Abstract
Let \(n\ge 2\) be an integer. The Grimaldi graph G(n) is defined by taking the elements of the set \(\{ 0, \ldots , n-1 \}\) as vertices. Two distinct vertices x and y are adjacent in G(n) if and only if \(\gcd (x+y, n) =1\). In this paper, we examine the Betti numbers of the edge ideals of these graphs and their complements.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Let \(R = \Bbbk [x_1,\ldots , x_n]\) be the polynomial ring over an arbitrary field \(\Bbbk \). It is well known that associated to any homogeneous ideal I of R there is a minimal graded free resolution
where \(R(-j)\) denotes the free R-module obtained by shifting the degrees of R by j. The quantity \(\beta _{i,j}(R/I)\), which is called the (i, j)th graded Betti number of R/I, is equal to the number of generators of degree j in the ith syzygy module. The graded Betti numbers are collected in the Betti table, in which the entry at column i and row j is \(\beta _{i,i+j}(R/I)\). The regularity and the projective dimension are two important invariants associated with R/I that can be read off from the minimal free resolution. The regularity of R/I is defined by \({{\,\textrm{reg}\,}}(R/I)=\max \{j-i\mid \beta _{i,j}(R/I)\ne 0\}\) and the projective dimension of R/I is defined by \({{\,\textrm{pd}\,}}(R/I) = \max \{i\mid \beta _ {i,j} (R/I) \ne 0 \text { for some } j\}.\)
1.1 Minimal Resolutions and the Edge Ideals
Let us now connect the above-mentioned notions with the ideals arising from graphs. In order to do this, let G be a finite simple graph with the vertex set \(V(G)=\{x_1, \ldots , x_n\}\) and the edge set E(G). One may associate to the graph G a quadratic squarefree monomial ideal
which is called the edge ideal of G. One of the central problems in combinatorial commutative algebra is a description of the minimal resolutions of these ideals. The structure of the resulting resolutions is very poorly understood. This problem has been studied by many authors (e.g., [8, 10, 13,14,15]). In particular, Katzman [15] and Kimura [16] have provided some results on the nonvanishing of the graded Betti numbers. For the Ferrers graphs, Corso and Nagel [5] have proved that their edge ideals have linear resolutions, and furthermore they have given an explicit formula for the Betti numbers of such ideals. Recently, Fröberg [8] has described the Betti numbers of the edge ideals of fat forests. We refer the reader to [7, 10, 17, 20] for other problems and results on this area.
1.2 A Class of Graphs Due to Grimaldi
Let \(n \ge 2\) be an integer. In 1990, Grimaldi [9] defined a graph G(n) based on the elements of \([n]=\{0,\ldots ,n-1\}\) and the notion of coprimeness, that is, a graph with a number-theoretical nature. The vertices of G(n) are the elements of [n] and distinct vertices x and y are defined to be adjacent if and only if \(\gcd (x+y,n)=1\). This graph is called a Grimaldi graph. By letting \(\varphi \) as the Euler’s phi function, it follows that when n is even, G(n) is a \(\varphi (n)\)-regular graph, whereas it is \((\varphi (n), \varphi (n)-1)\)-biregular when n is odd. This means by [1] that G(n) is an almost regular graph. Also, when \(n \ne 2\) is even, G(n) can be expressed as the union of \(\varphi (n)/2\) Hamiltonian cycles, that is, cycles containing all the vertices of the graph. The odd case is not quite so easy, but the structure is clear and the results are similar to the even case. Grimaldi has given an explicit formula for the chromatic polynomial of G(p) and has investigated some properties of the graph \(G(p^{ \alpha })\), where p is a prime and \(\alpha \ge 2\) is an integer. Hoang et al. [13] have characterized Cohen–Macaulay and Gorenstein properties for a kind of circulant graphs and their complements. When n is even, these latter graphs coincide with Grimaldi graphs and their complements. The complement of G(n), denoted by \(G'(n)\), is a graph whose vertex set is [n] in which two distinct vertices x and y are adjacent if and only if \(\gcd (x+y,n)\ne 1\). In [2], we have characterized when these graphs are well-covered, Cohen–Macaulay, vertex-decomposable, or Gorenstein.
1.3 The Aim of this Paper
In this paper, we focus on finding the Betti numbers of the edge ideals of G(n)’s and \(G'(n)\)’s. The rest of the paper is organized as follows. In Sect. 2, we provide some basic notation and terminology about the Betti numbers of the Stanley–Reisner rings and the edge ideals of graphs. In Sect. 3, we give a useful technique to compute the Betti numbers of the edge ideal of \(G(p^{\alpha } )\), where p is a prime. Finally, in Sect. 4, we give a formula for the Betti numbers of the edge ideal of \(G'(n)\), when n is either even or a prime power.
2 Preliminaries
In this section, we introduce some basic notation which will be used in the sequel. We refer the reader to [18, 21] for detailed information about the combinatorial and algebraic backgrounds.
