Fuzzy Threshold Graphs

Sovan Samanta and Madhumangal Pal

Abstract- In this paper, fuzzy threshold graphs, fuzzy alternating

4 -cycles, threshold dimension of fuzzy graphs and fuzzy Ferrers digraphs are defined. We show that fuzzy threshold graphs are fuzzy split graphs. Forbidden configurations of fuzzy Ferrers digraph are described. Also some basic theorems related to the stated graphs have been presented.
Keywords- Fuzzy Ferrers digraph, fuzzy graph, fuzzy threshold dimension, fuzzy threshold graph. 1. INTRODUCTION Threshold graphs play an important role in graph theory as well as in several applied areas such as psychology, computer science, scheduling theory, etc. These graphs can be used to control the flow of information between processors, much like the traffic lights used in controlling the flow of the traffic. Chvatal and Hammer [3] coined the name threshold graphs" and studied the graphs for their application inset packing problems. Ordman [10] found the use of graphs in resource allocation problems. Chvatal and Hammer defined threshold graph as follows. A graph G = (V , E ) is a threshold graph when there exists non-negative reals wv , v V and t such that W (U ) t if and only if

A split graph is a graph in which the vertices can be partitioned into a clique and an independent set. Alternating 4 -cycle of a graph G = (V , E ) is a

a, b, c, d such that ( a, b), (c, d ) E and ( a, c ), (b, d ) E . By considering the presence or absence of edges ( a, d ), (b, c ) , we see that the vertices of alternating 4 -cycle induce a path P4 , a square C4 , or a matching 2 K 2 . For the graph G = (V , E ) with n vertices and distinct positive vertex degrees 1 < 2 < < m , 0 = 0 (even no vertex of degree 0 exists) and m +1 =| V | 1 , degree partition is the sequence m +1 n , Di = {v V :deg (v) = i } for i = 0,1, , m .
configuration consisting of distinct vertices 1.1 Fuzzy graphs The fuzzy systems have been used with success in last few years, in problems that involve the approximate reasoning. The objective of the work is to specify the fuzzy systems with the addition of the interval theory in its components. A fuzzy set A on a set X is characterized by a mapping m : X [0,1] , called the membership function. We shall denote a fuzzy set as A = ( X , m) . Fuzzy graph theory was introduced by Azriel Rosenfeld [12] in 1975 . Though it is very young, it has been growing fast and has numerous applications in various fields. A fuzzy graph = (V , , ) is a non empty crisp set V together with a pair of functions : V [0,1] and : V V [0,1] such that for all

U V

is

stable

set

where

W (U ) = wv . So
vU

G = (V , E ) is a threshold graph whenever one can assign

vertex weights such that a set of vertices is stable if and only if its total weight does not exceed a certain threshold. The threshold dimension, t (G ) of a graph G is the minimum

k of threshold subgraphs T1 , T2 , , Tk of G that cover the edge set of G . Threshold partition number, denoted by tp (G ) , is the minimum number of edge disjoint threshold subgraphs needed to cover E (G ) . Formally, an edge cover of a graph G is a set of edges C E such that each vertex is incident with at least one edge in C . The set C is said to cover the vertices of G .
number

G = (V , E ) is said to be a Ferrers digraph if it does not contain vertices x, y , z , w , not necessarily distinct, satisfying and ( x, y ), ( z , w) E ( x, w), ( z , y ) E . For a digraph G = (V , E ) , the underlying loop less graph U (G ) = (V , E ) , where E = {(u , v) : u , v V , u v, (u, v ) E} .
graph. A digraph

Ferrers digraph [11] is a digraph related to threshold

x, y ; ( x, y ) ( x ) ( y ) . So is a fuzzy relation on . A fuzzy digraph = (V , , ) is similarly defined except : V V [0,1] need not be symmetric. We call the pairs (( x, y ), ( x, y )) fuzzy arcs to emphasize that the symmetry is not required. The fuzzy graph ' = (V , , ) is called a partial fuzzy subgraph of if and . Similarly, the fuzzy graph ' = ( P, , ) is called fuzzy subgraph of induced by P if P V , ( x) = ( x ) for all x P and ( x, y ) = ( x, y ) for all x, y P . The degree of a vertex u is d (u ) = (u , v) = (u , v ) . The minimum
u v (V )

