Vertex-Transitive Graphs Overview

• Vertex-transitive graphs are graphs whose automorphism group acts transitively on all vertices, making each vertex structurally identical.
• They encompass well-known families like Cayley, coset, and Haar graphs, constructed via group actions and offering rich examples in symmetry analysis.
• Ongoing research leverages these properties to classify graph isomorphisms, study coloring bounds, and explore Hamiltonicity in both finite and infinite cases.

A vertex-transitive graph is a finite or infinite (generally simple and undirected) graph whose automorphism group acts transitively on its vertex set; that is, for any two vertices, there exists an automorphism mapping one to the other. This high symmetry has profound implications across structural theory, algorithmic classification, coloring problems, and isomorphism phenomena. Vertex-transitive graphs encompass Cayley graphs and extend into non-Cayley families, with connections to group actions, coset graphs, and Haar graphs. Their study interfaces with permutation group theory, combinatorics, algebraic graph theory, and geometric configurations.

1. Definitions, Structural Framework, and Examples

Let $\Gamma = (V,E)$ be a simple undirected graph. The automorphism group $\mathrm{Aut}(\Gamma)$ is the set of adjacency-preserving bijections on $V$. $\Gamma$ is vertex-transitive when $\mathrm{Aut}(\Gamma)$ acts transitively; that is, for any $u,v \in V$, there exists $\phi \in \mathrm{Aut}(\Gamma)$ with $\phi(u)=v$.

Many vertex-transitive graphs are constructed via group actions:

• Cayley graphs: For group $G$, subset $S \subseteq G\setminus\{1\}$ (usually $S = S^{-1}$), $\mathrm{Cay}(G,S)$ has vertex set $G$ and edges $\{g,gs\}$ for $g \in G$, $s \in S$. $G$ acts regularly on vertices (Chen et al., 2016).
• Coset graphs: Generalized as $\mathrm{Cos}(G,H,HSH)$, with vertex set $G/H$ (right cosets) and edges $Hg \sim Hh$ iff $h g^{-1} \in H S H$.
• Haar graphs: For group $G$ and subset $S \subseteq G \setminus \{1\}$, the bipartite graph $H(G,S)$ has vertices $G \times \{0,1\}$ and edges between $(g,0)$ and $(sg,1)$ for $g \in G$, $s \in S$ (Conder et al., 2016).

Vertex-transitivity does not imply edge- or arc-transitivity; finer analysis of automorphism group action on ordered edges distinguishes arc-types (Conder et al., 2015). Important examples include:

• Cayley graphs of abelian/nonabelian groups;
• The Petersen, Coxeter, Dodecahedron graphs (Platonic graphs, generically vertex-transitive but not always Cayley);
• Bi-Cayley and Haar graphs with symmetry beyond Cayley regularity.

2. Classification, Isomorphism, Group Actions

Vertex-transitive graphs align closely with permutation group theory: given $\Gamma$ and $G \leq \mathrm{Aut}(\Gamma)$ acting transitively, $\Gamma$ is $G$-vertex-transitive. Orbitals (orbits of $G$ on $V \times V$) structure the possible edge frameworks. For cubic (valency $3$) and tetravalent ($4$) graphs, full enumerations up to moderate order have been computed, uncovering thousands to millions of nonisomorphic vertex-transitive structures (Potocnik et al., 2012, Holt et al., 2018).

Isomorphism theory initially focused on Cayley graphs via the CI (Cayley Isomorphism) property: $\mathrm{Cay}(G,S) \cong \mathrm{Cay}(G,T)$ iff $T = S^\alpha$ for some $\alpha \in \mathrm{Aut}(G)$. GI-Groups (Group-Automorphism Inducing Isomorphism) generalize this to coset graphs: two $G$-vertex-transitive graphs $\mathrm{Cos}(G,H,HSH)$ and $\mathrm{Cos}(G,H,HTH)$ are isomorphic iff there is $\alpha \in \mathrm{Aut}(G)$ fixing $H$ setwise and mapping $(H S H)^\alpha = H T H$ (Chen et al., 2016).

The coset-graph conjugacy criterion states: $\Gamma$ is a GI-graph of $G$ iff every permutation-isomorphic subgroup of $\mathrm{Aut}(\Gamma)$ is conjugate to $G$. Classification results indicate all DCI-groups (Directed Cayley Isomorphism) with normality are DGI (Directed GI) (Chen et al., 2016). Notably, new families of symmetric vertex-transitive graphs exist that are neither Cayley nor GI (e.g., construction on $40$ vertices with $\mathrm{Aut}=\mathrm{PSL}_4(3)$) (Chen et al., 2016).

3. Families Beyond Cayley: Haar, Bi-Cayley, and Non-Cayley Vertex-Transitive Graphs

Haar graphs generalize bipartite Cayley structures. Every bipartite Cayley graph is a Haar graph, but not every Haar graph is Cayley (Conder et al., 2016). Vertex-transitive Haar graphs arise via sophisticated group actions preserving two orbits. For trivalent graphs, infinite families $D(n,r)$ are constructed: $D(n,r)$ is vertex-transitive exactly when $n$ is even and $r^2 \equiv \pm 1 \pmod{n/2}$; it is non-Cayley when $r^2 \equiv -1 \pmod{n/2}$. The smallest such example is the Kronecker cover over the dodecahedral graph, $D(10,2)$ (order $40$), which is arc-transitive, non-Cayley Haar (Conder et al., 2016).