2.1 The Betti Numbers of the Stanley–Reisner Rings
A simplicial complex \(\Delta \) with the vertex set \(V(\Delta )= \{ v_1, \ldots , v_n\}\) is a collection of subsets of \(V(\Delta )\) such that \(F\in \Delta \) whenever \(F\subseteq F'\) for some \(F'\in \Delta \). The restriction of \(\Delta \) to a subset S of \(V(\Delta )\) is \(\Delta [S] = \{F\in \Delta \mid F \subseteq S\}\). For a given field \(\Bbbk \), we attach to \(\Delta \) the Stanley–Reisner ideal \(I_{\Delta }\) of \(\Delta \) to be the squarefree monomial ideal
in \(R =\Bbbk [x_1,\ldots ,x_n]\) and the Stanley–Reisner ring of \(\Delta \) to be the quotient ring \(\Bbbk [\Delta ] = R/I_{\Delta }\). This provides a bridge between combinatorics and commutative algebra (see [14]). We denote by \(\widetilde{H}_j(\Delta ; \Bbbk )\) the reduced homology group of \(\Delta \) with coefficients in \(\Bbbk \). A very useful result about the graded Betti numbers of a Stanley–Reisner ring is the following so-called Hochster’s formula (c.f. [11, Theorem 8.1.1]):
The dimension of a face \(F\in \Delta \) is given by \(\dim F = |F| - 1\). The dimension of \(\Delta \), denoted by \(\dim \Delta \), is the maximum dimension of all its faces. By letting \(d-1=\dim \Delta \), the f-vector of \(\Delta \) is the vector \((f_{-1}, f_0, \ldots , f_{d-1})\), where \(f_{-1}=1\) and \(f_i\) is the number of faces of dimension i. The reduced Euler characteristic of \(\Delta \), denoted by \(\widetilde{\chi }(\Delta )\), is defined to be \(\widetilde{\chi }(\Delta ) = \sum _{i=-1}^{d-1}(-1)^i f_i(\Delta ).\) The h-vector of \(\Delta \) is the vector \((h_0, \ldots , h_d)\), where
The Hilbert–Poincaré series of the R-module \(\Bbbk [\Delta ]\) is \(HP_{\Bbbk [\Delta ] }(t)= \sum _{i\ge 0} H_{\Bbbk [\Delta ]}(i) t^i\), where \(H_{\Bbbk [ \Delta ]}\) is the Hilbert function of \(\Bbbk [\Delta ]\). By [18, Corollary 1.5], this series can be expressed as follows:
Let \(\Gamma \) and \(\Lambda \) be two simplicial complexes with the disjoint vertex sets \(V(\Gamma )\) and \(V(\Lambda )\), respectively. We define their join on the vertex set \(V(\Gamma )\cup V(\Lambda )\) to be \(\Gamma * \Lambda = \{ \sigma \cup \tau \mid \sigma \in \Gamma ,\ \tau \in \Lambda \}.\) By using the Künneth’s formula (c.f. [3, Proposition 3.2]), we can describe the reduced homology of the join of two simplicial complexes in terms of the reduced homologies of the factors for each i as follows:
The following lemma gives us the Betti numbers of the Stanley–Reisner rings of simplicial complexes which are join of finitely many disjoint subcomplexes.
Lemma 2.1
([12], Lemma 1.2) Let \(\Delta \) be a simplicial complex. If \(\Delta =\Delta _1*\cdots *\Delta _m\), where the \(\Delta _i\)’s are disjoint subcomplexes of \(\Delta \), then
In particular, \({{\,\textrm{pd}\,}}(\Bbbk [ \Delta ]) = \sum _{k=1}^m {{\,\textrm{pd}\,}}(\Bbbk [\Delta _k])\) and \({{\,\textrm{reg}\,}}(\Bbbk [\Delta ]) = \sum _ {k=1}^m {{\,\textrm{reg}\,}}(\Bbbk [\Delta _k])\).
We now state and prove the following lemma for later use.
Lemma 2.2
Let \(\Delta _1\) and \(\Delta _2\) be two simplicial complexes. If \(V(\Delta _1) \cap V(\Delta _2) = \emptyset \), then \({{\,\textrm{reg}\,}}( \Bbbk [ \Delta _1 \cup \Delta _2 ] ) = \max \{ {{\,\textrm{reg}\,}}(\Bbbk [\Delta _1]), {{\,\textrm{reg}\,}}(\Bbbk [\Delta _2])\}.\)
Proof
Note that [11, Proposition 5.1.8] implies \(I_{\Delta _1} \cap I_{\Delta _2} = I_ {\Delta _1 \cup \Delta _2}\) and \(I_{\Delta _1} + I_{\Delta _2}= I_{ \Delta _1 \cap \Delta _2}\). Therefore, we obtain the following exact sequence:
Now, by [19, Proposition 18.6], \({{\,\textrm{reg}\,}}(\Bbbk [ \Delta _1 \cup \Delta _2]) = \max \{ {{\,\textrm{reg}\,}}(\Bbbk [\Delta _1]), {{\,\textrm{reg}\,}}( \Bbbk [ \Delta _2 ])\}\), as required. \(\square \)
Let \(\Delta _1\) and \(\Delta _2\) be two simplicial complexes with n and m vertices, respectively. It is known that if \(I_{\Delta _1} \subseteq \Bbbk [ x_1, \ldots , x_n ]\) is the Stanley–Reisner ideal of \(\Delta _1\) and \(I_{\Delta _2} \subseteq \Bbbk [ y_1, \ldots , y_m ]\) is the Stanley–Reisner ideal of \(\Delta _2\), then
is the Stanley–Reisner ideal of in \(\Bbbk [x_1,\ldots ,x_n,y_1,\ldots ,y_m]\). In this direction, Whieldon [22] has determined the graded Betti numbers of the edge ring . We close this subsection by stating it.
Lemma 2.3
([22], Lemma 5.4) Let \(\Delta _1\) and \(\Delta _2\) be two simplicial complexes with n and m vertices, respectively. If \(s \ge 2\) is an integer, then the following formulas hold true:
2.2 The Betti Numbers of the Edge Ideals of Graphs
In the sequel, by a graph we mean a finite undirected graph without loops or multiple edges. For a graph G, let V(G) denote the set of vertices of G and let E(G) denote the set of edges of G. A graph with just one vertex is referred to as trivial. All other graphs are nontrivial. A graph is called totally disconnected if it is either a null graph or it contains no edge. An edge \(e \in E(G)\) connecting two vertices x and y will also be written as \(\{x, y\}\). In this case, it is said that x and y are adjacent. For a subset S of V(G), we denote by G[S] the induced subgraph of G on the vertex set S and denote \(G\backslash S\) by \(G[V \backslash S]\). If \(S = \{x\}\), we write \(G\backslash x\) instead of \(G\backslash \{x\}\). The neighborhood of x in G is the set \(N_G(x)= \{y \in V(G) \mid \{x,y\} \in E(G)\},\) and the closed neighborhood of x is \(N_G[x] = \{x\} \cup N_G(x)\).
A bipartite graph is one whose vertex set can be partitioned into two disjoint parts in such a way that the two end vertices for each edge lie in distinct parts. A complete bipartite graph is one in which each vertex is joined to every vertex that is not in the same partition. The complete bipartite graph with exactly two parts of size m and n, is denoted by \(K_{m,n}\). A complete graph is a graph in which each pair of distinct vertices is joined by an edge. We denote the complete graph with n vertices by \(K_n\).
