Korean J. Math. 27 (2019), No. 3, pp.

563–580
https://doi.org/10.11568/kjm.2019.27.3.563

ON STRONG METRIC DIMENSION OF ZERO-DIVISOR

GRAPHS OF RINGS

M. Imran Bhat and Shariefuddin Pirzada∗

Abstract. In this paper, we study the strong metric dimension

of zero-divisor graph Γ(R) associated to a ring R. This is done by
transforming the problem into a more well-known problem of finding
the vertex cover number α(G) of a strong resolving graph Gsr . We
find the strong metric dimension of zero-divisor graphs of the ring
Zn of integers modulo n and the ring of Gaussian integers Zn [i]
modulo n. We obtain the bounds for strong metric dimension of
zero-divisor graphs and we also discuss the strong metric dimension
of the Cartesian product of graphs.

1. Introduction

Let G(V, E) be a simple graph with vertex set V (G) and edge set
E(G). The set of vertices adjacent to a vertex v ∈ V (G) is the neigh-
borhood of v and is denoted by N (v). Further N [v] = N (v) ∪ {v}. The
degree of v, denoted by dG (v), or more simply we write d(v) means the
cardinality of N (v). If the two vertices u and v are adjacent, we denote
it by u adj v. A graph is regular if each of its vertex has the same de-
gree. A path between two vertices x1 , xn ∈ V (G) is an ordered sequence
of distinct vertices x1 , x2 , . . . , xn of G such that xi−1 xi is an edge for
Received October 1, 2018. Revised July 9, 2019. Accepted August 7, 2019.
2010 Mathematics Subject Classification: 13A99, 05C78, 05C12.
Key words and phrases: Metric dimension, zero-divisor graph, strong metric
dimension.
This research is supported by the University Grants Commission, New Delhi with
research project number MRP-MAJOR-MATH-2013-8034.
∗ Corresponding author.
c The Kangwon-Kyungki Mathematical Society, 2019.
This is an Open Access article distributed under the terms of the Creative com-
mons Attribution Non-Commercial License (http://creativecommons.org/licenses/by
-nc/3.0/) which permits unrestricted non-commercial use, distribution and reproduc-
tion in any medium, provided the original work is properly cited.
564 M. Imran Bhat and S. Pirzada

2 ≤ i ≤ n. A closed path is a cycle. In G, the distance between two

vertices x and y, denoted by d(x, y), is the length of the shortest x − y
path in G. If there is no such path, we define d(x, y) = ∞. We say
that G is connected if there exists a path between every pair of vertices
in G. A graph that contains no cycles is called a tree. A cut vertex
of a connected graph is a vertex whose removal results in a graph hav-
ing two or more connected components. The diameter of a graph G is
diam(G) = sup{d(x, y) | x and y are distinct vertices of G}. A clique
is a maximal complete subgraph and the cardinality of its vertex set,
denoted by ω(G), is called the clique number of G. In a graph G, a set
S ⊂ V (G) is an independent set if the subgraph induced by S is totally
disconnected. We denote the complete graph on n vertices by Kn and
the complete bipartite graph on m and n vertices by Km,n . We will
sometimes call a K1,t a star graph. A vertex u of G is maximally distant
from v if for every vertex w ∈ N (u), d(u, v) ≥ d(v, w). If v is also maxi-
mally distant from u, then we say that u and v are mutually maximally
distant and denote this by uM M Dv. Also boundary of G(V, E) is de-
fined as ∂(G) = {u ∈ V : there exists v ∈ V with uMMDv}. A set T of
vertices of G is a vertex cover of G if every edge of G is incident with at
least one vertex of T . The vertex cover number of G, denoted by α(G),
is the cardinality of smallest vertex cover of G. For basic definitions, we
refer the reader to any standard graph theory book, such as [17, 26].

The idea of associating a graph to a ring is due to Beck [11], in

which the author is primarily concerned with colorings. In [5], Anderson
and Livingston defined the zero-divisor graph of a commutative ring
R, denoted Γ(R), to be the graph whose vertices are the nonzero zero-
divisors of R, and in which x and y are connected by an edge if xy = 0.
Since then, there have been many papers written on the subject of zero-
divisor graphs and and their variants (of which there are many). The
interrelation between the ring-theoretic structure of R and the graph-
theoretic structure of Γ(R) has brought out interesting results from the
perspective of both algebra and graph theory (cf. [2, 4–6], for example).
Zero-divisor graphs were initially defined for commutative rings and later
the concept of zero-divisor graphs was generalized to non-commutative
rings by Redmond [21] and to the modules (see for example [10]). The
concept widened the scope of this research area and many other graphs
have been defined like total graphs, co-maximal graphs, unit graphs,
Jacobson graphs, ideal based zero-divisor graphs, zero-divisor graphs of
 On strong metric dimension of zero-divisor graphs of rings 565

equivalence classes (cf. [1, 3, 7, 9, 22, 25]). For basic definitions, we refer
the reader to [8, 15].
Throughout, unless otherwise stated, R denotes a finite commutative
ring with 1 6= 0, the set of all non-zero zero-divisors of R is denoted by
Z ∗ (R) = Z(R) r {0}. A finite field on q number of elements is denoted
by Fq and the ring of integers modulo n is denoted by Zn . A ring R is
a local ring if and only if R has a unique maximal ideal. An element
x ∈ R is nilpotent if xn = 0 for some n ∈ N. A ring R is a reduced ring if
it contains no non-zero nilpotent element. An annihilator of an element
x of a ring R is defined as ann(x) = {r ∈ R | rx = 0}.

