Korean J. Math. Sciences Vol. 10 (2003) No. 2, pp.

712
Available free at mirror sites of http://www.kjms.or.kr KJMS
SOME FIXED POINT THEOREMS FOR MULTI-VALUED
A*-MAPPINGS

M.Akram A.A.Siddiqui and A.A.Zafar

Abstract. In [8], F.U. Rehman and B. Ahmad presented generalization

of some interesting results of [2],[4],[6] for M.S. Khan type multi-valued
maps (Cf [7]). Here, we introduce a general class of multi-valued maps,
called A -maps, containing all Kannan type multi-valued maps (Cf [?])
that in turn properly contains all M.S. Khan type multi-valued maps and
extend main results of [8] for general A -maps.

1. Introduction

Several authors studied fixed point theorems of multi-valued mappings on

complete metric spaces. The work of B. Fisher, K. Iseiki [3] and Rhodes [7] in
this connection is worth mentioning. F.U. Rehman and B. Ahmad presented
some interesting results on fixed point by improving certain results of [2],[4],[6].
Here, we shall extend the main results of [8] to a general A -maps. As usual
in any mathematical discourse, we need for the sequel to fix some Symbolism:
For setsA and B in a metric space X, we shall use the symbols D(A, B),
(A, B) and H(A, B) to mean inf {d(a, b) : a A and b B}, sup{d(a, b) :
a A and b B} and max{Sup{D(a, B) : a A}, Sup{D(A, b) : b B}},
respectively. Let R+ denote the set of all non-negative real numbers and A
stands for the set of all functions : R+ 3 R+ satisfying

(i) is continuous on the set R+ 3 of all triplets of non-negative reals (with

respect to the Euclidean metric on R3 );

(ii) is non-decreasing in each coordinate variables;
(iii) a kb for some k [0, 1) whenever a (a, b, b) or a (b, a, b) or
a (b, b, a) for all a, b.
A multi-valued mapping T : X 2X is said to be A -map:
if (T x, T y) ((T x, x), (T y, y), (x, y)) for some A and for all x, y
X. Through out the sequel, CB(X) would mean the set of all closed and
Received June 18, 2003. Revised October 8, 2003.
2000 Mathematics Subject Classification: Primary 54H25, secondary 47H10.
Key words and phrases: Complete metric spaces, fixed point and A*-maps.
8 M.Akram A.A.Siddiqui and A.A.Zafar

bounded subsets of X, where X is a complete metric space unless otherwise

stated. Recall (from [8], for instance) that a multi-valued mapping T : X
R(T ) is said to have x X as its fixed point if x T x where R(T ) is a set of
nonempty subsets of X. Now, we are in position to present some fixed point
results, which extend the correspondence results of [8] and some others.
Theorem 1.1. If S, T : X CB(X) are mappings such that
(Sx, T y) ((Sx, x), (T y, y), (x, y)) for some A and for all x, y
X. Then S and T have a common fixed point.
Proof. Let xo X, choose a point x2n1 in X2n1 = Sx2n2 and x2n in
X2n = T x2n1 for n = 1, 2, 3. Then for any n N ,
(X2n+1 , X2n+2 ) = (Sx2 n, T x2n+1 )
((Sx2n , x2n ), (T x2n+1 , x2n+1 ), (x2n , x2n+1 ))
((X2n+1 , x2n ), (X2n+2 , x2n+1 ), (x2n , x2n+1 ))
((X2n+1 , X2n ), (X2n+2 , X2n + 1), (X2n , X2n+1 ))
and by property (iii) of , we get

(X2n+1 , X2n+2 ) k(X2n+1 , X2n )for somek [0, 1).

Repeating this step, we get (X2n+1 , X2n+2 ) k 2n+1 (Xo , X1 ) n N .