To every graph G with the vertex set \(V(G)=\{x_1,\ldots ,x_n\}\) and the edge set E(G) , one may associates the edge ideal I(G) of the polynomial ring \(\Bbbk [V(G)] =\Bbbk [x_1,\ldots ,x_n]\). Let \(\Delta (G)\) be the set of all independent sets of G. Then \(\Delta (G)\) is a simplicial complex, which is called the independence complex of G. It is easy to see that \(I_{ \Delta (G)} = I(G)\). Note that \(\Delta (G[W]) = \Delta (G)[W]\) for any \(W \subseteq V(G)\). Therefore, the Hochster’s formula is also applied to compute the Betti numbers of edge ideals. We write \(\beta _{i,j} (G)\), \({{\,\textrm{pd}\,}}(G)\), and \({{\,\textrm{reg}\,}}(G)\) as shorthand for \(\beta _{i,j} (\Bbbk [ \Delta (G)])\), \({{\,\textrm{pd}\,}}(\Bbbk [\Delta (G)])\), and \({{\,\textrm{reg}\,}}( \Bbbk [ \Delta (G)])\), respectively.
We close this section by recalling the following result which gives us the Betti numbers of the complete graph \(K_n\) and the complete bipartite graph \(K_{n,m}\).
Proposition 2.4
([14], Theorems 5.1.1 and 5.2.4) The following statements hold true:
-
(1)
The edge ring of \(K_{n}\) has a 2-linear resolution and \(\beta _{i,i+1}(K_n) = i\left( {\begin{array}{c}n\\ i+1\end{array}}\right) .\)
-
(2)
The edge ring of \(K_{n,m}\) has a 2-linear resolution and
$$\begin{aligned} \beta _{i,i+1}(K_{n,m})= \sum \limits _{\begin{array}{c} s+t=i+1\\ s,t\ge 1 \end{array}} \left( {\begin{array}{c}n\\ s\end{array}}\right) \left( {\begin{array}{c}m\\ t\end{array}}\right) . \end{aligned}$$
3 The Betti Numbers of the \(G(p^{\alpha })\)’s
In this section, we find some invariants regarding the graphs \(G(p^{\alpha } )\)’s, where p is a prime. In particular, we find the Betti numbers of these graphs. The case \(p=2\) is easy to handle, while the odd case is not quite so easy. For this latter case, we give a useful technique to compute the Betti numbers. Let us start with the following purely combinatorial lemma, which is useful for later use.
Lemma 3.1
Let \(m, n, i, j, k \in {{\mathbb {N}}}\), \(0 \le i \le n\), and \(0 \le j \le m\). Then the following equalities hold true:
-
(1)
\(\sum \limits _ {\begin{array}{c} i+j=k \\ i, j \ge 0 \end{array}} \left( {\begin{array}{c}n\\ i\end{array}}\right) \left( {\begin{array}{c}m\\ j\end{array}}\right) = \left( {\begin{array}{c}n+m\\ k\end{array}}\right) \).
-
(2)
\(\sum \limits _{\begin{array}{c} i+j=k \\ i, j \ge 0 \end{array}} ij \left( {\begin{array}{c}n\\ i+1\end{array}}\right) \left( {\begin{array}{c}m\\ j+1\end{array}}\right) = (k+1) \left\{ \frac{kmn-m^2-n^2+m+n}{(m+n-1) (m+n)}\left( {\begin{array}{c}m+n\\ k+2\end{array}}\right) + \left( {\begin{array}{c}n\\ k+2\end{array}}\right) + \left( {\begin{array}{c}m\\ k+2\end{array}}\right) \right\} .\)
Proof
(1): By the binomial expansion theorem we have the following equalities:
and \((1+x)^{n+m} = \sum _{k=0}^{n+m} \left( {\begin{array}{c}n+m\\ k\end{array}}\right) x^{k}.\) It is noteworthy that the coefficients of the monomials of degree k of two above expressions are \(\sum _{\begin{array}{c} i+j=k\\ i,j\ge 0 \end{array}} \left( {\begin{array}{c}n\\ i\end{array}}\right) \left( {\begin{array}{c}m\\ j\end{array}}\right) \) and \(\left( {\begin{array}{c}n+m\\ k\end{array}}\right) \). Hence, they are equal, as required.
(2): Note that \(\sum _{i=0}^{n} \left( {\begin{array}{c}n\\ i\end{array}}\right) x^{i} = (1+x)^n\), and thus, we obtain that
Therefore, \(\sum _{i=1}^{n-1} i \left( {\begin{array}{c}n\\ i+1\end{array}}\right) x^{i+1} = nx (1+x)^ {n-1} -(1+ x)^ n+1.\) On the other hand,
Hence, the coefficients of the monomials of degree \(k+2\) (\(0\le k\le m+n-2\)) of the above expression are
Moreover, the product of two polynomials \(nx(1+x)^{n-1}-(1+x)^n+1\) and \(mx(1+x)^{m-1}-(1+x)^m+1\) is
This implies that the coefficient of the monomials of degree \(k+2\) \((0 \le k \le m+n-2)\) of the above polynomial is
We now get the result by comparing the two coefficients above. \(\square \)
The following proposition deals with the graphs \(G( 2^{\alpha } )\)’s and gives us the projective dimension of these graphs.
Proposition 3.2
Let \(\alpha \) be a positive integer. Then the edge ring of \(G( 2^{\alpha } )\) has a 2-linear resolution. Moreover, we have
In particular, \({{\,\textrm{pd}\,}}( G(2^{\alpha }) ) = 2^{\alpha }-1.\)
Proof
Let \(A = \{0,2, \ldots , 2^{\alpha }-2\}\) and \(B = \{1,3, \ldots , 2^{\alpha }-1 \}\). If either \(x,y\in A\) or \(x,y\in B\), then \(x+y = 0 \pmod 2\), and thus, \(\{x, y\} \notin E(G(2^{\alpha }))\). On the other hand, if \(x\in A\) and \(y\in B\), then \(x+y = 1 \pmod 2\), and thus, \(\{x, y\} \in E(G(2^{\alpha }))\). Therefore, A and B are the maximal independent sets of \(G(2^{ \alpha })\), and furthermore, \(G( 2^{\alpha } )\) is a complete bipartite graph with bipartition (A, B). By Proposition 2.4(2), we obtain that
This, together with Lemma 3.1(1), completes the proof. \(\square \)
We now continue the paper with the odd case.