2. Metric dimension of some graphs

Harary and Melter [14] introduced the concept of metric dimension

of graphs in the following way. Let G be a connected graph of order
n ≥ 1 and let W = {w1 , w2 , . . . , wk } be an (ordered) set of vertices. The
metric vector of a vertex v ∈ G relative to W is the vector r(v|W ) =
(d(v, w1 ), d(v, w2 ), . . . , d(v, wk )). The set W is a resolving set of G if
distinct vertices have distinct metric vectors and a minimum resolving
set is called a metric basis for G and its cardinality, denoted by dim(G),
is called the metric dimension of G. This invariant has been further
studied by a number of authors, including [12, 13, 18–20].
The strong metric dimension of a graph is defined as follows. In a
connected graph G, for two distinct vertices u and v, the interval I[u, v]
is the collection of all vertices that belong to some shortest u−v path. A
vertex w ∈ V (G) strongly resolves two vertices u and v if v ∈ I[u, w] or
u ∈ I[v, w]. In other words, two vertices u and v are strongly resolved by
w if d(w, u) = d(w, v) + d(v, u) or d(w, v) = d(w, u) + d(u, v). A set W of
vertices is a strong resolving set of G if every two distinct vertices of G are
strongly resolved by some vertex of W and a minimum strong resolving
set is called a strong metric basis and its cardinality is the strong metric
dimension of G, denoted by dims (G). If a vertex w strongly resolves u
and v, it is easy to see that w also resolves these vertices. Hence every
strong resolving set is a resolving set and dim(G) ≤ dims (G). In fact,
1 ≤ dims (G) ≤ n − 1. Oellermann and Peters-Fransen [16] showed that
the problem of finding the strong metric dimension of a graph G can be
transformed into a more well-known problem of finding the vertex cover
number α(Gsr ) of a strong resolving graph denoted by Gsr with vertex
566 M. Imran Bhat and S. Pirzada

set V (Gsr ) = ∂(G) and uv ∈ E(Gsr ) if and only if uMMDv in G. We

notice that every vertex of a strong resolving set is a boundary vertex.
Example 2.1. For positive integers m and n,
(i) (Kn )sr = Kn
(ii) (Km,n )sr = Km,n
Theorem 2.2. [16] For any connected graph G, dims (G) = α(Gsr ).
Now, we discuss the strong metric dimension of some useful graphs.
We start with the following lemma.
Lemma 2.3. For a connected graph G of order n ≥ 1, dims (G) = 1 if
and only if G ∼
= Pn , where Pn is the path on n vertices. Moreover, the
only end vertices belong to the strong resolving set.
Proof. Let Pn := u = v1 − v2 − · · · − vn = v be a path. To show
dims (G) = 1, by Theorem 2.2, it is enough to prove that V ((Pn )sr ) =
∂(Pn ) = {u, v}, that is, u and v are the only MMD vertices. First we
show u and v are MMD. Clearly, d(u, v) ≥ d(v, w) for all w ∈ N (u),
implies u is maximally distant from v. Also, d(v, u) ≥ d(u, w) for all w ∈
N (v), implies v is maximally distant from u. Therefore, by definition
uMMDv. Now, we show that there is no any other pair of vertices which
are M M D. Let vi , vj ∈ V (G) for 1 < i, j ≤ n − 1. We consider the
following three cases.
Case 1. If vi and u are adjacent, then d(u, vi ) > d(vi , w) for all w ∈ N (u).
But d(vi , u)  d(u, w) for all w ∈ N (vi ), therefore vi is not M M D to u.
Case 2. If vi and u are not adjacent, then d(vi , u) ≥ or ≤ d(u, w) for
all w ∈ N (vi ) and d(u, vi ) ≥ d(vi , w) for all w ∈ N (u), therefore vi is not
M M D to u.
Case 3. Now, consider the vertices vi and vj , we observe that d(vi , vj ) 
d(vj , w), for all w ∈ N (vi ) and d(vi , vj )  d(vi , w), for all w ∈ N (vj ),
which implies that vi is not M M D vj .
Thus u and v are the only MMD vertices in Pn . Hence, (Pn )sr ∼ = K2 ,
implies that dims (Pn ) = α(Pn )sr = 1.
On the other hand, let G be not a path, then either G is a tree (except
path) or contains a cycle. Since in either case dim(G) ≥ 2 and hence
dims (G) ≥ 2, as paths are the only graphs whose dimension is 1, a
contradiction.
Further, any vertex vi , 2 ≤ i ≤ n − 1 does not strongly resolve the
end vertices u = v1 and v = vn of Pn . Therefore, only the end vertex
forms a strong metric basis.
 On strong metric dimension of zero-divisor graphs of rings 567

The converse part also follows from the fact that 1 ≤ dim(G) ≤
dims (G), implying dim(G) = 1. Therefore, by [ [19], Lemma 2.1], G ∼
=
Pn .
Theorem 2.4. A connected graph G of order n ≥ 2 has strong metric
dimension n − 1 if and only if G ∼ = Kn .
Proof. First, assume that G ∼ = Kn . Since dim(G) ≤ dims (G) and
dim(G) = n − 1, it follows that dims (G) ≥ n − 1. Also, by definition,
dims (G) ≤ n − 1. Combining, we have dims (G) = n − 1.
0
For the converse, assume that dims (G) = n − 1. Let G = Kn − e,
where e = uv is an edge and let uui v be a path of length 2 in G − e. For
the strong resolving set of G − e, we consider the following three cases.
(i) W1 = V (G) r {u, v} (ii) W2 = V (G) r {ui , uj }, (|V (G)| ≥ 4) (iii)
W3 = V (G) r {ui , u} or V (G) r {ui , v}
Clearly, W1 is not a strong resolving set. If so, then u ∈ I[ui , v] or
v ∈ I[u, ui ], for any ui ∈ W1 which is not true as ui adj v. Also W2 is
not a strong resolving set, because neither ui ∈ / I[u, uj ] nor uj ∈
/ I[ui , u],
because u adj ui , uj . So, W3 is a strong resolving set, where ui and u are
strongly resolved by v; and ui and v are strongly resolved by u. Thus,
dims (G − e) ≤ n − 2. Therefore, G ∼ = Kn .
By using the fact (Kn )SR = Kn and Theorem 2.2, we note that
dims (G) = n − 1 if G ∼
= Kn .

