arXiv++ Combinatorics

These lecture notes are on automorphism groups of Cayley graphs and their applications to optimal fault-tolerance of some interconnection networks. We first give an introduction to automorphisms of graphs and an introduction to Cayley graphs. We then discuss automorphism groups of Cayley graphs. We prove that the vertex-connectivity of edge-transitive graphs is maximum possible. We investigate the automorphism group and vertex-connectivity of some families of Cayley graphs that have been considered for interconnection networks; we focus on the hypercubes, folded hypercubes, Cayley graphs generated by transpositions, and Cayley graphs from linear codes. New questions and open problems are also discussed.

Given a graph $Y$ on $n$ vertices and a desired level of fault-tolerance $k$, an objective in fault-tolerant system design is to construct a supergraph $X$ on $n + k$ vertices such that the removal of any $k$ nodes from $X$ leaves a graph containing $Y$. In order to reconfigure around faults when they occur, it is also required that any two subsets of $k$ nodes of $X$ are in the same orbit of the action of its automorphism group. In this paper, we prove that such a supergraph must be the complete graph. This implies that it is very expensive to have an interconnection network which is $k$-fault-tolerant and which also supports automorphic reconfiguration. Our work resolves an open problem in the literature. The proof uses a result due to Cameron on $k$-homogeneous groups.

Let $S$ be a set of transpositions generating the symmetric group $S_n$. The transposition graph of $S$ is defined to be the graph with vertex set $\{1,\ldots,n\}$, and with vertices $i$ and $j$ being adjacent in $T(S)$ whenever $(i,j) \in S$. In the present note, it is proved that two transposition graphs are isomorphic if and only if the corresponding two Cayley graphs are isomorphic. It is also proved that the transposition graph $T(S)$ is edge-transitive if and only if the Cayley graph $Cay(S_n,S)$ is edge-transitive.

\noindent The augmented cube graph $AQ_n$ is the Cayley graph of $\mathbb{Z}_2^n$ with respect to the set of $2n-1$ generators $\{e_1,e_2, \ldots,e_n, 00\ldots0011, 00\ldots0111, 11\ldots1111 \}$. It is known that the order of the automorphism group of the graph $AQ_n$ is $2^{n+3}$, for all $n \ge 4$. In the present paper, we obtain the structure of the automorphism group of $AQ_n$ to be \[ \Aut(AQ_n) \cong \mathbb{Z}_2^n \rtimes D_8~~(n \ge 4),\] where $D_8$ is the dihedral group of order 8. It is shown that the Cayley graph $AQ_3$ is non-normal and that $AQ_n$ is normal for all $n \ge 4$. We also analyze the clique structure of $AQ_4$ and show that the automorphism group of $AQ_4$ is isomorphic to that of $AQ_3$: \[ \Aut(AQ_4) \cong \Aut(AQ_3) \cong (D_8 \times D_8) \rtimes C_2.\] All the nontrivial blocks of $AQ_4$ are also determined.

The Johnson graph $J(n,i)$ is defined to the graph whose vertex set is the set of all $i$-element subsets of $\{1,\ldots,n\}$, and two vertices are joined whenever the cardinality of their intersection is equal to $i-1$. In Ramras and Donovan [\emph{SIAM J. Discrete Math}, 25(1): 267-270, 2011], it is conjectured that if $n=2i$, then the automorphism group of the Johnson graph $J(n,i)$ is $S_n \times \langle T \rangle$, where $T$ is the complementation map $A \mapsto \{1,\ldots,n\} \setminus A$. We resolve this conjecture in the affirmative. The proof uses only elementary group theory and is based on an analysis of the clique structure of the graph.

The modified bubble-sort graph of dimension $n$ is the Cayley graph of $S_n$ generated by $n$ cyclically adjacent transpositions. In the present paper, it is shown that the automorphism group of the modified bubble sort graph of dimension $n$ is $S_n \times D_{2n}$, for all $n \ge 5$. Thus, a complete structural description of the automorphism group of the modified bubble-sort graph is obtained. A similar direct product decomposition is seen to hold for arbitrary normal Cayley graphs generated by transposition sets.