Proposition 3.3
Let p be an odd prime. Then \(\alpha (G(p))=2\), the edge ring of G(p) has a 2-linear resolution, and we have
In particular, \({{\,\textrm{pd}\,}}(G(p))=p-1\).
Proof
Since p is a prime, it is easy to see that the vertex 0 is adjacent to all of the other vertices in G(p), and \(\{x,y\}\notin E(G(p))\) for some \(x,y\in [p]\) if and only if \(x+y=p\). Thus, all of the maximal independent sets of G(p) are \(\{0\}\) and \(\{ i, p-i \}\) for all \(1 \le i \le \frac{p-1}{2}\). Therefore, the independence complex of G(p) is a disjoint union of \(\frac{p-1}{2}\) edges and one vertex. This means that the Hilbert series of \(\Bbbk [\Delta (G(p) ]\) is
This implies that \(\beta _ {p-1,p} (G(p)) = \frac{p-1}{2}\), \(\beta _ {p-2,p-1} (G(p)) = \frac{p^2-p-2}{2}\), and
for all \(1\le i\le p-3\), as required. \(\square \)
The following lemma gives us a decomposition for a class of Grimaldi graphs involving the join of graphs. Let us recall that the join of two graphs G and H, denoted by \(G*H\), is a graph formed from disjoint copies of G and H by connecting each vertex of G to each vertex of H.
Lemma 3.4
Let p be an odd prime and \(\alpha > 1\) be an integer. Then we have
where \(I_i = \{x \in [p^{\alpha }] \mid x=i \pmod p\}\) for every \(0 \le i \le p-1\), \(G[I_0]\) is a totally disconnected graph, and \(G[ I_i \cup I_{p- i}]\) is an induced graph of \(G(p^{\alpha })\) with the edge set
Proof
In order to prove the lemma, it is enough to prove the following claims:
Claim 1: \(G[I_0]\) is a totally disconnected graph.
Claim 2: \(G[I_i]\) is a complete graph for \(1\le i\le p-1\).
Claim 3: \(\{x,y\} \notin E(G(p^{\alpha }))\) for \(x\in I_{i}\) and \(y\in I_{p-i}\) with \(1\le i\le \frac{p-1}{2}\).
Claim 4: \(\{x,y\} \in E(G(p^{\alpha }))\) for \(x\in I_0\) and \(y\in I_{i} \cup I_{ p-i}\) with \(1\le i\le \frac{p-1}{2}\).
Claim 5: \(\{x,y\} \in E(G(p^{\alpha }))\) for \(x\in I_{i} \cup I_{p-i}\) and \(y\in I_{j} \cup I_{p-j}\) with \(1\le i\ne j\le \frac{p-1}{2}\).
Claim 1 is true since for every \(x, y \in I_0\), \(x+y=0 \pmod p\), and thus, x and y are not adjacent in \(G(p^{\alpha })\). Claim 2 is true since for every \(x, y \in I_i\) with \(1\le i\le p-1\), \(\gcd ( x+y, p^{ \alpha } )=1\), and thus, x is adjacent to y in \(G(p^{\alpha })\). Claim 3 is true since by the assumption, \(x+y=0 \pmod p\), and thus, x is not adjacent to y in \(G(p^{\alpha })\). Claim 4 is true since by the assumption, \(x+y = y \pmod p\), and thus, x is adjacent to y in \(G(p^{\alpha })\). Claim 5 is true since by the assumption, \(\gcd (x+y, p^ {\alpha }) = 1\), and thus, x is adjacent to y in \(G (p^ {\alpha } )\). \(\square \)
Proposition 3.5
Let p be an odd prime and \(\alpha > 1\) be an integer. Then \(\alpha (G(p^{ \alpha })) = p^ {\alpha -1}\) and the f-vector of \(\Delta ( G(p^ {\alpha }))\) is \((f_{-1},f_0,\ldots ,f_{p^{\alpha -1}-1})\), where \(f_{-1}=1\), \(f_0= p^{\alpha }\), \(f_1 = \frac{ (p^{\alpha }-1) p^{ \alpha -1}}{2}\), and \(f_i = \left( {\begin{array}{c}p^ {\alpha -1}\\ i+1\end{array}}\right) \) for all \(2 \le i \le p^{ \alpha -1 } - 1\). In particular,
Proof
By Lemma 3.4, all of the maximal independent sets of \(G(p^{\alpha })\) are \(I_0\) and \(I_{uv}=\{\{u,v\}\mid 0\le u<v\le p^{\alpha }-1, u+v=0\pmod p \}\). Thus, \(\alpha (G(p^{\alpha })) = p^{\alpha -1}\). We also obtain that \(f_i = \left( {\begin{array}{c}p^{\alpha -1}\\ i+1\end{array}}\right) \) for all \(2\le i\le p^{\alpha -1}-1\). Moreover, we claim that \(f_1=\frac{(p^{\alpha }-1)p^{\alpha -1}}{2}\). Indeed, by the structure of the graph \(G(p^{\alpha })\), for \(x\in V(G(p^{\alpha }))\), we have
Hence,
This implies that \( f_1 = \left( {\begin{array}{c}p^{\alpha }\\ 2\end{array}}\right) - |E(G(p^{\alpha }))| = \frac{(p^{\alpha }-1)p^{\alpha -1}}{2}.\) Now, these computations lead to
We know that \( \sum \limits _{i=0}^ {p^{\alpha -1}} (-1)^{i} \left( {\begin{array}{c}p^{\alpha -1}\\ i\end{array}}\right) = (1-1)^ {p^ {\alpha -1}} =0\), and so,
Hence, we conclude that
as required. \(\square \)
Lemma 3.6
Let p be an odd prime and \(\alpha > 1\) be an integer. Then \({{\,\textrm{reg}\,}}(G( p^{ \alpha } )) = 2\) and \({{\,\textrm{pd}\,}}(G(p^ {\alpha })) = p^{\alpha }-1\).