Proposition 2.5. For a graph G, dim(G) = dims (G) if

(i) G ∼
= Pn .
∼
(ii) G = Kn .
Theorem 2.6. For any complete bipartite graph Km,n , dims (Km,n ) =
m − n − 2.
Proof. Consider a complete bipartite graph G = Km,n with partite
sets V1 = {u1 , u2 , . . . , um } and V2 = {v1 , v2 , . . . , vn }. Let W be a strong
resolving set of G. Then W = W1 ∪ W2 , Wi ⊆ Vi , (i = 1, 2) with
|W1 | = m − 1 and |W2 | = n − 1. We claim that W = V (G) r {um , vn }.
For if, W ∗ = V (G) r {um , vn , uk } or W ∗ = V (G) r {um , vn , vk }, then
both do not form a strong metric basis of G. If W ∗ = V (G)r{um , vn , uk }
forms a strong metric basis, then a pair of vertices um and uk are not
strongly resolved by any vertex of W ∗ . The same argument applies to
the other case. Hence, W is a strong metric basis of G.
568 M. Imran Bhat and S. Pirzada

Definition 2.7. Two distinct vertices u and v of a connected graph

G with |V (G)| ≥ 2 are distance similar if d(u, x) = d(v, x), for all x ∈
V (G)r{u, v}. It can be easily seen that two distinct vertices are distance
similar if either uv ∈
/ E(G) and N (u) = N (v) or uv ∈ E(G) and N [u] =
N [v].
Theorem 2.8. Let G be a connected graph whose vertex set is parti-
tioned into k distinct distance similar classes V1 , V2 , . . . , Vk and m is the
number of distance similar equivalence classes that consist of a single
vertex. Then |V (G)| r k ≤ dims (G) ≤ |V (G)| − k + m.
Proof. Let V (G) be partitioned into k distinct distance similar classes
V1 , V2 , . . . , Vk . Clearly, each Vi , 1 ≤ i ≤ k, is either an independent set
or induces a complete subgraph of G. If W is a strong resolving set of
G, then W contains all except one vertex in each of the equivalence class
Vi , otherwise there exists a pair of vertices u, v (u ∼ v) not resolved by
any vertex w ∈ W . Thus, dims (G) ≥ |V (G)| − k.
If W is a minimal strong resolving set for G, we prove that W contains
at most |Vi | − 1 vertices of Vi , |Vi | > 1. Without loss of generality,
suppose that |V1 | > 1 and W be a strong resolving set for G such that
0
V1 ⊂ W . Let x ∈ V1 . We show that either (a) W = W − {x} is a strong
resolving set for G or (b) there exists an element t ∈ V (G) − V1 such
0
that W ∗ = W ∪ {t} = W ∪ {t} − {x} is a strong resolving set for G.
That is, W ∗ is a strong resolving set of cardinality no larger than W ,
where V1 ( W .
0
Define W = W − {x} and without loss of generality, choose W =
0
{x, w1 , w2 , . . . } and W = {w1 , w2 , . . . }. Let u, v ∈ V (G). If both u, v ∈
0 0
W , then clearly u and v are strongly resolved by a vertex of W , that is,
0 0
for any w ∈ W either u ∈ I[w, v] or v ∈ I[u, w]. Again, if u or v ∈ W ,
then by definition, there exists some shortest w − u path containing v
or some shortest w − v path containing u.
Suppose u, v ∈ / W . Then u and v are strongly resolved by a vertex of
0
W . If u and v are not strongly resolved by a vertex of W , it must be
the case that u and v are strongly resolved by x. However, there exist
0
some z ∈ W ∩ V1 such that u and v are not strongly resolved, would
imply u and v are not strongly resolved by z. Since x, z ∈ V1 , imply u
0
and v are strongly resolved by a vertex of W .
0
If there does not exist any u ∈ W such that u and x are not strongly
0 0
resolved by a vertex of W , then W is a strong resolving set of G.
0
So, assume there does not exist any u ∈ W such that u and x are
 On strong metric dimension of zero-divisor graphs of rings 569
0 0
not strongly resolved by a vertex of W . Let r ∈ W ∩ V1 and let there
0
be some other element v ∈ W such that v and x and u and x are not
0
strongly resolved by a vertex of W . Thus, v ∈ I[u, x] or u ∈ I[v, x].
Since W is a strong resolving set, u, v are strongly resolved by a vertex
0
of W . However, u and v not being strongly resolved by a vertex of W
and x, r ∈ V1 imply x, r are not strongly resolved by u and v. Also,u, v
are not strongly resolved by r, a contradiction. Hence, if there exists an
0
element v ∈ W such that v and x are not strongly resolved by a vertex
0
of W , then v is unique.
Case 1. Suppose |Vi | > 1 for each i. For any y ∈ V (G) − {x, u}, choose
0
q ∈ W such that y ∼ q. Then d(y, u) = d(q, u) = d(q, x) = d(y, x).
Thus, u ∼ x, which is a contradiction.
Case 2. Suppose |Vi | = 1 for some i. Since vi and x are not distance
similar, there is some s ∈ V (G) − {x, u} with d(s, x) 6= d(u, x). Note
0
that Vj = {s} for some j, s ∈ Vj , because if not, there is some t ∈ Vj ∩W
∗ 0
and d(u, s) = d(u, t) = d(x, t) = d(x, s). Define, W = W ∪ {s}. Then
0 0
|W ∗ | = |W | and u and x are strongly resolved by a vertex of W ∗ . Since
0 0
W ⊂ W ∗ and using the same argument as above, a and b are strongly
0
resolved by a vertex of W for any two distinct vertices a, b ∈ V (G).
0
Hence, W ∗ is a strong resolving set for G. Combining these facts we
have dims (G) ≤ |V (G)| − k + m.

3. Strong metric dimension of zero divisor graphs of rings

We start this section with the following observation.

Theorem 3.1. Let R be a finite commutative ring. Then
(i) dims (Γ(R)) is finite if and only if R is finite.
(ii) dims (Γ(R)) is undefined if and only if R is an integral domain.
Proof. (i) If R is finite, then |Z ∗ (R)| is finite and therefore dims (Γ(R))
is finite. Now assume that dims (Γ(R)) is finite. Let W be a minimal
strong metric basis for Γ(R) with |W | = k, where k is some positive
integer. Then dim(Γ(R)) ≤ dims (Γ(R)) = k implies that dim(Γ(R)) ≤
k. Now, since the diameter of Γ(R) is not more than 3, so by [ [5],
Theorem 2.3] Γ(R) is finite. Therefore, d(x, y) ∈ {0, 1, 2, 3}, for every
x, y ∈ Z ∗ (R). Hence, |Z ∗ (R)| ≤ 4k . This implies that Z ∗ (R) is finite
and hence R is finite.
570 M. Imran Bhat and S. Pirzada

(ii) This follows from the fact that the strong metric basis of Γ(R) is
undefined if and only if the vertex set of Γ(R) is empty.
Theorem 3.2. Let R be a commutative ring with unity. Then dims (Γ(R)) =
1 if and only if R is isomorphic to one of the following rings.
(i) Z6 , Z8 , Z2 × Z2
(ii) Z(x3 [x] Z2 [x] Z4 [x]
2 ) , (x3 ) or (2x,x2 −2)

Proof. Observe that the zero-divisor relation is not transitive for these
rings, implies their Γ(R) is a path P2 or P3 . Therefore, by Lemma 2.3,
dims (G) = 1. On the other hand, since paths are the only graphs for
which the strong metric dimension is 1, so |Z ∗ (R)| ≤ 3. The only if
direction follows.