The complete transposition graph is defined to be the graph whose vertices are the elements of the symmetric group $S_n$, and two vertices $α$ and $β$ are adjacent in this graph iff there is some transposition $(i,j)$ such that $α=(i,j) β$. Thus, the complete transposition graph is the Cayley graph $\Cay(S_n,S)$ of the symmetric group generated by the set $S$ of all transpositions. An open problem in the literature is to determine which Cayley graphs are normal. It was shown recently that the Cayley graph generated by 4 cyclically adjacent transpositions is not normal. In the present paper, it is proved that the complete transposition graph is not a normal Cayley graph, for all $n \ge 3$. Furthermore, the automorphism group of the complete transposition graph is shown to equal \[ \Aut(\Cay(S_n,S)) = (R(S_n) \rtimes \Inn(S_n)) \rtimes \mathbb{Z}_2, \] where $R(S_n)$ is the right regular representation of $S_n$, $\Inn(S_n)$ is the group of inner automorphisms of $S_n$, and $\mathbb{Z}_2 = \langle h \rangle$, where $h$ is the map $α\mapsto α^{-1}$.

Let $S$ be a set of transpositions such that the girth of the transposition graph of $S$ is at least 5. It is shown that the automorphism group of the Cayley graph of the permutation group $H$ generated by $S$ is the semidirect product $R(H) \rtimes \Aut(H,S)$, where $R(H)$ is the right regular representation of $H$ and $\Aut(H,S)$ is the set of automorphisms of $H$ that fixes $S$ setwise. Furthermore, if the connected components of the transposition graph of $S$ are isomorphic to each other, then $\Aut(H,S)$ is isomorphic to the automorphism group of the line graph of the transposition graph of $S$. This result is a common generalization of previous results by Feng, Ganesan, Harary, Mirafzal, and Zhang and Huang. As another special case, we obtain the automorphism group of the extended cube graph that was proposed as a topology for interconnection networks.

These lecture notes provide an introduction to automorphism groups of graphs. Some special families of graphs are then discussed, especially the families of Cayley graphs generated by transposition sets.

Let $S$ be a set of transpositions that generates the symmetric group $S_n$, where $n \ge 3$. The transposition graph $T(S)$ is defined to be the graph with vertex set $\{1,\ldots,n\}$ and with vertices $i$ and $j$ being adjacent in $T(S)$ whenever $(i,j) \in S$. We prove that if the girth of the transposition graph $T(S)$ is at least 5, then the automorphism group of the Cayley graph $\Cay(S_n,S)$ is the semidirect product $R(S_n) \rtimes \Aut(S_n,S)$, where $\Aut(S_n,S)$ is the set of automorphisms of $S_n$ that fixes $S$. This strengthens a result of Feng on transposition graphs that are trees. We also prove that if the transposition graph $T(S)$ is a 4-cycle, then the set of automorphisms of the Cayley graph $\Cay(S_4,S)$ that fixes a vertex and each of its neighbors is isomorphic to the Klein 4-group and hence is nontrivial. We thus identify the existence of 4-cycles in the transposition graph as being an important factor in causing a potentially larger automorphism group of the Cayley graph.

A problem of practical and theoretical interest is to determine or estimate the diameter of various families of Cayley networks. The previously known estimate for the diameter of Cayley graphs generated by transposition trees is an upper bound given in the oft-cited paper of Akers and Krishnamurthy (1989). In this work, we first assess the performance of their upper bound. We show that for every $n$, there exists a tree on $n$ vertices, such that the difference between the upper bound and the true diameter value is at least $n-4$. Evaluating their upper bound takes time $Ω(n!)$. In this paper, we provide an algorithm that obtains an estimate of the diameter, but which requires only time $O(n^2)$; furthermore, the value obtained by our algorithm is less than or equal to the previously known diameter upper bound. Such an improvement to polynomial time, while still performing at least as well as the previous bound, is possible because our algorithm works directly with the transposition tree on $n$ vertices and does not require examining any of the permutations. We also provide a tree for which the value computed by our algorithm is not necessarily unique, which is an important result because such examples are quite rare. For all families of trees we have investigated so far, each of the possible values computed by our algorithm happens to also be an upper bound on the diameter.

