You are on page 1of 7

BULL. AUSTRAL. MATH. SOC. VOL. 18 ( 1978), 137-143.

54E40,

54E99

An analogue of Banach's contraction principle for 2-metric spaces

S.N. Lai and A.K. Singh


In this paper we establish a fixed point theorem for spaces. 2-metric

Some interesting particular cases of this theorem are

also obtained.

1 .
Just as a metric abstracts the properties of the length function, a 2-metric space has its topology given by a real function of point triples which abstracts the properties of the area function for euclidean triangles. 2-metric on numbers R Let X X be a set consisting of at least three points. p of X x X x X A

is a mapping

into the set of real

that satisfies the following conditions: if at least two of p(a, b, a) = 0 point a a, b, a are equal, then a # b , there exists a

(1.1)

and, for any

such that

p(a, b, a) t 0 ; for all a, b, a in

(1.2)

p(a, b, a) = p(>, a, a) = p(e, a, b) X ;

(1.3) p(a, b, a) < p(a, b, d) + p(a, d, a) + p(d, b, a)


a, b, a, d
The pair (X, p) (X, p)

for all

in X .
is called a 2-metric space. A sequence <x > m in and

is said to be a Cauchy sequence if a X . The sequence

p(x , x , a) * 0 <x >

as

n * for every point x X if

is said to converge to the a (. X . A 2-metric

lim p[x , x, a) = 0

for every

Received 9 December 1977. 137

138

S.N.

Lai

and A . K .

Singh

space {X, p) is said to be complete if every Cauchy sequence in it is convergent. Further the 2-metric space (X, p) is said to be bounded if there exists a constant K such that p(a, b, c) K for a l l a, b, a in X ([2], [5]).
2.

As is well-known, there are a large number of generalisations of Banach's contraction principle and analogous results in the literature (see for example [ 7 ] , [3], [6]). In this paper we establish the following analogue of Banach's contraction principle for 2-metric spaces. THEOREM. Let two self-maps (2.1) (X, p) be a complete 2-metric spaoe and $ X , ^^y), a) and < > t

on X such that for all

x, y, a in

p(* 1 (x), * 2 (t/), a) 5 ap{x, +


<Z 3 P(X,

^ ( x ) , a) + a2p[y,

$ 2 (y), a) + akp{y, are non-negative Then $

^ ( x ) , a) + a^pix, y, a) , numbers such that and $ 2 have a unique

where
1, X.

a , a , a , a, , and a

a. < 1 and [a-,-) (a - a , ) 2 0 . point.


3.

common fixed

Proof of the theorem. write

Let [x] denote the integral part of x and

Take any x_ X and define

For any non-negative integer

n we have

using ( l . l ) and condition (2.1) of the theorem.

As a 2 < 1 , in view of

Banach's

contraction

principle

139

(1.2) the above inequality gives > X 2n + 1' Similarly


(3

-2) For any a t X ,

^2n+l'

2 + 2'

> X2+2'
W+J X
+

)
x

P(X2+1'
X

2M+2>

))

2 W + 1' a)

+ a x

A 2n>

2n + 1' a ) '

and therefore, using (3.1), we get P( X 2 + 1'


X

2+2> a) ~ P K n '

2 + l' a)

Similarly, using (3.2), we get P( X 2n + 2'


X

2n + 3' a ) ~ B P^ X 2n + r

2n + 2' a )

With the help of the above two inequalities i t follows that P( x 2n + 1' x 2n + 2' a) ~ ( ^ ) ( 6 ) and P ( X 2 + 2 ' *sn+y Hence
(3 a [(2n+l)/2]

p ( x Q , x, a) ,

) ~ < 1+a > ( a6) [ ( & I + 2 ) / 2 ] p ( x Q , x ^ a) .

-3) Note that

Plxm' x m + l' a ) " ( l ^ ) ( B ) t m / 2 ] p ( x 0 , xv

a) .

(3.U) for m = 0, 1, 2, . . . .

p(* 0 , *

xj

=0 m =0 and m = 1 . Suppose now

This is true for m in

that i t holds for every

2 m 5 k-1 .

Using (1.3) we have


+

p(x 0 ,

xk) - p(x 0 , x x , a r ^ J + p{xQ, a r ^ , xfc)


/ 2 ]

p ^ , a:fe, x j

l x 0 , x x , x o )4 P (x o , x x , x j ] ,

140

S . N .Lai andA . K . Singh

and t h i s proves ( 3 . M . Since

P(V Xm+1>
i t follows that (3-5)

(^He) [ m / 2 ] p(x 0 , r x j .

p(xm> xm+1, xn) = 0 , m and n .

for a l l non-negative integers

Note t h a t for any a X and m < n we have


p

K,' V

fl

) - p(Xm' X m + 1' a )

?(V

m+l'

P( x m + 1'

and t h e r e f o r e

i n v i e w o f ( 3 . 5 ) a n d ( 3 . 3 ) we h a v e

v
- V 2 - a)
+

+ p ( v r

As

a6 < 1 , the right hand side of the above inequality tends to zero as <x > is a Cauchy sequence and it converges to some

m -> . Hence x X .

