Abstract
We construct a series of finitely presented semigroups. The centers of these semigroups encode uniquely up to rigid ambient isotopy in 3-space all non-oriented spatial graphs. This encoding is obtained by using three-page embeddings of graphs into the product of the line with the cone on three points. By exploiting three-page embeddings we introduce the notion of the three-page complexity for spatial graphs. This complexity satisfies the properties of finiteness and additivity under natural operations.
Introduction: From Knotted Networks to Algebra
Spatial graphs—networks of vertices and edges knotted and linked in three-dimensional space—represent a significant challenge in topology. Unlike simple knots (closed loops), spatial graphs can have complex branching structures, making their classification notoriously difficult. A major hurdle is translating these "rubbery" 3D shapes into a formal language suitable for algebraic manipulation or computational analysis. While Reidemeister moves describe how knot diagrams can be deformed without changing the underlying topology, a universal algebraic language for graphs has been lacking.
This paper bridges the gap between topology and algebra by showing that any spatial graph can be uniquely represented as a word in a finitely presented semigroup. By embedding these graphs into a "three-page book" structure, the research introduces a new "three-page complexity" measure that is both finite and additive, offering a systematic way to classify and compare complex knotted structures.
The Three-Page Encoding Method
The core of this approach is the "three-page book" structure, building on techniques originally developed by Dynnikov for links. A three-page book consists of a central axis with three half-planes (pages) radiating from it. The paper proves that any spatial graph can be embedded in this structure such that all vertices lie on the axis and all edges are "arches" passing through the pages.
Semigroup Construction and Topological Relations
The author defines an alphabet where letters represent local graph components: crossings, vertices of a specific degree, and arches. A set of algebraic relations is then established to correspond exactly to topological deformations, specifically rigid ambient isotopy. Rigid isotopy is a crucial distinction for physical modeling (e.g., proteins), as it prevents edges from "spinning" around a vertex like a propeller.
The resulting semigroup, $RSG_n$, has a remarkable property: the elements that represent valid spatial graphs are exactly those in the center of the semigroup (elements that commute with all others). Furthermore, the paper provides an algorithm to verify if a word represents a valid spatial graph in linear time relative to the word's length.
Three-Page Complexity Theory
The "three-page complexity" ($tp(G)$) is defined based on the "arch number"—the minimum number of arches required to represent the graph in the three-page book. This complexity measure is remarkably well-behaved, exhibiting additivity under natural operations such as disjoint unions ($G \sqcup H$) or vertex sums ($G * H$). This means the complexity of a large, tangled network is essentially the sum of the complexities of its constituent parts.
Key Results and Classification
The theoretical framework leads to several significant results:
- Unique Encoding: Theorem 1 proves that the center of the constructed semigroup $RSG_n$ uniquely encodes all non-oriented spatial graphs up to rigid isotopy.
- Complexity Properties: Theorem 3 establishes the finiteness and additivity of the $tp(G)$ function.
- New Classifications: Using this method, the author successfully classified all singular knots (knots allowed to intersect themselves at finite points) with one singular point and up to 6 crossings.
- Generalization of Brunn’s Result: The work extends a 19th-century theorem, demonstrating that any spatial graph can be projected onto a plane with only a single singular point, provided non-rigid deformations are allowed.
Limitations and Future Directions
While the paper successfully maps graphs to words, the "word problem"—deciding if two different words represent the same graph—remains a significant computational challenge. The author notes that while groups are often easier to study than semigroups, a group-based version of this theory would lose critical information about edge tangling.
Future research is directed toward finding better lower bounds for the three-page complexity using other topological invariants, such as the Alexander polynomial, to further refine the classification of complex spatial networks.
Summary updated