Let $G$ be a simple undirected graph. The regular number of $G$ is defined to be the minimum number of subsets into which the edge set of $G$ can be partitioned so that the subgraph induced by each subset is regular. In this work, we obtain the regular number of some families of graphs and discuss some general bounds on this parameter. Also, some of the lower or upper bounds proved in \cite{Kulli:Janakiram:Iyer:2001} are shown here to hold with equality.

Let $Γ$ be a Cayley graph of the permutation group generated by a transposition tree $T$ on $n$ vertices. In an oft-cited paper \cite{Akers:Krishnamurthy:1989} (see also \cite{Hahn:Sabidussi:1997}), it is shown that the diameter of the Cayley graph $Γ$ is bounded as $$\diam(Γ) \le \max_{π\in S_n}{c(π)-n+\sum_{i=1}^n \dist_T(i,π(i))},$$ where the maximization is over all permutations $π$, $c(π)$ denotes the number of cycles in $π$, and $\dist_T$ is the distance function in $T$. In this work, we first assess the performance (the sharpness and strictness) of this upper bound. We show that the upper bound is sharp for all trees of maximum diameter and also for all trees of minimum diameter, and we exhibit some families of trees for which the bound is strict. We then show that for every $n$, there exists a tree on $n$ vertices, such that the difference between the upper bound and the true diameter value is at least $n-4$. Observe that evaluating this upper bound requires on the order of $n!$ (times a polynomial) computations. We provide an algorithm that obtains an estimate of the diameter, but which requires only on the order of (polynomial in) $n$ computations; furthermore, the value obtained by our algorithm is less than or equal to the previously known diameter upper bound. This result is possible because our algorithm works directly with the transposition tree on $n$ vertices and does not require examining any of the permutations (only the proof requires examining the permutations). For all families of trees examined so far, the value $β$ computed by our algorithm happens to also be an upper bound on the diameter, i.e. $$\diam(Γ) \le β\le \max_{π\in S_n}{c(π)-n+\sum_{i=1}^n \dist_T(i,π(i))}.$$

Let $Γ$ be a Cayley graph generated by a transposition tree. A natural problem is to understand how the properties of the Cayley graph depend on those of the underlying transposition tree. We focus here on diameter and distance related questions. We examine some related bounds and prove some properties about them. This paper has been withdrawn and a new version with these results and further extensions is at abs/1111.3114

For a given undirected graph $G$, an \emph{ordered} subset $S = {s_1,s_2,...,s_k} \subseteq V$ of vertices is a resolving set for the graph if the vertices of the graph are distinguishable by their vector of distances to the vertices in $S$. While a superset of any resolving set is always a resolving set, a proper subset of a resolving set is not necessarily a resolving set, and we are interested in determining resolving sets that are minimal or that are minimum (of minimal cardinality). Let $Q^n$ denote the $n$-dimensional hypercube with vertex set ${0,1}^n$. In Erdös and Renyi (Erdos & Renyi, 1963) it was shown that a particular set of $n$ vertices forms a resolving set for the hypercube. The main purpose of this note is to prove that a proper subset of that set of size $n-1$ is also a resolving set for the hypercube for all $n \ge 5$ and that this proper subset is a minimal resolving set.

Given a weighted graph $G_\bx$, where $(x(v): v \in V)$ is a non-negative, real-valued weight assigned to the vertices of G, let $B(G_\bx)$ be an upper bound on the fractional chromatic number of the weighted graph $G_\bx$; so $χ_f(G_\bx) \le B(G_\bx)$. To investigate the worst-case performance of the upper bound $B$, we study the graph invariant $$β(G) = \sup_{\bx \ne 0} \frac{B(G_\bx)}{χ_f(G_\bx)}.$$ \noindent This invariant is examined for various upper bounds $B$ on the fractional chromatic number. In some important cases, this graph invariant is shown to be related to the size of the largest star subgraph in the graph. This problem arises in the area of resource estimation in distributed systems and wireless networks; the results presented here have implications on the design and performance of decentralized communication networks.