degree of

is ( ) = {d (v ) | v V } . The maximum

( u ,v ) > 0

 is ( ) = {d (v) | v V } . The strength of connectedness between two vertices u and v is (u , v ) =

degree of

( x, y ) = { (u , v) | u , v supp( )} . ( , ) is a clique if ( supp ( ), supp( )) is a clique. ( , ) is fuzzy

clique if it is a clique and every cycle in it is a fuzzy cycle. 1.2 Review of previous works Andelic and Simic [1] presented some results on the threshold graphs. Bhutani, Moderson and Rosenfeld [2] discussed on degrees of end nodes and cut nodes in fuzzy graphs. Mathew and Sunitha [4] defined types of arcs in a fuzzy graph. Mordeson and Nair [5] have given the details of fuzzy graphs and hypergraphs. Nagoorgani and Malarvizhi [6] decribed the isomorphism properties of strong fuzzy graphs. Nagoorgani and Radha [7] proved some results on regular fuzzy graphs. Nair and Cheng [8] discussed cliques and fuzzy cliques in fuzzy graphs. Natarajan and Ayyasawamy [9] described on strong (weak) domination in fuzzy graphs. 1.3 Our works In this paper, we define fuzzy threshold graph and investigated some of its important properties. Here we use the notation Cn, Pn, Kn , K for cycle, path and complete graph of length n in fuzzy graph and fuzzy clique respectively. 2. FUZZY THRESHOLD GRAPHS Definition 1 A fuzzy graph = (V , , ) is called a fuzzy threshold graph if there exists non negative real number T such that

sup{ k (u , v) | k = 1,2,...} where k (u , v ) = sup{ (u, u1 ) (u1 , u2 ) (uk 1 , v) | u1 , u2 , , uk 1 V } . An edge (u, v) is a fuzzy bridge of if deletion of (u , v)
reduces the strength of connectedness between the pair of vertices. A vertex u is a fuzzy cut vertex of if deletion of u reduces the strength of connectedness between other pair of vertices. The order of fuzzy graph We use the notation of order of graph

is O( ) = (u ) .
uV

is S ( ) =

( u ,v ) > 0

as | | . The size of fuzzy (u , v) . In a fuzzy graph an arc

(u, v) is said to be strong arc [5] or strong edge, if (u , v) (u , v) and the node v is said to be strong neighbour of u . If (u , v) is not strong arc then v is called isolated node or isolated vertex. D V is said to fuzzy dominating set of if for every v V D , there exists u in D such that (u , v) = (u ) (v) . The minimum scalar cardinality of D is called fuzzy domination number and it is denoted by f ( ) . Scalar cardinality of a fuzzy set D is | D | f = (u ) . Two nodes in a fuzzy graph are said
uV

to be fuzzy independent if there is no strong arc between them. Let u be a node in fuzzy graph then N (u ) = {v : (u, v) is a strong arc} is called open neighborhood of u and N [u] = N(u) {u} is called closed neighborhood of u . A path in fuzzy graph is a sequence of distinct

x0 , x1 , , xn such that ( xi 1 , xi ) > 0,1 i n . Here n > 0 is called the length of the path. The strength of is defined as in=1 ( xi 1 , xi ) . In other words, strength of
nodes a path is the weight of the weakest arc of the path. If the path has length 0 , then its strength is ( x0 ) . We call a cycle if x0 = xn and n 3 . We recall that a graph without cycle is called acyclic or forest, a connected forest is tree. We call a fuzzy graph a forest if the graph consisting of its nonzero arcs is a forest. We call fuzzy graph a fuzzy forest if it has a fuzzy spanning subgraph ' = ( , ) which is a forest, where for

uU

(u) T

if and only if U V is stable set in

= (V , , ) be a fuzzy graph with vertex set V = {a, b, c, d , e} such that (a) = 0.1, (b) = 0.5, (c) = 0.2, (d ) = 0.2, (e) = 0.2 . Also let T = 0.5 and all cycles of the graph are fuzzy cycles.
Example 1 Let This graph is an example of fuzzy threshold graph (see Fig. 1).

