Dice Question Streamline Icon: https://streamlinehq.com

Generation of CA(G,A) by Lazy and Invertible Cellular Automata

Prove or disprove that the monoid of all cellular automata over A^G equals the submonoid generated by invertible cellular automata together with lazy cellular automata; specifically, establish whether CA(G,A) = ⟨ICA(G,A) ∪ L(G,A)⟩. If this equality does not hold, determine properties and structure of the submonoids ⟨ICA(G,A) ∪ L(G,A)⟩ and ⟨L(G,A)⟩.

Information Square Streamline Icon: https://streamlinehq.com

Background

Cellular automata over AG form a monoid under composition, with invertible CAs forming the group ICA(G,A). In classical semigroup theory, the full transformation monoid on a finite set is generated by idempotents of defect 1 together with the symmetric group. Lazy CAs with minimal neighborhood {e} correspond to such idempotents, suggesting a possible generation result for CA(G,A) using lazy and invertible CAs.

Motivated by this analogy, the authors ask whether every CA can be expressed as a composition of lazy and invertible CAs; if not, they seek structural insights into the submonoids generated by these classes.

References

In this section, we propose two open problems related to the study of lazy CAs. If \mathcal{L}(G,A) is the set of all lazy cellular automata over AG, prove or disprove the following: \text{CA}(G,A) = \langle \text{ICA}(G,A) \cup \mathcal{L}(G,A) \rangle. If the above does not hold, what can we say of the submonoids \langle \text{ICA}(G,A) \cup \mathcal{L}(G,A) \rangle and \langle \mathcal{L}(G,A) \rangle?

On the order of lazy cellular automata (2510.14841 - Alcalá-Arroyo et al., 16 Oct 2025) in Problem 2, Section 4 (Open problems)