Since p[x, $ 1 ( x ) , a) (
P(X2M+2'
X>

V*^ (2' '


X

+ p X X

^ 2n+2' x ' a^

+ a p x

i(>

i ^ x ^ ' a3

' a)
$ (a;

' a^+p(x'

) ' a)}

+ a 5

taking the limit as

n -* , the above inequality gives

p(x, $ 1 ( x ) , a) < (a1-h21(]p(x, $ 1 ( x ) , a) ,

and therefore p(x, $ 1 (x), a) = 0 ,


for all a X . Axiom (l.l) now gives If y is a fixed point of $.(x) = x . Similarly < t > (x) =

$ , then using (2.1) we get x

p(x, y, a) = 0

for all a X . Hence

x = y . This proves that

is a unique fixed

Banach's c o n t r a c t i o n

principle

141

point of 4 >

and similarly i t i s a unique fixed point of < J >

as well.

Hence the theorem.

A point i s an unique fixed point of a map $ : X -* X i f and only i f i t i s an unique fixed point of any positive power of * . This observation leads us to the following: T H E O R E M A. Let (X, p) be a complete $ two self-maps p, q , on X such that for all 2-metric space and $ and x, y, a in X and positive

integers

, a} < axp\x, P(x), a] + a^y, $J(y), a


+ a3p\x, where a , a , a , a, j a. < l
3

$|(z/), a + a^ply, are nan-negative Then $

$^(x), a + a^x, constants such that

y, a) ,

and a

and (a -a ) [a -a ) 2 0 . point.

and $ have a unique and

common fixed

C O R O L L A R Y 1. Let {X, p) be a complete

2-metric

space and f.

( i = 1, 2, 3, . . . ) a family of mappings of X into itself. Suppose there <m.) and non-negative numbers exists a sequence of positive integers a , a , a , a, , a i, 3 > i * 3 , such that for all x, y, a in X and every pair

n.

m.

m.

>

[
where i
</.>
\r~A-

m.

-\

em.

x, //(/), a + fliPJ, / / ( * ) , a + a pU, y, a) , 5 Y. a. < 1 and [a -a ) [a -a/) 2 0 . Then the sequence of mappings
has a unique common fixed point.

It is interesting to note that the particular case a = a = a ,

142

S . N . Lai

and A . K .

Singh

= a, = 0 and

a_ = 3

of this result has been recently established [4] 2-metric space is bounded. / . and / . have t 3 are arbitrary, the

with the additional assumption that the Proof. Take any pair i t j .

Then, by Theorem A, Since i and j

an unique and common fixed point. corollary follows.

C O R O L L A R Y 2. Let (X, p) be a complete 2-metric space and *. and 9two self maps on X satisfying (a) the following conditions: a., a^ , and a such

there exist nan-negative constants that 2[a +a ) + a < 1 and

aJp\x,

*P*|(y), al+plj/, <^*|(a;), a | + a 3 p(x, j / , a) , p and

for all x, y, a in X and any positive integers


(b) Then $ $ and < > i commute. point.

and $ ? ?wue a unique common fixed

Proof. By our theorem of Section 2, the map Kvi has a unique fixed o w point say u N

for

and $

commute. Hence $ (u) i s a fixed point of $^$| and

so $.() = u . Similarly $ (w) = M . Observe that if a; is a common fixed point of $, and $ x = u . Hence the result. then x is a fixed point of 4>^$| and so

Banach's

contraction

principle

143

References
['] [2] [3] [4] [5] [6] A.T. Bharucha-Reid, "Fixed point theorems in probabilistic analysis",
Bull. Amer. Math. Soo. 82 ( 1 9 7 6 ) , 61*1-657.

Siegfried (Shier, "2-metrische Raume und ihre topologische Struktur", Math. Naohr. 26 ( 1 9 6 3 M ) , 115-lW. G.E. Hardy and T.D. Rogers, "A generalization of a fixed point theorem of Reich", Canad. Math. Bull. 16 (1973), 201-206. K. Iseki, P.L. Sharma and B.K. Sharma, "Contraction type mapping on 2-metric space", Math. Japon. 21 (1976), 67-70. Albert George White, J r . , "2-Banach spaces", Math. Naohr. 42 (1969), 1*3-60. Chi Song Wong, "Common fixed points for tvo mappings", Pacific Math. 48 (1973), 299-312. J.

Department of Mathematics, Banaras Hindu U n i v e r s i t y , Varanasi, India.