Figure 1. dims (Γ(Z3 × Z3 )) = 2

Proposition 3.3. Let R be a commutative ring with 1 6= 0. Then

dims (Γ(R)) = 2 if
(i) R ∼
= Z3 × Z3
(ii) R ∼ F [x] Z4 [x]
= (x4 2 ) , (x2 +x+1)
Z4 [x] Z2 [x,y]
, (2,x)2 , (x,y)2 .

Proof. (i) If R ∼ = Z3 × Z3 , then Γ(R) is a cycle on four vertices as

shown in Figure 1. The set W = {a, b} is in fact a strong resolving
set of Γ(R). Since all the possible sets I[u, v], where u ∈ Γ(R) and
v ∈ W have the form I[a, b] = {a, b}, I[a, c] = {a, c}, I[a, d] = {a, b, c, d},
I[b, c] = {a, b, c, d}, therefore each pair of vertices which contain vertex
a or vertex b is strongly resolved by a or b. Vertices c and d are strongly
resolved by both a and b, since c ∈ [a, d] and d ∈ [b, c]. Hence, 2 =
dim(Γ(Z3 × Z3 )) ≤ dims (Γ(Z3 × Z3 )) ≤ 2. Thus, dims (Γ(Z3 × Z3 )) = 2.
(ii) We know that Γ(R) ∼ = K3 only if R is isomorphic to the rings men-
tioned above. Now, by Example 2.1, (K3 )SR = K3 and it is easy to see
that α(K3 ) = 2. Hence, dims (K3 ) = 2, by Theorem 2.2.
 On strong metric dimension of zero-divisor graphs of rings 571

For any zero-divisor graph Γ(R) of a commutative ring R with vertex

set V (Γ(R)) containing at least four vertices, dim(Γ(R)) = |Γ(R)| − 2
implies that dims (Γ(R)) = |Γ(R)| − 2. However, the converse is not
true. For example, consider the ring R ∼ Z [x]
= (x4 2 ) . Then its Γ(R) is shown
in Figure 2(b) with 3 = dim(Γ(R)) ≤ dims (Γ(R)) = 5.

Figure 2. 3 = dim(Γ( Z(x4 [x] Z4 [x]

2 ) )) ≤ dims (Γ( (x2 ) )) = 5

Theorem 3.4. Let R be a finite commutative ring with unity 1 such

that |Z ∗ (R)| ≥ 3. If Γ(R) has a cut vertex but no degree 1 vertex, then
dims (Γ(R)) = 5
Proof. By [ [23], Theorem 3], if Γ(R) has a cut vertex but no degree
one vertex, then R is isomorphic to one of the following rings
Z4 [x, y]/(x2 , y 2 , xy−2, 2x, 2y), Z2 [x, y]/(x2 , y 2 ), Z4 [x]/(x2 ), Z4 [x]/(x2 +2x),
Z8 [x]/(2x, x2 +4), Z2 [x, y]/(x2 , y 2 −xy), Z4 [x]/(x2 , y 2 −xy, xy −2, 2x, 2y).
The zero-divisor graphs associated to these rings with dims (Γ(R)) = 5
are shown in Figure 2.
Theorem 3.5. Let R be a commutative ring with unity. Then (Γ(R))sr ∼ =
∼ Z3 [x] Z2 [x] Z4 [x]
K1,t if and only if R = Z6 , Z8 , Z9 , Z2 ×Z2 , (x2 ) , (x3 ) or (2x,x2 −2) and t = 1.
Proof. Let R be isomorphic to one of these rings. Then Γ(R) is a path
with at most three vertices. Since the only end vertices of a path are
M M D from each other, see Lemma 2.3, it follows that |V (Γ(R))sr | =
|∂(Γ(R))| = 2. Thus, (Γ(R))sr ∼
= K2 . On the other hand, let (Γ(R))sr ∼=
K1,t . Then, by Lemma 2.3, dims (Γ(R)) = 1 implies Γ(R) is a path.
Hence R is isomorphic to one of the rings mentioned above.
The above discussions lead to the following problem.
Problem 3.6. Do there exist rings R whose strong resolving graph
Γ(R))sr is isomorphic to P3 .
572 M. Imran Bhat and S. Pirzada

Theorem 3.7. If R is a finite commutative ring with unity and R ∼ =

Z2 × F for some finite field F, then dims (Γ(R)) = |Γ(R)| − 2. Moreover,
if R is a local ring such that Γ(R) has no cycles, then dims (Γ(R)) = 1.

Proof. Firstly, if R is a local ring, the only zero divisor graphs with
no cycles have three or fewer vertices [ [24], Theorem 2.1]. Hence,
dims (Γ(R)) = 1 in this case. Now, if R is a non local ring and R ∼ = Z2 ×F,
then its zero-divisor graph has a vertex adjacent to all other vertices, that
is, Γ(R) is a star graph K1,|Z ∗ (R)|−1 of order |Z ∗ (R)|. Let u be the center
vertex adjacent to the set of all other |Z ∗ (R)|−1 vertices vi , (1 ≤ i ≤ n),
n = |Z ∗ (R)| − 1 which is an independent set. Clearly, the path between
the two vertices vi and vj is not contained in any other shortest path
and therefore every strong resolving set must contain at least one of
them. In other words, each vi is mutually maximally distant with vj ,
i 6= j, (1 ≤ i, j ≤ n), as d(vi , vj ) ≥ d(vj , u) for every u ∈ N (vi ) and
d(vi , vj ) ≥ d(u, vi ) for every u ∈ N (vj ). Therefore, any strong resolving
set of K1,|Z ∗ (R)|−1 must contain either vi or vj , i 6= j.
We claim that W = {v1 , v2 , . . . , vn−1 } is a strong resolving set. For,
0
if W = {v1 , v2 , . . . , vn−2 } is a strong resolving set, then by definition
0
each pair of vertices is resolved by any vertex of W . Choose vn−1 and
0
vn . Then vn−1 ∈ I[vi , vn ] or vn ∈ I[vi , vn−1 ] for any vi ∈ W which
is not true. Thus, W is a strong resolving set. Hence, dims (Γ(R)) =
dims (K1,|Z ∗ (R)|−1 ) = |Z ∗ (R)| − 2 = |Γ(R)| − 2.