Proof
By considering \(G=G(p^{\alpha })\) and applying Lemma 3.4, we have \(\Delta (G) = \left<I_0\right> \cup \Delta _1 \cup \cdots \cup \Delta _{ \frac{p-1}{2}},\) where \(\left<I_0\right>\) is the simplex over the set \(I_0\) and each \(\Delta _i\) denotes the independence complex of the induced subgraph of \(G[I_{i} \cup I_{p-i}]\). Note that at this point, for every \(1 \le i \le \frac{p-1}{2}\), \(\dim \Delta _i=1\), and so, \(\Delta _i\) is regarded as a complete bipartite graph. Hence, \(G [I_i \cup I_{p-i}]\) is a disjoint union of two complete graphs. Thus, \({{\,\textrm{reg}\,}}(\Delta _i) = {{\,\textrm{reg}\,}}(G [I_i \cup I_{p-i}])=2\). Now, by Lemma 2.2, we obtain that \( {{\,\textrm{reg}\,}}( \Delta (G) ) = \max \{ {{\,\textrm{reg}\,}}( \left<I_0\right>), {{\,\textrm{reg}\,}}(\Delta _1), \ldots , {{\,\textrm{reg}\,}}( \Delta _ { \frac{p-1}{2}} ) \}= 2,\) as required. \(\square \)
Theorem 3.7
Let p be an odd prime and \(\alpha > 1\) be an integer. If \(G=G(p^{\alpha })\) and \( I_0 = \{x\in [p^{\alpha }]\mid x=0\pmod p\} \), then
and
Proof
By Lemma 3.4, we have \(G =G[I_0]* G[I_{1} \cup I_{p-1}] * \cdots * G [I_ {\frac{p-1}{2}} \cup I_{\frac{p+1}{2}}],\) where \(G[I_0]\) is a totally disconnected graph and for every \(1 \le i \le \frac{p-1}{2}\), \(G[I_i \cup I_ {p -i}]\) is an induced subgraph of G with the edge set
This implies that \(G \backslash I_0 = G[ I_{1} \cup I_{p-1} ] * \cdots *G [I_ {\frac{p-1}{2}} \cup I_{\frac{p+1}{2}}].\) Also, for every \(x\in I_0\), \(N_ G (x) = I_1 \cup \cdots \cup I_{p-1}\), and so, \(G\backslash N_G[x] = G [I_0 \backslash \{x\}]\) is a totally disconnected graph. Now, by [6, Lemma 3.1], we have \(I(G):x = I(G \backslash N_G[x]) + (N_G(x))\) and \((I(G),x) = I(G\backslash x) + (x).\) Thus, we obtain that
On the other hand, by applying the long exact sequence theorem to the short exact sequence \(0 \longrightarrow R/ (I(G):x) (-1) \overset{\cdot x}{\longrightarrow } R/I(G) \longrightarrow R/(I(G),x) \rightarrow 0,\) we obtain the following long exact sequence:
This, together with the fact that for \(j=i+1\) with \(i\ge 0\), \({{\,\textrm{Tor}\,}}_{i+1} (R/(I(G),x); \Bbbk )_{i+1}=0\) and \({{\,\textrm{Tor}\,}}_ {i-1}(R/(I(G):x)(-1);\Bbbk )_{i+1} = 0,\) lead us to the following short exact sequence:
By using the above observations, we have for every \(i \ge 1\),
Also, by [4, Remark 2.1], the equality \(\beta _ {i, i+1} (R/(I(G), x)) = \beta _ {i,i+1} (G \backslash x) + \beta _{i-1,i} (G \backslash x)\) holds true for every \(i \ge 1\). This implies the following recurrence formula for computing the Betti numbers of the graph G, where \(\beta _{0,1}(G\backslash x)=0\) and \(\beta _{1,2}(G) = (p^{\alpha }-p^{\alpha -1}) + \beta _ {1,2} (G \backslash x)\), as follows:
As a result of applying this process consecutively by replacing G by \(G \backslash x\), we obtain that \(\beta _ {0,1} (G) = \beta _ {0,1} (G \backslash I_0)=0\) and
In order to simplify the formula, we apply Lemma 3.1(1) and obtain
Moreover, if \(i\ge p^{\alpha -1}>t\), we have
This implies that
For the second formula, we have
This implies that \({{\,\textrm{Tor}\,}}_ i (R/I(G); \Bbbk )_ {i+2} \cong {{\,\textrm{Tor}\,}}_ i (R/(I(G),x); \Bbbk )_ {i+2}\), and so, by [4, Remark 2.1], \(\beta _ {i, i + 2} (G) = \beta _{i,i+2}(R/(I(G),x)) = \beta _ {i-1, i+1}( G \backslash x) + \beta _ {i, i+2} (G \backslash x).\) As a result of applying this process consecutively by replacing G by \(G \backslash x\), we obtain that
as required. \(\square \)
By keeping the notation of the previous theorem, its proof shows that \(G \backslash I_0 = G[I_{1} \cup I_{p-1}] * \cdots *G[I_{\frac{p-1}{2}} \cup I_{ \frac{p+1}{2}}].\) Also, by Lemma 3.4, we have where for every \(1 \le k \le \frac{p-1}{2}\), \(\Delta _k = \Delta (G[I_k \cup I_{p-k}])\). Thus, in order to finding the Betti numbers of \(G(p^{\alpha })\) by Lemma 2.3, we need to compute the Betti numbers of \(G[I_k \cup I_{p-k}]\) for all \(1\le k\le \frac{p-1}{2}\). To this end, we state and prove the following proposition.
Proposition 3.8
Let p be an odd prime and \(\alpha > 1\) be an integer. If \(G=G(p^{\alpha })\) and \( I_i = \{x\in [p^{\alpha }]\mid x=i \pmod p\}\) for \(0\le i\le p-1\), then
Proof
By the proof of Lemma 3.6, for every \(1\le k\le \frac{p-1}{2}\), the graph \(G[I_k\cup I_{p-k}]\) is a disjoint union of two complete graphs. Thus, \(\Delta (G[I_k\cup I_{p-k}]) = \Delta (G[I_k])* \Delta (G[I_{p-k}])\), where \(\Delta (-)\) denotes the independence complex of −. Since \({{\,\textrm{reg}\,}}(G[I_k\cup I_{p-k}])=2\) and \(\beta _{i,j}(G[I_k\cup I_{p-k}])=0\) for all \(j>i+2\), by Lemma 2.1, we obtain that
Now, Proposition 2.4(1) implies that
Note that, by Lemma 3.1(2),
and thus,
as required. \(\square \)
We conclude this section with the following example.