For tetravalent graphs of order $6p$ ($p$ prime), a complete classification identifies all such connected vertex-transitive non-Cayley graphs: an infinite family of bi-Cayley graphs $X_{3,p,t}$ for $p \ge 11$ with $t^2 \equiv -1\pmod p$, and nine sporadic graphs for $p\le7$ including Coxeter, Desargues, and dodecahedron graphs (Arezoomand et al., 2022). These constructions rely on semiregular group actions and intricate automorphism arguments.

4. Coloring and Chromatic Properties

Vertex-transitive graphs are highly constrained in coloring; let $\chi(G)$ denote chromatic number, $\omega(G)$ clique number, $\Delta(G)$ maximum degree. The sharp conjecture states

$$\chi(G) \leq \max\left\{ \omega(G), \left\lceil \frac{5\Delta(G) + 3}{6} \right\rceil \right\}$$

proved tight for line graphs of odd cycles with duplicated edges. The fractional relaxation $\chi_f(G) = |V(G)|/\alpha(G)$ also admits strong bounds: $$\chi_f(G) \leq \max \left\{ \omega(G), \frac{5}{6}(\Delta(G) + 1) \right\}$$ (Cranston et al., 2014). For large maximum degree, Borodin-Kostochka holds: if $\Delta(G)\geq 13$ and $K_{\Delta(G)}\not\subseteq G$, then $\chi(G) \leq \Delta(G) - 1$. These results exploit clique partitioning, independent transversal lemmas, and connectivity arguments peculiar to vertex-transitivity.

5. Hamiltonicity, Motion, and Matching Phenomena

Vertex-transitive graphs are natural candidates for Hamilton cycles and decompositions. Nevertheless, there are infinite families of vertex-transitive (and Cayley) graphs with no Hamilton decomposition, constructed via $K(mX)$ method: taking $m$ parallel copies of $X$, then the corresponding arc graph is vertex-transitive if $X$ is arc-transitive. For example, $K(2F_8)$ (order $48$, valency $6$) has no Hamilton decomposition (Bryant et al., 2014). Such results refute any naive generalization of Hamilton decomposition to nonabelian Cayley graphs.

For motion (minimal number of vertices moved by any nontrivial automorphism), classified constructions show that, except for the trivial cases, only lexicographic products over cyclic circulants have prime motion, and motion-$4$ graphs have explicit product or matched block structures (Montero et al., 2024). This links motion to the primitivity and imprimitive block structure of the automorphism group.

In infinite settings, every countable, connected, vertex-transitive graph admits a perfect matching. The proof leverages exhaustion, automorphism adjustments, maximally matchedness, and compactness-style arguments (Georgakopoulos et al., 2020).

6. Cores, CIS Property, and Identifying Codes

The core of a vertex-transitive graph, defined as a minimal induced subgraph to which there is a homomorphism, must have size dividing the order of the whole graph (Hahn-Tardif theorem). For normal Cayley graphs and those with core of half their size, the vertex set can be partitioned into disjoint copies of the core. If the core is less than half, partitions may fail (e.g., line graphs of $K_{2n}$) (Roberson, 2013).

Vertex-transitive CIS (Cliques Intersect Stable sets) graphs belong to an intersection-theoretic subclass, characterized by being well-covered, co-well-covered, and $\alpha(G)\omega(G)=|V(G)|$, with infinite families and classification up to valency $7$ (Dobson et al., 2014).

For identifying codes, vertex-transitive graphs permit explicit calculation of the fractional code number via symmetry: the uniform distribution attains the LP optimum, and for generalized quadrangles $GQ(s,t)$, code sizes attain exponents $1/3$, $1/4$, or $2/5$, much below general upper bounds, indicating exceptional efficiency; the separation and domination requirements formalize these codes (Gravier et al., 2014).

7. Extensions, Arc-Types, and Quantum Symmetry

Arc-type, marking the partition of valency into self-paired and paired orbits of the automorphism group, provides refined structural typology for vertex-transitive graphs beyond classical cubic and quartic cases. Almost all feasible arc-type partitions are realisable via suitable product constructions, except for trivial cases (Conder et al., 2015).

Uniform vertex-transitivity is a sharpened notion requiring n automorphisms such that each maps every base vertex to each target exactly once collectively—a property strictly intermediate between Cayley and vertex-transitive graphs. Its equivalence to the existence of a size-$n$ clique in the derangement graph of automorphisms connects to quantum automorphism theory and combinatorial blocking phenomena (Schmidt et al., 2019).

Partite-presented graphs subsume classical Cayley graphs: every connected, countable vertex-transitive graph admits a group-like presentation, possibly with vertex-dependent relators, extending algebraic techniques to the broad vertex-transitive universe (Georgakopoulos et al., 2020).

---

Vertex-transitive graphs serve as a benchmark for the study of finite permutation group actions, combinatorial optimization, symmetric coverings, coloring, and isomorphism classification. Ongoing research pursues complete classification within specific valencies/orders, the quantification of non-Cayley families, implications for quantum symmetry, and the intersection of group-theoretic and combinatorial invariants. Open problems include classification of GI-groups, detailed enumeration at higher valencies, and extension of code and partition theories to infinite and mixed symmetry settings (Chen et al., 2016).