Finite type as fundamental objects even non-single-valued and non-continuous
Zhengyu Yin
TL;DR
The paper develops a unified framework linking multivalued dynamics to finite-type graph models, shadowing, and shifts of finite type via inverse limits under the Mittag-Leffler condition. It proves that every closed relation on a compact totally disconnected space is an ML inverse limit of finite graph homomorphisms, and that a multivalued system has shadowing precisely when its orbit space is conjugate to an ML inverse limit of shifts of finite type. The work extends the Good–Meddaugh paradigm to multivalued maps, showing richer dynamical behavior and density of shadowing within upper semicontinuous maps, and connects graph covers to Shimomura’s graph-cover framework. It also furnishes a general method to approximate multivalued dynamics by finite-type models and extends shadowing results from zero-dimensional to general spaces via inverse-limit constructions, yielding broad structural insight into multivalued shadowing phenomena.
Abstract
In this paper, inspired by the elegant work of Good and Meddaugh \cite{GM} and the graph models for zero-dimensional systems developed by several authors, like Gambaudo and Martens \cite{GM06}, Shimomura \cite{Sh14}. We try to discover a connection among some objects, such as finite directed graph, shift of finite type and shadowing property by employing the Closed Graph Theorem for multivalued maps. From the perspective of structure theorems, we demonstrate that every closed relation (multivalued map) on a compact, totally disconnected space is represented as an inverse limit of finite directed graph homomorphisms satisfying the Mittag-Leffler condition. Moreover, from dichotomy-theorem point of view, we prove that an inverse limit of finite directed graph homomorphisms possesses the shadowing property if and only if its induced space of infinite graph walks (as a shift of finite type) satisfies the Mittag-Leffler condition. As an application, a question raised by Boroński, Bruin and Kucharski \cite{BBK} is also concerned. Furthermore, we show that under a multivalued dynamical system, the resulting dynamical behaviors exhibit greater diversity and counterintuitively compared to those observed in single-valued continuous systems.