Example 3.9
Consider the graph \(G=G(3^2)\) and let \(I_0 = \{0,3,6\}\), \(I_1 = \{1,4,7\}\), and \(I_2 = \{2,5,8\}\). By applying Proposition 3.8 with \(p=3\) and \(\alpha =2\), we obtain the Betti table of \(G\backslash I_0 = G[I_1 \cup I_2]\) as follows:
Now, by Theorem 3.7, we have
and \(\beta _ {i,i+2}(G) = \beta _{i-3,i-1}(G \backslash I_0) + 3 \beta _ {i-2,i} (G\backslash I_0)+3\beta _{i-1,i+1}(G\backslash I_0)+\beta _{i,i+2}(G \backslash I_0).\) Thus, we obtain the Betti table of G as follows:
4 The Betti Numbers of the \(G'(n)\)’s
In this section, we focus on the \(G'(n)\)’s, the complements of the G(n)’s. Let us recall that \(G'(n)\) is the graph whose vertex set is [n] in which two distinct vertices x and y are adjacent if and only if \(\gcd (x+y,n)\ne 1\). It is obvious that \(G'(p)\), p prime, contains one isolated vertex and disjoint edges. By [2, Lemma 4.1], if n is even, \(G'(n)\) is well-covered. Because of this, we first consider the graph \(G'(n)\) with even n.
Lemma 4.1
Let n be an even integer. Then \(\alpha (G'(n)) = 2\) and the f-vector of \(\Delta (G'(n))\) is \((1, n, \frac{n}{2} \varphi (n))\), where \(\varphi \) is the Euler’s phi function.
Proof
By the proof of [2, Lemma 4.1], we obtain that \(\alpha (G'(n))=2\) and the f-vector of \(\Delta (G'(n))\) is \((f_{-1}, f_0, f_1)\), where \(f_{-1}=1\), \(f_0=n\), and \(f_1= |E(G(n))| = \frac{n}{2}\varphi (n)\), as required. \(\square \)
Proposition 4.2
Let n be an even integer. Then the following statements hold true:
-
(1)
The Hilbert series of \(R/I(G'(n))\) is \(HP_{R/I(G'(n))}(t) = \frac{1+(n-2)t + (1-n+\frac{n}{2}\varphi (n))t^2}{(1-t)^2}.\)
-
(2)
\({{\,\textrm{reg}\,}}(G'(n)) =2\) and \( {{\,\textrm{pd}\,}}(G'(n)) = n-2\).
-
(3)
\(\beta _{n-2, n} (G'(n)) = 1 - n + \frac{n}{2} \varphi (n)\) is the unique extremal Betti number, where \(\varphi \) is the Euler’s phi function.
Proof
(1): By Lemma 4.1, the components of the h-vector of \(\Delta (G'(n))\) are \(h_0 = \left( {\begin{array}{c}2\\ 0\end{array}}\right) f_{-1} = 1\), \(h_1=(-1)^ 1 \left( {\begin{array}{c}2-0\\ 1-0\end{array}}\right) f_{-1} + (-1)^ 0 \left( {\begin{array}{c}2-1\\ 1-1\end{array}}\right) f_0 = n-2,\) and
This implies that the Hilbert–Poincaré series of \(R/I(G'(n) )\) is
as required.
(2): By [2, Theorem B], the graph \(G'(n)\) is Cohen–Macaulay with \(\alpha (G'(n) ) = 2\), and also, by the Auslander–Buchsbaum formula, we have \({{\,\textrm{pd}\,}}(G'(n)) = n - \dim R/I(G'(n) ) = n - \alpha (G'(n)) = n-2.\) Moreover, since \(1-n+\frac{n}{2} \varphi (n) \ne 0\), by part (1), we obtain that \({{\,\textrm{pd}\,}}( G'(n))+{{\,\textrm{reg}\,}}(G'(n))=n\), and thus, \({{\,\textrm{reg}\,}}(G'(n) ) = 2\). (3): It is followed from (1) and (2). \(\square \)
Next, we consider the Betti numbers of the graph \(G'(p^{\alpha })\), p prime. We recall the structure of the graph \(G'(p^{\alpha })\) as follows:
Lemma 4.3
([2], Lemma 4.2) Let p be an odd prime and \(\alpha \) be a positive integer. Then the graph \(G'( p^{\alpha }) \) is a disjoint union of one complete graph \(K_{ p^{\alpha -1}}\) and \(\frac{p-1}{2}\) complete bipartite graphs \(K_{p^{\alpha -1}, p^{\alpha -1}}\), that is,
Lemma 4.4
Let p be an odd prime and \(\alpha \) be a positive integer. Then the following statements hold true:
-
(1)
\(\alpha (G'(p^{\alpha })) = \frac{(p-1)p^{\alpha -1}}{2} +1\).
-
(2)
The f-vector of \(\Delta (G'(p^{\alpha }))\) is \((1,p^{\alpha }, f_ {1}, \ldots , f_{\alpha (G'(p^{\alpha }))-1})\), where \(f_1=\frac{1}{2}(p^{\alpha }-p^{\alpha -1})(p^{\alpha }-1)\) and
$$\begin{aligned} f_i= & {} p^ {\alpha -1} \cdot \left[ \underset{ 0\le c_1,\ldots , c_{\frac{p-1}{2}}\le p^{\alpha -1}}{\underset{c_1+\cdots + c_{\frac{p-1}{2}}=i}{\sum }} \left\{ \prod _{j=1}^{\frac{p-1}{2}} \left( {\begin{array}{c}p^{\alpha -1}\\ c_j\end{array}}\right) \right\} 2^{|\{j\mid c_j>0\}|} \right] \\{} & {} + \underset{ 0\le c_1,\ldots , c_{\frac{p-1}{2}}\le p^{\alpha -1}}{\underset{c_1+\cdots + c_{\frac{p-1}{2}}=i}{\sum }} \left\{ \prod _{j=1}^{\frac{p-1}{2}} \left( {\begin{array}{c}p^{\alpha -1}\\ c_j\end{array}}\right) \right\} 2^{|\{j\mid c_j>0\}|} \end{aligned}$$\( \text { for all } 2\le i \le \frac{(p-1)p^{\alpha -1}}{2}.\)
Proof
By Lemma 4.3, the graph \(G'(p^{\alpha })\) is well-covered and \(\alpha (G'(p^{ \alpha })) = \frac{(p-1)p^ {\alpha -1}}{2} +1\). It is clear that \(f_{-1}=1, f_0=p^{\alpha }\). Applying the proof of Proposition 3.5, \(f_1= |E(G(p^{\alpha }))|=\frac{1}{2}(p^{\alpha }-p^{\alpha -1})(p^{\alpha }-1)\).