Corollary 3.8. If R is a reduced ring and Γ(R) has a vertex adjacent

to every other vertex, then either Γ(R) ∼
= K2 or dims (Γ(R)) = |Γ(R)|−2.
Theorem 3.9. Let R be a ring and let Γ(R) be a regular graph. Then
dims (Γ(R)) = |Z ∗ (R)| − 1 if and only if either R ∼
= F × A, where A is an
integral domain, or Z(R) is an annihilator ideal (and hence is prime).

Proof. Suppose that R ∼ = F × A, where A is an integral domain.

Then, for 0 6= a, vertex (a, 0) is adjacent to every other vertex. But
Γ(R) is regular graph, therefore Γ(R) is complete regular and hence
dims (Γ(R)) = |Z ∗ (R)| − 1. Conversely, assume that dims (Γ(R)) =
|Z ∗ (R)| − 1. Since Γ(R) is regular, so Γ(R) is a complete graph. Thus
there exists a vertex adjacent to every other vertex. Now, let Z(R) be
not an annihilator ideal (and hence is prime). Then, by [ [5], Theorem
2.5], the result follows.
 On strong metric dimension of zero-divisor graphs of rings 573

Proposition 3.10. If R is a finite commutative ring with unity 1 such

that R = F1 × F2 , where F1 and F2 are finite fields with |F1 | = m ≥ 3
and |F2 | = n ≥ 3, then dims (Γ(R)) = |F1 | + |F2 | − 2ω(Γ(R)).
Proof. If R = F1 × F2 , then the vertex set of Γ(R) can be parti-
tioned into two distinct vertex sets V1 = {(u, 0) : u ∈ F∗1 } and
V2 = {(0, v) : v ∈ F∗2 }, where each (u, 0) is adjacent to every ver-
tex (0, v). Thus, Γ(R) is a complete bipartite graph Km−1,n−1 . Since
ω(Γ(R)) = 2, by Theorem 2.6, dims (Γ(R)) = |F1 | + |F2 | − 2ω(Γ(R)).
Proposition 3.11. If R is a finite local ring with maximal ideal m
and m2 = {0}, then dims (Γ(R)) = |Γ(R)| − 1.
Proof. Recall that the Jacobian radical J(R) of R is the intersection
of maximal ideals of R. Since R is finite local ring, so J(R) = Z(R) and
Z(R) = m. Thus Z(R) is a nilpotent ideal and R is not a field, implies
ann(Z(R)) 6= {0}. As m2 = {0}, so ann(Z(R)) = Z ∗ (R) and therefore
Γ(R) is complete and thus the result follows by Theorem 2.4.
Theorem 3.12. Let R be a reduced ring and I1 and I2 be two distinct
prime ideals such that I1 ∩ I2 = {0}. Then dims (Γ(R)) = |I1 | + |I2 | − 4.
Proof. Let x ∈ Z(R) r I1 ∪ I2 . Then there exists 0 6= b ∈ R such that
ab = 0 ∈ I1 ∩ I2 . So, y ∈ I1 ∩ I2 , a contradiction, because I1 ∩ I2 = {0}.
Also, I1 ∩I2 ⊆ Z(R). So, Z(R) = I1 ∪I2 . Now, take V1 = |I1 |−{0} and
V2 = |I2 | − {0}. We claim that Γ(R) is a complete bipartite graph with
partite sets V1 and V2 . Indeed, if a, b ∈ V1 with ab = 0, then ab ∈ I2 and
therefore a or b ∈ V2 , a contradiction. Thus Γ(R) is a bipartite graph.
Now, we take a ∈ V1 and b ∈ V2 . So ab ∈ I1 and ab ∈ I2 , since I1
is an ideal and I2 is an ideal. Then ab ∈ I1 ∩ I2 = {0} implies that
ab = 0. Thus, Γ(R) is a complete bipartite graph. Hence, by Theorem
2.6, dims (Γ(R)) = |I1 | + |I2 | − 4.
Theorem 3.13. Let R1 and R2 be commutative rings with R1 ∼ = Zp2
and R2 ∼ Z p [x]
= 2 , where p is a prime. Then dims (Γ(R1 )) = dims (Γ(R2 )) =
(x )
p − 2.
Proof. Considering the ring R1 = Zp2 , its set of non-zero zero-divisors
is Z ∗ (R1 ) = {kp : 1 ≤ k ≤ p − 1, k ∈ N} such that k1 pk2 p = 0 for all
1 ≤ k1 k2 ≤ p − 1. Thus, Γ(R) ∼= Kp−1 .
Zp [x]
Now, consider R2 = (x2 ) = {a + bx : a, b ∈ Zp }. So, Z ∗ (R2 ) = {bx :
1 ≤ b ≤ p − 1}. We see that Γ(R2 ) ∼= Kp−1 . Therefore, by Theorem 2.4,
dims (Γ(R1 )) = dims (Γ(R2 )) = p − 2.
574 M. Imran Bhat and S. Pirzada

From the above theorem we have the following consequence.