Which in fact gives that d(xn , xn+1 ) (Xn , Xn+1 ) k n (Xo , X1 ) n N .
Pni
Thus, for any m, n N with m < n, we see that d(xm , xn ) i=m d(xi , xi+1 )
Pni i
(Xo , X1 ) i=m k 0 as m, n since k [0, 1). Therefore, (xn ) is a
Cauchy sequence in X. Since X is complete so xn x in X.
Since

(Sx, x2n ) = (Sx, T x2n1 ) ((Sx, x), (T x2n1 , x2n1 ), (x, x2n1 ))
((Sx, x), (X2n , X2n1 ), (x, x2n1 ))
((Sx, x), k 2n1 (X0 , X1 ), (x, x2n1 ))
when n , we have, (Sx, x) ((Sx, x), 0, 0)). Hence (Sx, x) k(0)
with k [0, 1) by property (iii) of . Thus (Sx, x) = 0. D(Sx, x) (Sx, x)
implies D(Sx, x) 0. This implies D(x, Sx) = 0. Which gives x Sx . But
Sx = Sx as Sx CB(X). Therefore, x Sx. Similarly, one can show that
x T x. Thus, x is a common fixed point of S and T .
Theorem 1.2. If S, T : X CB(X) are mappings such that
 A fixed point theorem 9

(1) d(Sx, T y) h[d(x, Sx) + d(y, T y)]

1
for all x, y in X and 0 < h < 2 , then S and T have common fixed point.
3 R as (u, v, w) = h(u + v) where 0 < h < 1 .
Proof. By defining : R+ + 2
One can easily see that A . So by taking u = d(Sx, x), v = d(T x, x) and
w = d(x, y), (1) becomes d(Sx, T y) (d(x, Sx), d(y, T y), d(x, y)). Hence by
Theorem 1.1 we have the conclusion.
Theorem 1.3. (see Theorem 1 of [8]) If S, T : X CB(X) are mappings
such that
1
(2) (Sx, T y) h[(x, Sx)(y, T y)] 2
for all x, y in X and 0 < h < 1, then S and T have common fixed point .
1
3 R+ as (u, v, w) = h(uv) 2 where 0 < h < 1 ,
Proof. If we define : R+
then one can easily see that A . Now taking u = (x, Sx), v = (y, T y)
1
and w = (x, y) we have ((x, Sx), (y, T y), (x, y)) = h[(x, Sx)(y, T y)] 2
and by condition (2),
1
(Sx, T y) h[(x, Sx)(y, T y)] 2 = ((x, Sx), (y, T y), (x, y)).
Hence by using previous Theorem 1.1 we get the required conclusion.
We recall from [4] that in a metric space X, the mapping f : X X and
T : X CB(X) are said to be weakly commuting if for each x in X and
f T x in CB(X), H(f T x, T f x) D(T x, f y). It is known that every pair of
commuting mappings is weakly commuting but the converse is not true. Next
result extends the Theorem 2 of [8] to a general class of multi-valued mappings
called A -maps.
Theorem 1.4. Let f : X X and S, T : X CB(X) be mappings such
that
(i) f, S or T are continuous;
(ii) Sx f (X) and T x f (X);
(iii) f weakly commutes with S and T ;
(iv) (Sx, T y) ((Sx, x), (T y, y), (x, y)) for all x, y in X and for some
A .
Then f , S and T have common fixed point.
Proof. Let xo X by (ii) we may define a sequence (xn ) in X as
x2n1 = f x2n2 X2n1 = Sx2n2
and x2n = f x2n1 nX2n = T x2n1 where n = 1, 2, 3, .
10 M.Akram A.A.Siddiqui and A.A.Zafar

d(f x2n , f x2n + 1) (X2n+1 , X2n+2 ) = (Sx2n , T x2n+1 )

((Sx2n , x2n ), (T x2n+1 , x2n+1 ), (x2n , x2n+1 ))
((X2n+1 , X2n ), (X2n+2 , X2n+1 ), (X2n , X2n+1 ))
k(X2n+1 , X2n )by property (iii) of .
Hence (X2n+1 , X2n+2 ) k(X2n+1 , X2n ). Repeating this step, we get
d(f x2n , f x2n+1 ) (X2n+1 , X2n+2 ) k 2n+1 (Xo , X1 )n N . Which in fact
gives that d(f xn1 , f xn ) (Xn , Xn+1 ) k n (Xo , X1 )n N . By using the
same argument as in Theorem 1.1, (f xn ) is a Cauchy sequence in X and hence
f xn x in X. Then
D(x2n , Sx) (x2n , Sx) (T x2n1 , Sx)
((T x2n1 , x2n1 ), (Sx, x), (x2n1 , x))
((X2n , X2n1 ), (Sx, x), (x2n1 , x))
(k 2n1 (X0 , X1 ), (Sx, x), (x2n1 , x))
and so by letting n , we get that

