Symbolic Dynamics and Markov Partitions

Roy Adler
DOI: https://doi.org/10.48550/arXiv.math/9607214
1996-07-02
Abstract:The decimal expansion real numbers, familiar to us all, has a dramatic generalization to representation of dynamical system orbits by symbolic sequences. The natural way to associate a symbolic sequence with an orbit is to track its history through a partition. But in order to get a useful symbolism, one needs to construct a partition with special properties. In this work we develop a general theory of representing dynamical systems by symbolic systems by means of so-called Markov partitions. We apply the results to one of the more tractable examples: namely hyperbolic automorphisms of the two dimensional torus. While there are some results in higher dimensions, this area remains a fertile one for research.
Dynamical Systems
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how and to what extent a dynamical system can be represented by a symbolic system. Specifically, the paper explores the theories and methods of representing dynamical systems through so - called Markov partitions. These Markov partitions have special properties, enabling the orbits of the dynamical system to be traced by symbol sequences, thus providing a method of simplifying complex dynamical systems into a form that is easier to analyze. In the paper, particular attention is paid to how to construct such Markov partitions and apply them to specific examples of dynamical systems, such as hyperbolic automorphisms of the two - dimensional torus. Through this method, deep insights into the behavior of the dynamical system can be obtained, including the identification of periodic orbits, almost - periodic orbits and dense orbits, etc. In addition, the paper also discusses how to construct Markov partitions in reverse from the symbolic representation, which involves an interesting question: should Markov partitions be constructed first to obtain the symbolic representation, or should the symbolic representation be generated first to obtain the Markov partition? The answer to this question is crucial for understanding the relationship between dynamical systems and symbolic dynamics.