Corollary 3.14. The graph Γ(Zn ) is Hamiltonian if and only if
dims (Γ(Zn )) = |Z ∗ (Zn )| − 1.
Proof. By Corollary 1 of [2], we know that the graph Γ(Zn ) is Hamil-
tonian graph if and only if n = p2 , where p is a prime larger than 3 and
Γ(Zn ) is isomorphic to Kp−1 . Thus the result follows.
Proposition 3.15. dims (Γ(Zn )) = |Z ∗ (Zn )| − 2, if n = 2p, where p
is prime larger than 2.
Proof. If p > 2, the zero-divisor set of Zn is Z ∗ (Zn ) = {2k, 1 ≤ k ≤
p; k ∈ N} such that 2k1 2k2 = 0. It follows that p is adjacent to all
other vertices. Thus, Γ(Zn ) ∼ = K1,|Z ∗ (Zn )|−1 . Therefore, by Theorem 2.6,
dims (Γ(Zn )) = |Z ∗ (Zn )| − 2.
Theorem 3.16. Let p be a prime number and n ∈ N. Then dims (Γ(Zn )) =
∗
|Z (Zn )| − 2 if
(i) n = pq, where p and q are distinct primes.
(ii) n = 22 p, where p is any odd prime.
Proof. (i) If n = pq, we partition the zero-divisor set of Zn into sets
V1 = {kp : (k, q) = 1} and V2 = {kq : (k, p) = 1}. Clearly, Γ(Zn ) is
a bipartite graph. Also, uv = 0 for every u ∈ V1 and v ∈ V2 . Hence,
Γ(Zn ) is a complete bipartite graph. Therefore, by Theorem 2.6, the
result follows.
(ii) If n = 22 p, where p is any odd prime, we partition the vertex set
into sets V1 = {2k, 1 ≤ k ≤ n, k 6= p} and V2 = {kp : kp < n} =
{p, 2p, 3p}. Since, p - 2k for any 1 ≤ k ≤ n, none of the elements of V1
are adjacent. Also, since 2 - p and 2 - 3p, no elements of V2 are adjacent.
Furthermore, we see that uv = 0 for every u ∈ V1 and v ∈ V2 . Hence,
Γ(Zn ) is a complete bipartite graph. Therefore, by Theorem 2.6, the
result follows.
Remark 3.17. To construct the zero-divisor graph of Zn and hence
to find strong metric dimension of Γ(Zn ), it is best to break down n into
prime factorization. Here we discuss some cases, when 1 < n < 100.
Case 1. If n is a single prime, the graph Γ(Zn ) is trivial with no vertices
or edges.
Case 2. If n = pq. The numbers in this case are 6, 10, 14, 15, 21, 22,
26, 33, 34, 35, 38, 39, 46, 51, 55, 57, 58, 62, 65, 69, 74, 77, 82, 85, 86, 87,
 On strong metric dimension of zero-divisor graphs of rings 575

91, 93, 94, 95. The zero-divisor graph is the complete bipartite graph by
taking all the multiples of p in one partite set and the remaining zero-
divisors as the multiples of q in another partite set. Thus, by Theorem
2.6, dims (Γ(Zn )) = number of multiples of p + number of multiples of
q - 2. Clearly this case is discussed in Theorem 3.16.
Case 3. If n = p2 , the numbers in this case are 9, 25 and 49. This case
is also discussed in Theorem 3.13.
Case 4. If n = p3 , the zero-divisor graph is a complete bipartite graph
by taking the vertices which are multiples of p2 in one class and the
remaining vertices being all multiples of p in the other. Thus this case
also follows from Theorem 2.6.
Definition 3.18. The set√of Gaussian integers is denoted by Z[i] =
{a + ib | a, b ∈ Z and i = − 1}. Clearly Z[i] is a ring under the
usual complex operations. The factor ring Z[i]/hni is isomorphic to
Zn [i] = {a + ib | a, b ∈ Zn }, where hni is a principal ideal generated by n
for some positive integer larger than 1 in Z[i]. Obviously, Zn [i] is a ring
with addition and multiplication modulo n. This ring is called the ring
of Gaussian integers modulo n.
We now determine the strong metric dimension of Γ(Zn [i]).
Theorem 3.19. (i) dims (Γ(Zn [i])) = 0, if n = 2.
(ii) dims (Γ(Zn [i])) is undefined, if n = q ≡ 3 modulo 4.
(iii) dims (Γ(Zn [i])) = q 2 − 2, if n = q 2 .
(iv) dims (Γ(Zn [i])) = 2p − 4, if n = p ≡ 1 modulo 4.
(v) dims (Γ(Zq1 q2 [i])) = q1 2 + q2 2 − 4, if qj ≡ 3 modulo 4, j = 1, 2.
Proof. (i). Z2 [i] is isomorphic to the local ring Z[i]/h(1 + i)2 i, with
unique maximal ideal {0, 1+i}. So we have V (Γ(Z2 [i])) = {1+i}, which
implies that Γ(Z2 [i]) is a graph on a single vertex and no edge and the
result holds.
(ii). In this case Zq [i] is a field, therefore Γ(Zq [i]) is an empty graph. So
dims (Γ(Zq [i])) is undefined.
(iii). Γ(Zq2 [i]) is a complete graph isomorphic to Kq2 −1 implies that
dims (Γ(Zq2 [i])) = q 2 − 2.
(iv). Γ(Zp [i]) is a complete bipartite graph Kp−1,p−1 with partite sets
V1 =< a + ib > −{0} and V2 =< a − ib > −{0}, since Zp [i] ∼ = Z[i] ∼
=
Z[i]/ha + ibi × Z[i]/ha − ibi.
(v). Since Zqj [i] is a field and Γ(Zq1 q2 [i]) ∼
= Γ(Zq1 [i])×Γ(Zq2 [i]), therefore
Γ(Zq1 q2 [i]) ∼ K
= q1 −1,q2 −1
2 2 is a complete bipartite graph.
576 M. Imran Bhat and S. Pirzada