( x, y ) not in the subgraph ' , we have ( x, y ) < ( x, y ) . Thus if ( x, y ) but ( x, y ) ' , there is a path in ' between x and y whose strength is greater than ( x, y ) . ( , ) is a cycle if and only if ( supp ( ), supp( )) is a cycle. ( , ) is a fuzzy cycle if and only if ( supp ( ), supp( )) is a cycle and there does not exist unique ( x, y ) , x, y supp( ) such that
all arcs

Fig. 1: Example of fuzzy threshold graph. Now, we define fuzzy alternating

4 cycle.

Definition 2 Let = (V , , ) be a fuzzy graph and V = {a, b, c, d } . Also let (a ), (b), (c), (d ) , (a, b), (c, d ) be positive and (a, c) = (b, d ) = 0 . This configuration of four vertices is called fuzzy alternating 4 -cycle. This fuzzy alternating 4 -cycle induces a path P 4

= (V , , ) be a fuzzy graph whose distinct positive vertex degrees are 1 < 2 < < p , and
Definition 4 Let let

(a, d ), (b, c ) is zero and other is nonzero), a square C 4 (when both ( a, d ), (b, c ) are nonzero) or a matching 2K2 (when both ( a, d ), (b, c ) are
(when one of zero). Fuzzy alternating

0 = 0 (even if no isolated vertex exists), p +1 =| V | 1 . Let D i = {v V : i d (v) < i + 1 } for non negative integer i p . The sequence D0 , D1,, Dp is called degree partition of the fuzzy graph .
Theorem 1 A fuzzy threshold graph does not have a strong fuzzy alternating 4 -cycle.

4 -cycle may also induce a fuzzy

C 4 if the arcs membership values don't have exactly one minimum value. The fuzzy alternating 4 -cycle and its
induced fuzzy graphs are shown in Fig. 2. Definition 3 A strong alternating 4 -cycle is an alternating -cycle if fuzzy C4 be induced from it.

= (V , , ) be a fuzzy threshold graph. Let, if possible, has a strong fuzzy alternating 4 -cycle. So there exists vertices a, b, c, d with ( a, b) > 0 , (c, d ) > 0 and ( a, d ) = 0 = (b, c) . As the graph is fuzzy threshold graph with threshold T , then (a) + (b) > T , (c) + (d ) > T [as the a, b, c, d construct strong fuzzy alternating 4 -cycle] (a) + (d ) T , (b) + (c) T [as (a, d ) = 0 = (b, c) ]. These inequalities are inconsistent. Hence a, b, c, d does not construct a strong fuzzy alternating 4 -cycle. So a
Proof. Let fuzzy threshold graph does not have a strong fuzzy alternating 4 -cycle. Theorem 2 A fuzzy threshold graph is a fuzzy split graph.

P4

= (V , , ) be a fuzzy threshold graph and K =(K, , ) , K V, (x)= (x) for all x K, (x,y)= (x,y) for all x,y K , be the largest clique in . If (a, b) be a strong arc in | V K | , then by maximality of K , there exists distinct vertices c , d in K such that (a, d ) = 0 = (b, c) . These vertices create a strong fuzzy alternating 4 -cycle. So contradiction arises. To avoid strong fuzzy alternating 4 -cycle, | V K | must be stable set. Hence is a fuzzy split graph.
Proof. Let Theorem 3 If is a fuzzy threshold graph, then can be constructed from the one vertex graph by repeatedly adding a fuzzy isolated vertex or a fuzzy dominating vertex.

C4

2K2 4-

Fig. 2: Induced subgraphs of fuzzy alternating cycle.

Degree partition is very important term in graph theory. Here we define degree partition in a fuzzy graph.

