Table of Contents
Fetching ...

Some properties of Higman-Thompson monoids and digital circuits

J. C. Birget

TL;DR

The paper builds a unified framework tying Higman–Thompson monoids to acyclic circuit models, showing that monoids such as $M_{2,1}$ and ${\sf tot}M_{2,1}$ naturally arise as right-ideal morphisms and as circuit-input/output maps. It proves finite generation for ${\cal RM}^{\sf fin}_2$ and related monoids, while establishing non-finite generation for several fixed-length and plep/tlep variants. The work also develops completion procedures (deterministic, inverse-homomorphic, generator-based, and circuit-based), and provides a robust representation theory: any element can be decomposed into a finite unambiguous union of partial circuits with disjoint domains, bridging algebraic and circuit viewpoints. Collectively, these results reveal a deep equivalence between Thompson monoids and circuit-model computations, with concrete algorithms for completion and decomposition and a detailed complexity landscape for related decision problems.

Abstract

We define various monoid versions of the R. Thompson group $V$, and prove connections with monoids of acyclic digital circuits. We show that the monoid $M_{2,1}$ (based on partial functions) is not embeddable into Thompson's monoid ${\sf tot}M_{2,1}$, but that ${\sf tot}M_{2,1}$ has a submonoid that maps homomorphically onto $M_{2,1}$. This leads to an efficient completion algorithm for partial functions and partial circuits. We show that the union of partial circuits with disjoint domains is an element of $M_{2,1}$, and conversely, every element of $M_{2,1}$ can be decomposed efficiently into a union of partial circuits with disjoint domains.

Some properties of Higman-Thompson monoids and digital circuits

TL;DR

The paper builds a unified framework tying Higman–Thompson monoids to acyclic circuit models, showing that monoids such as and naturally arise as right-ideal morphisms and as circuit-input/output maps. It proves finite generation for and related monoids, while establishing non-finite generation for several fixed-length and plep/tlep variants. The work also develops completion procedures (deterministic, inverse-homomorphic, generator-based, and circuit-based), and provides a robust representation theory: any element can be decomposed into a finite unambiguous union of partial circuits with disjoint domains, bridging algebraic and circuit viewpoints. Collectively, these results reveal a deep equivalence between Thompson monoids and circuit-model computations, with concrete algorithms for completion and decomposition and a detailed complexity landscape for related decision problems.

Abstract

We define various monoid versions of the R. Thompson group , and prove connections with monoids of acyclic digital circuits. We show that the monoid (based on partial functions) is not embeddable into Thompson's monoid , but that has a submonoid that maps homomorphically onto . This leads to an efficient completion algorithm for partial functions and partial circuits. We show that the union of partial circuits with disjoint domains is an element of , and conversely, every element of can be decomposed efficiently into a union of partial circuits with disjoint domains.
Paper Structure (28 sections, 58 theorems, 8 equations)

This paper contains 28 sections, 58 theorems, 8 equations.

Key Result

Lemma 3.4

(equivalent definition of input-length). Let $\Gamma$ be a finite set such that $\,\Gamma \cup \tau$ generates ${\cal RM}^{\sf fin}$ (or ${\sf tot}{\cal RM}^{\sf fin}$, ${\sf tfl}{\cal RM}^{\sf fin}$, ${\sf pfl}{\cal RM}^{\sf fin}$), and let $\,u = u_k \,\ldots\,u_1 \in$$(\Gamma \cup \tau)^*$ with $

Theorems & Definitions (80)

  • Definition 2.1
  • Definition 2.2
  • Definition 2.3
  • Definition 2.4
  • Definition 3.1
  • Definition 3.2
  • Definition 3.3
  • Lemma 3.4
  • Definition 3.5
  • Definition 3.6
  • ...and 70 more