The following result gives a relation between the maximum degree,

diameter and metric dimension of Γ(R).
Theorem 3.20. [19] Let R be a finite commutative ring with unity
1 such that |Z ∗ (R)| ≥ 2 with diameter d. Then
dlog3 (∆ + 1)e ≤ dim(Γ(R)) ≤ |Z ∗ (R)| − d,
where ∆ is the maximum degree of Γ(R).
We observe that the lower and upper bounds of Theorem 3.20 also
hold when dim(Γ(R)) is replaced by dims (Γ(R)).
Theorem 3.21. Let R be a finite commutative ring with unity 1 such
that |Z ∗ (R)| ≥ 2 with diameter d. Then
dlog3 (∆ + 1)e ≤ dims (Γ(R)) ≤ |Z ∗ (R)| − d,
where ∆ is the maximum degree.
Proof. We first establish the upper bound. Let u and v be the vertices
for which d(u, v) = sup{d(x, y) | x, y ∈ Z ∗ (R)}, that is, d(u, v) is the
diameter of Γ(R) and let u = v1 , v2 , . . . , vd = v be u − v path of length
d. Since W = V (Γ(R)) − {ui | 1 ≤ i ≤ d} forms a strong resolving set
for Γ(R) with |W | = n − d, so dims (Γ(R)) ≤ n − d.
Now, for the lower bound, since dlog3 (∆+1)e ≤ dim(Γ(R)) ≤ dims (Γ(R)),
it follows that dims (Γ(R)) ≥ dlog3 (∆ + 1)e.
Theorem 3.22. If R is a finite commutative ring, then dims (Γ(R)) ≤
|∂(Γ(R))| − 1.
Proof. If R is a finite commutative ring and Γ(R) be its corresponding
zero-divisor graph with vertex set |Z ∗ (R)|, then dim(Γ(R)) ≤ |Z ∗ (R)|−1
implies dims (Γ(R)) = α((Γ(R))SR ) ≤ |∂(Γ(R))| − 1.
Definition 3.23. For a commutative ring R with 1 6= 0, a compressed
zero-divisor graph of a ring R is the undirected graph ΓE (R) with vertex
set Z(RE ) r {[0]} = RE r {[0], [1]} defined by RE = {[x] : x ∈ R}, where
[x] = {y ∈ R : ann(x) = ann(y)} and the two distinct vertices [x] and
[y] of Z(RE ) are adjacent if and only if [x][y] = [xy] = [0], that is, if and
only if xy = 0.
The authors in [18] have discussed the metric dimension of compressed
zero-divisor graphs ΓE (R). We have the following observations.
 On strong metric dimension of zero-divisor graphs of rings 577

Theorem 3.24. If R is a finite commutative ring, then dimE (Γ(R)) ≤

dim(Γ(R)) ≤ dims (Γ(R)).
Proof. Since ΓE (R) has a vertex set constructed by taking equivalence
of zero-divisors of a ring R, therefore [x] ⊂ Z(R) r {0} implies that each
vertex of ΓE (R) is a representative of a distinct class of zero-divisors
actually in R. Hence, dim(ΓE (R)) ≤ dim(Γ(R)). Also, we know that
dim(Γ(R)) ≤ dims (Γ(R)).
Proposition 3.25. (i) dims (ΓE (R)) = 0 if ΓE (R) ∼ = Kn unless
R∼= Z2 × Z2 .
(ii) dims (ΓE (R)) = 1 if ΓE (R) ∼
= Km,n , m or n ≥ 2.
(iii) dims (ΓE (R)) = n − 1 if ΓE (R) = K1,n , n ≥ 2.

4. Strong metric dimension of Cartesian products

The Cartesian product of two graphs G1 and G2 is the graph G1 × G2

whose vertex set is V = V (G1 ) × V (G2 ) and the two vertices w1 =
(u1 , v1 ) and w2 = (u2 , v2 ) in V ; u1 , v1 ∈ V (G1 ) and u2 , v2 ∈ V (G2 ) are
adjacent in G1 × G2 if and only if (a) either u1 = u2 and v1 v2 ∈ E(G2 )
or v1 = v2 and u1 u2 ∈ E(G1 ).
Let S be a set of vertices in the Cartesian product G1 × G2 . The projec-
tion of S onto G1 is the set of vertices a ∈ V (G1 ) for which there exists
a vertex (a, v) ∈ S. The same is defined similarly for G2 .

We have the following observation about cartesian product of two

graphs.
Lemma 4.1. For any graphs G1 and G2 , ∂(G1 × G2 ) = ∂(G1 ) × ∂(G2 ).
Proof. Suppose (u, v) ∈ ∂(G1 × G2 ) and u ∈ / ∂(G1 ). Then, for every
u1 ∈ V (G1 ), there exists u2 ∈ NG1 (u) such that dG1 (u, u1 ) < dG1 (u1 , u2 ).
Now, consider a vertex (u2 , v) ∈ NG1 ×G2 (u, v). Then, for arbitrary
v1 ∈ V (G2 ), we have dG1 ×G2 ((u1 , v1 ), (u2 , v)) = dG1 (u1 , u2 )+dG2 (v1 , v) >
dG1 (u1 , u) + dG2 (v1 , v) = dG1 ×G2 ((u1 , v1 ), (u, v)), a contradiction to the
assumption (u, v) ∈ ∂(G1 × G2 ). Thus, u ∈ ∂(G1 ). Similarly, we can
prove that v ∈ ∂(G2 ).
Now, let u ∈ ∂(G1 ) and v ∈ ∂(G2 ). Thus there exists a vertex u1 ∈
V (G1 ) such that for every u2 ∈ NG1 (u), we have dG1 (u, u1 ) ≥ dG1 (u1 , u2 )
and there is a vertex v1 ∈ V (G2 ) such that for every v2 ∈ NG2 (v), we
578 M. Imran Bhat and S. Pirzada

have, dG2 (v, v1 ) ≥ dG2 (v1 , v2 ). Let (u2 , v2 ) be an arbitrary vertex from
NG1 ×G2 (u, v). Without loss of generality, assume that u2 u ∈ E(G1 ) and
v2 = v. Then dG1 ×G2 ((u1 , v1 ), (u2 , v2 )) = dG1 (u1 , u2 ) + dG2 (v1 , v2 ) ≤
dG1 (u1 , u) + dG2 (v1 , v) = dG1 ×G2 ((u1 , v1 ), (u, v)) and (u, v) ∈ ∂(G1 ×
G2 ).
Here, we observe that V ((G1 × G2 )SR ) = ∂(G1 × G2 ) = ∂(G1 ) ×
∂(G2 ) = V ((G1 )SR × V (G2 )SR ).
Theorem 4.2. Let R be a finite commutative ring with unity 1 6= 0.
Then dims (Γ(R) × Γ(Z2 × Z2 )) = 2 if and only if Γ(R) is a path.
Proof. If Γ(R) is a path, then
dims (Γ(R) × Γ(Z2 × Z2 )) = α((Γ(R) × Γ(Z2 × Z2 ))SR )
= α((Γ(R))SR × Γ(Z2 × Z2 )SR )
= α(K2 × K2 ) = 2.
Now, Let G = Γ(R) × Γ(Z2 × Z2 ) and let W = {(u, v), (u1 , v1 )} be a
strong metric basis of G. We consider the following two cases.
Case 1. If u 6= u1 , let w1 be a neighbor of u1 on a u − u1 path. Since
W is a strong metric basis, each pair of vertices of G by definition is re-
solved by a vertex of W . We choose (u1 , v) and (w1 , v1 ). Then we have
dG ((u1 , v), (u, v)) = dΓ(R) (u, u1 ) + dΓ(Z2 ×Z2 ) (v, v) = dΓ(R) (u, w1 ) + 1 +
dΓ(Z2 ×Z2 ) (v, v1 ) − 1 = dΓ(R) (u, w1 ) + dΓ(Z2 ×Z2 ) (v, v1 ) = dG ((w1 , v1 ), (u, v)).
Thus,(u1 , v) ∈ / IG [(w1 , v1 ), (u, v)] and (w1 , v1 ) ∈
/ IG [(u, v), (u1 , v)]. More-
over,
dG ((u1 , v), (u1 , v1 )) = dΓ(Z2 ×Z2 ) (v, v1 ) = 1
= dΓ(R) (u1 , w1 ) = dG ((w1 , v1 ), (u1 , v1 )).
Thus, (u1 , v) ∈ / IG [(w1 , v1 ), (u1 , v1 )] and (w1 , v1 ) ∈
/ IG [(u1 , v1 ), (u1 , v)].
Therefore, S = {(u, v), (u1 , v1 )} does not strongly resolve (u1 , v) and
(w1 , v1 ) and so u = u1 .
Case 2. If u = u1 , then the projection of W onto Γ(R) is a single
vertex and therefore the projection of W onto Γ(R) strongly resolves
Γ(R). Hence Γ(R) is a path.

