\omega, v)). \] If the multimap \(\mathcal{N}\) contains a \(uC\) sub-correspondence \(\eta (\cdot,\cdot)\) taking closed connected values in \(X\) (i.e., \(\operatorname{Gr}\eta(\omega, \cdot) \subset \operatorname{Gr} \mathcal{N}(\omega, \cdot)\) for all \(\omega\)) then there exists \(v^\star \in \mathcal{L}^\infty_{Y}\) such that \[ v^\star \in \mathcal{P}(\omega, v^\star)\quad \mu - \text{a.e. }\omega. \] Some applications to the existence of stationary Markov perfect equilibria in disconnected stochastic games are considered.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    multifunction
    0 references
    measurable selection
    0 references
    fixed point
    0 references
    upper Caratheodory correspondence
    0 references
    selection valued correspondence
    0 references
    approximate Caratheodory selection
    0 references
    discounted stochastic game
    0 references
    stationary Markov perfect equilibria
    0 references
    0 references