On involution kernels and large deviations principles on $ \beta $-shifts

Abstract

<p style='text-indent:20px;'>Consider <inline-formula><tex-math id="M2">\begin{document}$ \beta &gt; 1 $\end{document}</tex-math></inline-formula> and <inline-formula><tex-math id="M3">\begin{document}$ \lfloor \beta \rfloor $\end{document}</tex-math></inline-formula> its integer part. It is widely known that any real number <inline-formula><tex-math id="M4">\begin{document}$ \alpha \in \Bigl[0, \frac{\lfloor \beta \rfloor}{\beta - 1}\Bigr] $\end{document}</tex-math></inline-formula> can be represented in base <inline-formula><tex-math id="M5">\begin{document}$ \beta $\end{document}</tex-math></inline-formula> using a development in series of the form <inline-formula><tex-math id="M6">\begin{document}$ \alpha = \sum_{n = 1}^\infty x_n\beta^{-n} $\end{document}</tex-math></inline-formula>, where <inline-formula><tex-math id="M7">\begin{document}$ x = (x_n)_{n \geq 1} $\end{document}</tex-math></inline-formula> is a sequence taking values into the alphabet <inline-formula><tex-math id="M8">\begin{document}$ \{0,\; ...\; ,\; \lfloor \beta \rfloor\} $\end{document}</tex-math></inline-formula>. The so called <inline-formula><tex-math id="M9">\begin{document}$ \beta $\end{document}</tex-math></inline-formula>-shift, denoted by <inline-formula><tex-math id="M10">\begin{document}$ \Sigma_\beta $\end{document}</tex-math></inline-formula>, is given as the set of sequences such that all their iterates by the shift map are less than or equal to the quasi-greedy <inline-formula><tex-math id="M11">\begin{document}$ \beta $\end{document}</tex-math></inline-formula>-expansion of <inline-formula><tex-math id="M12">\begin{document}$ 1 $\end{document}</tex-math></inline-formula>. Fixing a Hölder continuous potential <inline-formula><tex-math id="M13">\begin{document}$ A $\end{document}</tex-math></inline-formula>, we show an explicit expression for the main eigenfunction of the Ruelle operator <inline-formula><tex-math id="M14">\begin{document}$ \psi_A $\end{document}</tex-math></inline-formula>, in order to obtain a natural extension to the bilateral <inline-formula><tex-math id="M15">\begin{document}$ \beta $\end{document}</tex-math></inline-formula>-shift of its corresponding Gibbs state <inline-formula><tex-math id="M16">\begin{document}$ \mu_A $\end{document}</tex-math></inline-formula>. Our main goal here is to prove a first level large deviations principle for the family <inline-formula><tex-math id="M17">\begin{document}$ (\mu_{tA})_{t&gt;1} $\end{document}</tex-math></inline-formula> with a rate function <inline-formula><tex-math id="M18">\begin{document}$ I $\end{document}</tex-math></inline-formula> attaining its maximum value on the union of the supports of all the maximizing measures of <inline-formula><tex-math id="M19">\begin{document}$ A $\end{document}</tex-math></inline-formula>. The above is proved through a technique using the representation of <inline-formula><tex-math id="M20">\begin{document}$ \Sigma_\beta $\end{document}</tex-math></inline-formula> and its bilateral extension <inline-formula><tex-math id="M21">\begin{document}$ \widehat{\Sigma_\beta} $\end{document}</tex-math></inline-formula> in terms of the quasi-greedy <inline-formula><tex-math id="M22">\begin{document}$ \beta $\end{document}</tex-math></inline-formula>-expansion of <inline-formula><tex-math id="M23">\begin{document}$ 1 $\end{document}</tex-math></inline-formula> and the so called involution kernel associated to the potential <inline-formula><tex-math id="M24">\begin{document}$ A $\end{document}</tex-math></inline-formula>.</p>