References
[1] M. Aijaz and S. Pirzada, On annihilating ideal graphs of commutative rings,
Asian-European J. Math. to appear.
 On strong metric dimension of zero-divisor graphs of rings 579

[2] S. Akbari and A. Mohammadian, On zero divisor graphs of a commutative ring,

J. Algebra 274 (2004), 847–855.
[3] D. F. Anderson and A. Badawi, The total graph of a commutative ring, J. Al-
gebra 320 (7) (2008) 2706–2719.
[4] D. F. Anderson, A. Frazier, A. Lauve and P. S. Livingston, The Zero-Divisor
Graph of a Commutative Ring, II, Lecture Notes in Pure and Appl. Math.,
Marcel Dekker, New York, 220 (2001), 61–72.
[5] D. F. Anderson and P. Livingston, The zero divisor graph of a commutative ring,
J. Algebra 217 (1999), 434–447.
[6] D. F. Anderson, R. Levy and J. Shapiro, Zero-divisor graphs, von Neumann
regular rings, and Boolean algebras, J. Pure Appl. Algebra 180 (2003), 221–241.
[7] N. Ashrafi, H. R. Maimani, M. R. Pournaki and S. Yassemi, Unit graphs asso-
ciated with rings, Comm. Algebra 38 (2010), 2851–2871.
[8] M. F. Atiyah and I. G. MacDonald, Introduction to Commutative Algebra,
Addison-Wesley, Reading, MA (1969).
[9] A. Azimi, A. Erfanian, and M. Farrokhi D.G., The Jacobson graph of commu-
tative rings, J. Algebra Appl. 12 (3) (2013), 1250179.
[10] M. Baziar, E. Momtahan, and S. Safaeeyan, A zero-divisor graph of modules
with respect to their (first) dual, J. Algebra Appl. 12 (3) (2013), 93–106.
[11] I. Beck, Coloring of commutative rings, J. Algebra 116 (1988) 208–226.
[12] J. Cceres, C. Hernando, M. Mora, I. M. Pelayo, M. L. Puertas, C. Seara and
D. R. Wood, On the metric dimension of cartesian products of graphs, SIAM J.
Discrete Math. 21 (2) (2007), 423–441.
[13] G. Chatrand, L. Eroh, M. A. Johnson and O.R. Ollermann, Resolvability in
graphs and the metric dimension of a graph, Discrete Appl. Math. 105 (2000),
99–113.
[14] F. Harary, and R. A. Melter, On the metric dimension of a graph, Ars Combin.
2 (1976) 191–195.
[15] I. Kaplansky, Commutative Rings, rev. ed., Univ. of Chicago Press, Chicago,
(1974).
[16] Ortrud R. Oellermann and Joel Peters-Fransen, The strong metric dimension of
graphs and digraphs, Discrete Appl. Math. 155 (2007), 356–364.
[17] S. Pirzada, An Introduction to Graph Theory, University Press, Orient Black-
swan, Hyderabad, India, (2012).
[18] S. Pirzada and M. Imran Bhat, Computing metric dimension of compressed zero
divisor graphs associated to rings, Acta Univ. Sap. Mathematica 10 (2) (2018),
298–318.
[19] S. Pirzada and Rameez Raja and S. P. Redmond, On locating numbers and codes
of zero-divisor graphs associated with commutative rings, J. Algebra Appl. 13
(7) (2014), 1450047.
[20] S. Pirzada, M. Aijaz and S. P. Redmond, Upper dimension and bases of zero
divisor graphs of commutative rings, AKCE International J. Graphs Comb. to
appear.
[21] S. Redmond, The zero-divisor graph of a non-commutative ring, International
J. Comm. Rings 1 (4) (2002) 203–211.
580 M. Imran Bhat and S. Pirzada

[22] S. P. Redmond, An Ideal based zero divisor graph of a commutative ring, Comm.
Algebra 31 (9) (2003), 4425–4443.
[23] S. P. Redmond, Cut vertices and degree one vertices of zero-divisor graphs,
Comm. Algebra 40 (8) (2012), 2749–2756.
[24] S. P. Redmond, Recovering rings from zero-divisor graphs, J. Algebra Appl. 12
(8) (2013), 1350047 (9 pages).
[25] S. Spiroff and C. Wickham, A zero divisor graph determined by equivalence
classes of zero divisors, Comm. Algebra 39 (7) (2011), 2338–2348.
[26] D. B. West, Introduction Graph Theory, 2nd ed. USA: Prentice Hall, (2001).

M. Imran Bhat
Department of Mathematics
University of Kashmir
Srinagar, 190006, Srinagar, Kashmir, India
E-mail : imran [email protected]
Shariefuddin Pirzada
Department of Mathematics
University of Kashmir
Srinagar, 190006, Kashmir, India
E-mail : [email protected]