Based on the structure of the graph \(G'(p^{\alpha })\), we now assume that
Let \((U^{(j)}, V^{(j)})\) be the bipartition of \(K_{p^ {\alpha -1 }, p^{ \alpha -1}}^{(j)}\) for \(1\le j\le \frac{p-1}{2}\). Let \(I_i\) be the set of independent sets of \(K_{p^{\alpha -1}, p^{ \alpha -1}}^{(1)} \cup K_{p^{\alpha -1},p^{\alpha -1}}^{(2)} \cup \cdots \cup K_{p^ {\alpha -1 }, p^{ \alpha -1}}^{(\frac{p-1}{2})}\) with size i. We see that \(f_i= p^{\alpha -1}\cdot |I_i| + |I_j|.\) We observe that each element of \(I_i\) has the following form
where \(0\le |X^{(j)}|\le p^{\alpha -1}\) and \(|X^{(1)}| + \cdots + |X^{(\frac{p-1}{2})}|=i\) and either \(X^{(j)}\subset U^{(j)}\) or \(X^{(j)}\subset V^{(j)}\) if \(|X^{(j)}|>0\) for all \(1\le j \le \frac{p-1}{2}\). By virtue of this observation, we have
Therefore, we obtain
for every \(1 \le i \le \frac{(p-1)p^{\alpha -1}}{2}\), as required. \(\square \)
Proposition 4.5
Let p be an odd prime. Then \(\beta _{i,j}(G'(p)) = {\left\{ \begin{array}{ll} \left( {\begin{array}{c}\frac{p-1}{2}\\ i\end{array}}\right) &{}\text { if }\ j=2i, \\ 0 &{}\text { if }\ j\ne 2i. \end{array}\right. } \) In particular, \({{\,\textrm{pd}\,}}(G'(p)) = {{\,\textrm{reg}\,}}(G'(p))= \frac{p-1}{2}\).
Proof
Since the graph \(G'(p)\) is a disjoint union of \(\frac{p-1}{2}\) graphs \(K_2\) and one graph \(K_1\),
Now, by Lemma 2.1 and Proposition 2.4(1), we obtain that
Therefore, \({{\,\textrm{pd}\,}}(G'(p)) = {{\,\textrm{reg}\,}}(G'(p))= \frac{p-1}{2}\), as required. \(\square \)
Theorem 4.6
Let p be an odd prime and \(\alpha > 1\) be an integer. Let i and j be positive integers. If \(1\le j-i\le \frac{p+1}{2}\), then
In the case \(j-i>\frac{p+1}{2}\), we have \(\beta _{i,j} (G'(p^{\alpha }))=0\).
Proof
By Lemma 4.3, the graph \(G'(p^{\alpha })\) is a disjoint union of one complete graph \(K_{p^{\alpha -1}}\) and \(\frac{p-1}{2}\) complete bipartite graphs \(K_{p^{\alpha -1},p^{\alpha -1}}\). This implies that
Note that \(\beta _{0,0} (K_{p^ {\alpha -1}}) = \beta _{0,0} (K_{p^{\alpha -1}, p^{\alpha -1}})=1\). Also, by Proposition 2.4, the edge rings of \(K_{p^{\alpha -1}}\) and \(K_{p^{\alpha -1}, p^{ \alpha - 1}}\) have 2-linear resolutions. Now, \(\beta _ {u_0, u_0+1} (K_{ p^{ \alpha -1}}) = u_0 \left( {\begin{array}{c}p^{ \alpha -1}\\ u_0+1\end{array}}\right) \), together with Lemma 3.1(1), implies that
Therefore, by Lemma 2.1, we obtain that for every i and j,
where \(u_k, v_k\ge 0\) are integers.
First, suppose that \(j-i=1\). By Eq. 1, we obtain that
Second, suppose that \(j-i=\frac{p+1}{2}\). Therefore, there exists \(0\le k\le \frac{p-1}{2}\) such that \(u_k, v_k \ge 1\) and \(v_k \ne u_k+1\). Thus, \(\beta _ {u_k,v_k} (K_{p^{ \alpha -1}}) = 0\) and \(\beta _ {u_k, v_k} (K_{p^ {\alpha -1}, p^{ \alpha -1}} ) = 0\). Now, by Eq. 1, we obtain that
Third, suppose that \(1<j-i< \frac{p+1}{2}\). This implies that \(j-i=\sum _{k=0}^ {\frac{p-1}{2}} (v_k-u_k) < \frac{p+1}{2}.\) Therefore, there exists \(0\le k \le \frac{p-1}{2}\) such that \(v_k-u_k<1\). Note that if \(\beta _{i, j} (K_{p^ {\alpha - 1}})\ne 0\) then \(i=j=0\) or \(j=i+1\ge 2\). This is also true for \(\beta _{i, j} (K_{p^ {\alpha - 1}, p^ {\alpha - 1}})\ne 0\). Hence, if \(1<j-i< \frac{p+1}{2}\), it implies that
Hence, we obtain that
Since \(\beta _{0,0}(K_{p^{\alpha -1}})=1\) and \(\beta _{u_0,u_0+1}(K_{p^{\alpha -1}}) =u_0\left( {\begin{array}{c}p^{\alpha -1}\\ u_0+1\end{array}}\right) \), we obtain that
Forth, suppose that \(j-i>\frac{p+1}{2}\). Therefore, there exists \(0\le k \le \frac{p-1}{2}\) such that \(v_k>u_k+1\). This implies that \(\beta _{u_k,v_k}(H)= 0\), where H is either the complete graph \(K_{p^{\alpha -1}}\) or the complete bipartite graph \(K_{p^{\alpha -1}, p^{\alpha -1}}\). Now, by Eq. 1, we obtain that \(\beta _{i,j} (G'(p^{\alpha })) = 0\), and so, the proof is completed. \(\square \)
As an immediate corollary, we obtain that
Corollary 4.7
If p is an odd prime and \(\alpha > 1\) is an integer, then \({{\,\textrm{reg}\,}}(G'(p^{\alpha })) = \frac{p+1}{2} \text { and } {{\,\textrm{pd}\,}}(G'(p^{\alpha })) = p^{\alpha } -\frac{p+1}{2}.\)
We close this paper with the following example.
