Selections and their absolutely continuous invariant measures (Q2019217)

The authors study multivalued maps whose graphs are defined by single-valued maps \(\tau_1, \tau_2:\mathbb{I}\to\mathbb{I}\) with \(\tau_1(x)\leq\tau_2(x)\) for any \(x\in\mathbb{I}\). Let \(G\subset\mathbb{I}\times\mathbb{I}\) be the region between the boundaries of the graphs of \(\tau_1\) and of \(\tau_2\). Any map \(\tau:\mathbb{I}\to\mathbb{I}\) with the graph in \(G\) is called a selection of the multivalued map defined by \(G\). The main result: If \(\tau_1\), \(\tau_2\) are piecewise monotonic, their invariant distribution functions \(F^{(1)}\) and \(F^{(2)}\) are continuous, and for any \(0<\lambda<1\) the convex combination \(F=\lambda F^{(1)}+(1-\lambda)F^{(2)}\) is an automorphism of \(\mathbb{I}\), then there exists a piecewise monotonic selection \(\tau\), \(\tau_1\leq\tau\leq \tau_2\), preserving the distribution function \(F\). The paper contains also a big number of motivating examples. In the last part the authors study relationships between dynamics of such defined multivalued maps and random maps.