1: \begin{abstract}

2: Let $A$ be a finite set and $n \geq 2$. This paper introduces the concept of \emph{universal simulation} in the context of semigroups of transformations of $A^n$, also known as finite state-homogeneous automata networks. For $m \geq n$, a transformation of $A^m$ is defined as \emph{$n$-universal of size $m$} if it may simulate every transformation of $A^n$ by updating one coordinate (or register) at a time. Using tools from memoryless computation, it is established that there is no $n$-universal transformation of size $n$, but there is such a transformation of size $n+2$. An $n$-universal transformation is defined as \emph{complete} if it may sequentially simulate every finite sequence of transformations of $A^n$; in this case, minimal examples and bounds for the size and time of simulation are determined. It is also shown that there is no $n$-universal transformation that updates all the registers in parallel, but that there exists a complete one that updates all but one register in parallel. This illustrates the strengths and weaknesses of parallel models of computation, such as cellular automata.

3: \end{abstract}