Proof. In Theorem 2, we proved that a fuzzy threshold graph is a fuzzy split graph. It is enough to show that has a fuzzy isolated vertex or a fuzzy dominating vertex, if for such a vertex is removed, then the graph will be still fuzzy split graph. Let S be a stable set of the fuzzy threshold graph = ( K , S). Let S be nonempty. If S contains fuzzy isolated vertices only, then the result holds. If S has no fuzzy isolated

vertices, then vertex u S with smallest neighborhood has some neighbor v K. As K is a fuzzy clique, so the vertex v is dominating vertex of . Comparability in fuzzy graph is important topic. We now define strong comparability of vertices in fuzzy graph. Definition 5 Let = (V , , ) be a fuzzy graph. Two vertices x and y are said to be strong comparable if there exists a path from x to y or y to x whose every arc is strong arc. We now give the definition of threshold dimension in fuzzy graph.

weak fuzzy arc of stable set to the stars then they satisfy the condition of fuzzy threshold graph. So all such stars together with weak arcs of stable sets covers the edge set of . Thus

~ t ( ) | V S | . Again we know that | V |= n and ( ) | S | , V , S being the crisp sets. Thus ~ t ( ) (n ( )) . ~ We know |S|=|supp(S)|. So t ( ) ( n | supp ( S ) | . If in addition is fuzzy triangle free, then every fuzzy

threshold graph is a star or star together with weak edges. So ~ t ( ) (n | supp ( S ) |) . Hence ~ t ( ) = (n | supp ( S

~ Definition 6 The threshold dimension t ( ) of a fuzzy graph = (V , , ) is the minimum number k of fuzzy threshold subgraphs T 1 , T 2 , , T k of that cover the edge set of

) |) .

that is if ' = T1 T 2, Tk then ' = (V , , ) .

Fig. 3: A fuzzy graph with

~ t ( ) = 2 and ~ t p ( ) = 3 .

Since every fuzzy edge along with fuzzy isolated vertices is a fuzzy threshold subgraph, the threshold dimension of fuzzy graph is well defined and is bounded by the number of edges of the graph. We denote ( ) as stability number of a fuzzy graph

Fig. 4: Forbidden configurations in fuzzy Ferrers digraph. Fuzzy threshold partition is as important as threshold dimension. ~ Definition 7 The fuzzy threshold partition number t p ( ) of the fuzzy graph is the minimum number of fuzzy threshold subgraphs, not containing common strong arcs, cover edge set of . Example 2 Here we give an example of fuzzy graph whose fuzzy threshold dimension number is 2 and fuzzy partition number is 3 . The graph has vertices with membership values

, i.e., the order of the largest stable set of

= (V , , ) on n ~ vertices we have t ( ) ( n ( )) . Furthermore, if is ~ triangle-free, then t ( ) = ( n | supp ( S ) |) where S is

Theorem 4 For every fuzzy graph the stable set with largest number of vertices. Proof. Let S = (S, ) be a stable set with largest number of vertices of the fuzzy graph with n vertices. For each vertex u V S , we consider the star centered at u . Each such star is a fuzzy threshold graph. If we add one or more

a (0.1), a1 (0.1), b(0.5), b1 (0.5), c(0.4), c1 (0.4), e(0.4),

d(0.2), d1(0.2) and strong arcs shown in Fig. 3.

Theorem 5 If

is a fuzzy triangle free graph, then ~ ~ t ( ) = t p ( ) .

Proof. Let

REFERENCES [1]
M. Andelic and S.K. Simic, Some notes on the threshold graphs, Discrete Mathematics, 310 , 2241 2248 , 2010 . [2] K. R. Bhutani, J. Moderson and A. Rosenfeld, On degrees of end nodes and cut nodes in fuzzy graphs, Iranian Journal of Fuzzy Systems, 1(1) , 53 60 , 2004 . [3] V. Chvatal and P. L. Hammer, Set-packing problems and threshold graphs, CORR 73 21 , University of Waterloo, Canada, 1973 . [4] S. Mathew and M. S. Sunitha, Types of arcs in a fuzzy graph, Information Sciences, 179 , 1760 1768 , 2009 . [5] J. N. Mordeson and P. S. Nair, Fuzzy graphs and hypergraphs, Physica Verlag, 2000 . [6] A. Nagoorgani and J. Malarvizhi, Isomorphism properties of strong fuzzy graphs, International Journal of Algorithms, Computing and Mathematics, 2(1) , 39 47 , 2009 . [7] A. Nagoorgani and K. Radha, On regular fuzzy graphs, Journal of Physical Sciences, 12 , 33 40 , 2008 . [8] P. S. Nair and S. C. Cheng, Cliques and fuzzy cliques in fuzzy graphs, IFSA World Congress and 20th NAFIPS International Conference, 4 , 2277 2280 , 2001 . [9] C. Natarajan and S. K. Ayyasawamy, On strong (weak) domination in fuzzy graphs, World Academy of Science, Engineering and Technology, 67 , 526 528 , 2010 . [10] E. T. Ordman, Threshold coverings and resource allocation, In 16 th Southeastern Conference on Combinatorics, Graph Theory and Computing, 99 113 , 1985. [11] U. N. Peled and N. V. Mahadev, Threshold graphs and related topics, North Holland, 1995 . [12] A. Rosenfeld, Fuzzy graphs, in: L.A. Zadeh, K.S. Fu, M. Shimura (Eds.), Fuzzy Sets and Their Applications, Academic Press, New York, 77 95 , 1975 .