Example 4.8
Consider the graph \(G=G'(3^2)\). By Theorem 4.6, we obtain
Moreover, \(\beta _{i,i+2}(G) = \sum _{\begin{array}{c} i_0+i_1=i\\ i_0,i_1\ge 1 \end{array}} i_0\left( {\begin{array}{c}3\\ i_0+1\end{array}}\right) \left\{ \left( {\begin{array}{c}6\\ i_1+1\end{array}}\right) - 2\left( {\begin{array}{c}3\\ i_1+1\end{array}}\right) \right\} .\) Thus, we obtain the Betti table of G as follows:
Availability of Data, Code and Materials
Data sharing not applicable to this work as no data sets were generated or analysed during the current study.
References
Alon, N., Friedland, S., Kalai, G.: Regular subgraphs of almost regular graphs. J. Comb. Theory Ser. B 37(1), 79–91 (1984)
Ashitha, T., Asir, T., Hoang, D.T., Pournaki, M.R.: Cohen–Macaulayness of a class of graphs versus the class of their complements. Discrete Math. 344(10), 112525 (2021)
Baclawski, K., Garsia, A.M.: Combinatorial decompositions of a class of rings. Adv. Math. 39(2), 155–184 (1981)
Bruns, W., Conca, A., Römer, T.: Koszul homology and syzygies of Veronese subalgebras. Math. Ann. 351(4), 761–779 (2011)
Corso, A., Nagel, U.: Monomial and toric ideals associated to Ferrers graphs. Trans. Am. Math. Soc. 361(3), 1371–1395 (2009)
Dao, H., Huneke, C., Schweig, J.: Bounds on the regularity and projective dimension of ideals associated to graphs. J. Algebr. Comb. 38(1), 37–55 (2013)
Engström, A., Go, C., Stamps, M.T.: Betti numbers and anti-lecture hall compositions of random threshold graphs. Pac. J. Math. 319(1), 75–98 (2022)
Fröberg, R.: Betti numbers of fat forests and their Alexander dual. J. Algebr. Comb. 56(4), 1023–1030 (2022)
Grimaldi, R.P.: Graphs from rings. In: Proceedings of the Twentieth Southeastern Conference on Combinatorics, Graph Theory, and Computing (Boca Raton, FL, 1989). Congressus Numerantium, vol. 71, pp. 95–103 (1990)
Hà, H.T., Van Tuyl, A.: Resolutions of square-free monomial ideals via facet ideals: a survey. In: Algebra, Geometry and Their Interactions. Contemporary Mathematics, vol. 448. American Mathematical Society, Providence, pp. 91–117 (2007)
Herzog, J., Hibi, T.: Monomial Ideals. Graduate Texts in Mathematics, vol. 260. Springer, London (2011)
Hoang, D.T.: On the Betti numbers of edge ideals of skew Ferrers graphs. Int. J. Algebra Comput. 30(1), 125–139 (2020)
Hoang, D.T., Maimani, H.R., Mousivand, A., Pournaki, M.R.: Cohen–Macaulayness of two classes of circulant graphs. J. Algebr. Comb. 53(3), 805–827 (2021)
Jacques, S.: Betti numbers of graph ideals. PhD. thesis (2004). arXiv:math/0410107v1
Katzman, M.: Characteristic-independence of Betti numbers of graph ideals. J. Comb. Theory Ser. A 113(3), 435–454 (2006)
Kimura, K.: Non-vanishingness of Betti numbers of edge ideals. In: Harmony of Gröbner Bases and the Modern Industrial Society. World Scientific Publishing, Hackensack, pp. 153–168 (2012)
Martínez-Bernal, J., Pizá-Morales, O.A., Valencia-Bucio, M.A.: Nonvanishing Betti numbers of edge ideals of weakly chordal graphs. J. Algebr. Comb. 58(1), 279–290 (2023)
Miller, E., Sturmfels, B.: Combinatorial Commutative Algebra. Graduate Texts in Mathematics, vol. 227. Springer, New York (2005)
Peeva, I.: Graded Syzygies. Algebra and Applications, vol. 14. Springer, London (2011)
Rather, S.A., Singh, P.: On Betti numbers of edge ideals of crown graphs. Beitr. Algebra Geom. 60(1), 123–136 (2019)
Stanley, R.P.: Combinatorics and Commutative Algebra, vol. 41, 2nd edn. Birkhäuser Boston, Inc., Boston (1996)
Whieldon, G.: Jump sequences of edge ideals (2010). arXiv:1012.0108v1
Acknowledgements
A part of this work was completed during the third author’s visit to the Vietnam Institute for Advanced Study in Mathematics (VIASM). He wishes to express his gratitude toward VIASM for its support and hospitality. The research of D. T. Hoang was in part supported by a grant from Hanoi University of Science and Technology, Vietnam (Ref. T2023-PC-082). The research of T. Asir was in part supported by a grant from The Council of Scientific and Industrial Research, India (CSIR Project—Ref. 25/0323/23/EMR-II). The research of M. R. Pournaki was in part supported by a grant from The World Academy of Sciences, Italy (TWAS–UNESCO Associateship—Ref. 3240295905).
Author information
Authors and Affiliations
Contributions
All authors have contributed equally to this work.
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no Conflict of interest
Ethical Approval and Consent to Participate
Not applicable.
Consent for Publication
Not applicable.
Additional information
Communicated by Siamak Yassemi.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Ashitha, T., Asir, T., Hoang, D.T. et al. Betti Numbers of Edge Ideals of Grimaldi Graphs and Their Complements. Bull. Malays. Math. Sci. Soc. 47, 136 (2024). https://doi.org/10.1007/s40840-024-01731-2
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s40840-024-01731-2