D(x, Sx) (x, Sx) (0, (Sx, x), (x, x))

(0, (Sx, x), 0) k(0) by (iii) of ,
hence D(x, Sx) 0. This implies D(x, Sx) = 0. Which gives x Sx . But
Sx = Sx as Sx CB(X). Therefore x Sx. Similarly we can show that
x T x. Thus x is a common fixed point of S and T .
Suppose that f and S are continuous. Since f weakly commutes with S, so
we have
d(x, f x) d(x, f f x2n ) + d(f f x2n , f x) qH(Sx, f Sx2n ) + d(f f x2n , f x), q > 1
q{H(Sx, Sf x2n ) + H(Sf x2n , f Sx2n )} + d(f f x2n , f x)
q{H(Sx, Sf x2n ) + D(f x2n , Sx2n )} + d(f f x2n+1 , f x)
q{H(Sx, Sf x2n ) + 0} + d(f f x2n+1 , f x)
when n we have d(x, f x) qH(Sx, Sx) + d(f x, f x) = 0 This gives
d(x, f x) = 0 or x = f x. Thus x = f x Sx T x.
Theorem 1.5. Let f : X X and S, T : X CB(X) be mappings satisfying
conditions (i) to (ii) of theorem 1.4 and with the condition
(3) (Sx, T y) h((Sx, x) + (T y, y))
 A fixed point theorem 11

1
for all x, y in X and 0 < h < 2 . Then f, S and T have common fixed
point.
Proof. Defining as in Theorem 1.2, (iv) of Theorem 1.4 similarly obtained
by (3). Thus we have conclusion.
Theorem 1.6. ( see Theorem 2 of [8]) Let f : X X and S, T : X CB(X)
be mappings satisfying conditions (i) to (iii) of Theorem 1.4 and with the
condition
1
(4) (Sx, T y) ((Sx, x)(T y, y)) 2
for all x, y in X. Then f, S and T have common fixed point.
1
Proof. Let us define a functions : R+3 R by (u, v, w) = (uv) 2 then
+

one can easily see that A . Now taking u = d(x, Sx), v = d(y, T y) and
w = d(x, y) condition (4) of this theorem becomes similar to corresponding
condition of above Theorem 1.4. Hence by using Theorem 1.4 we have the
conclusion.

References
[1] Beg, I. and Azam, A; Fixed point of asymptotically regular multi- valued mappings,
Austral. Math. Soc. 53Series(A) (1992), 313326
[2] K.J Chung,Some common fixed point theorem , Math Japonica , 23(4) (1978), 401408
[3] B.Fisher and K.Iseki,Fixed point for set valued mappings om complete and compact
metric spaces, Math Japonica , 28(5) (1983), 639646
[4] H. Kanecko,A common fixed point of weakly commuting multi-valued map pings, Math
Japonica, 33(5) (1988), 714744
[5] M.S Khan M.D Khan and I.Kubiaczyk,Some common fixed point theorems for multi-
valued mappings, Demonst, Math 17(4) (1984), 9971002
[6] M.S Khan and I.Kubiaczyk,Fixed point theorem for point to set maps, Math Japonica,
33(3) (1988), 409415
[7] K.L sing and J.H.M Whitefield,Fixed points of contractive type mult ivalued mappings,
Math Japonica, 27 (1982), 117124
[8] F.U.Rehman and B,Ahmad,Multi-valued mappings and fixed point theorems, Journal of
natural sciences and mathematics, 36(2) (1996), 19319 7

M.Akram
Department of Mathematics
G. C. University
Lahore-54000, Pakistan
E-mail : [email protected]
12 M.Akram A.A.Siddiqui and A.A.Zafar

Akhlaq A. Siddiqui
Department of Mathematics
G. C. University
Lahore-54000, Pakistan
E-mail : [email protected]

A.A.Zafar
Department of Mathematics
G. C. University
Lahore-54000, Pakistan