= (V , , ) be a fuzzy graph. We know that ~ the edge set can be covered by t ( ) number of stars. If a
strong arc belongs to more than one stars then we delete it from all but remain one of the stars. This gives a fuzzy ~ threshold partition of size t ( ) . Fuzzy Ferrers digraph is an important graph related to fuzzy threshold graph.

= (V , , ) is said to be a fuzzy Ferrers digraph if it does not contain vertices x, y , z , w , not necessarily distinct, satisfying ( x, y ) , ( z , w) are non zero and ( x, w) , ( z , y ) are zero.
Definition 8 A fuzzy digraph

Forbidden configurations are shown in Fig. 4. Here solid arrow represents that edge membership values are non zero and dotted arrow represents that edge membership values are zero.

= (V , , ) be a fuzzy digraph. The underlying fuzzy graph of is denoted by U ( ) and is defined as U ( ) = (V , , ) where (u , v) = min { (u , v), (v, u )} for all u , v V .
Definition 9 Let

is a symmetric fuzzy Ferrers digraph, then its underlying fuzzy undirected loop less fuzzy graph U ( )
Theorem 6 If may or may not be fuzzy threshold graph.

is a symmetric fuzzy Ferrers digraph, then its underlying fuzzy undirected loop less fuzzy graph U ( ) has no fuzzy alternating 4 cycle. So it may contain fuzzy cycle. Hence U ( ) may be fuzzy threshold graph. If U ( ) does not have strong fuzzy alternating 4 cycle then U ( ) must be fuzzy threshold graph using the result of theorem 1 .
Proof. If 3. CONCLUSION In this paper we have presented the relations between fuzzy threshold graphs and fuzzy split graphs, fuzzy threshold dimension and fuzzy partition number. Also we discussed some properties of fuzzy Ferrers digraphs. These graphs will help to solve resourse allocation problems in fuzzy approach. Also these graph are helpful to control the flow of information using fuzzy properties. We hope our study will enable us to extend the fuzzy graph classes like fuzzy difference graphs, fuzzy matroidal graphs, fuzzy matrogenic graphs.

Madhumangal Pal is a Professor of Applied Mathematics with Oceanology and Computer Programming, Vidyasagar University, India. He received University Silver Medal for rank second in B.Sc. (Honours) in the year 1988 from Vidyasagar University, India. He received University Gold Medal for rank first in M.Sc. in the year 1990 from the same university. He received Computer Division medal from Institution of Engineers (India) in the year 1996 for the best research work published in the Institution journal jointly with Prof. G.P.Bhattacharjee .

He is also Editor-in-Chief of Journal of Physical Sciences. He is member of Editorial Board of International Journal of Computer Sciences, Systems Engineering and Information Technology, International Journal of Fuzzy Systems & Rough Systems, Advanced Modeling and Optimization, Romania International Journal of Logic and Computation, Malaysia. He is a reviewer of several international journals. He has written several books on Mathematics and Computer Science. His research interest includes computational graph theory, fuzzy matrices, game theory and regression analysis, parallel and genetic algorithms, etc.

Sovan Samanta received his B.Sc. degree in 2007 and M.Sc. degree in 2009 in Applied Mathematics from Vidyasagar University, India. He is now currently a research scholar in the Department of Applied Mathematics, Vidyasagar University since 2010 . His research interest includes fuzzy graph theory .