We will consider a family of cellular automata $\Phi: \{ 1,2,...,r\}^\mathbb{ N}\circlearrowright$ that are not of algebraic type. Our first goal is to determine conditions that result in the identification of probabilities that are at the same time $\sigma$-invariant and $\Phi$-invariant, where $\sigma$ is the full shift. Via the use of versions of the Ruelle operator $\mathcal{ L}_{ A,\sigma}$ and $\mathcal{ L}_{ B,\Phi}$ we will show that there is an abundant set of measures with this property; they will be equilibrium probabilities for different Lispchitz potentials $A,B$ and for the corresponding dynamics $\sigma$ and $\Phi$. Via the use of a version of the involution kernel $W$ for a $(\sigma,\Phi)$-mixed skew product $\hat{ \Phi}: \{ 1,2,...,r\}^\mathbb{ Z}\circlearrowright$, given $A$ one can determine $B$, in such way that the integral kernel $e^W$ produce a duality between eigenprobabilities $\rho_A$ for $(\mathcal{ L}_{ A,\sigma})^*$ and eigenfunctions $\psi_B$ for $\mathcal{ L}_{ B,\Phi}$. In another direction, considering the non-mixed extension $\hat{ \Phi}_n : \{ 1,2,...,r\}^\mathbb{ Z}\circlearrowright$ of $\Phi$, given a Lispchitz potential $\hat{ A} : \{ 1,2,...,r\}^\mathbb{ Z}\to \mathbb{ R}$, we can identify a Lipschitz potential $A:\{ 1,2,...,r\}^\mathbb{ N} \to \mathbb{ R} $, in such away that relates the variational problem of $\hat{ \Phi}_n$-Topological Pressure for $\hat{ A}$ with the $\Phi$-Topological Pressure for $A$. We also present a version of Livsic's Theorem. Whether or not $\Phi$ (or $\hat{ \Phi})$ can eventually be conjugated with another shift of finite type is irrelevant in our context.
Categories: math.DS math-ph math.MP math.PR nlin.CG