Open Access

PPF dependent common fixed point theorems for mappings in Banach spaces

Journal of Inequalities and Applications20132013:287

https://doi.org/10.1186/1029-242X-2013-287

Received: 19 October 2012

Accepted: 22 May 2013

Published: 6 June 2013

Abstract

We prove the existence of the PPF dependent common fixed point theorems and the PPF dependent coincidence points for a pair of mappings satisfying some contractive conditions in Banach spaces where the domains and ranges of the mappings are not the same. Our results extend and generalize the results in the literature.

MSC:54H25, 55M20.

Keywords

PPF dependent common fixed pointsPPF dependent coincidence pointsRazumikhin classesgeneralized ψ-contractions

1 Introduction

The fixed point theory in Banach spaces plays an important role and is useful in mathematics. It can be applied for solving various problems, for instance, variational inequalities, optimization and approximation theory. The common fixed point theorems for mappings satisfying certain contractive conditions have been continually studied for a decade (see [17] and references contained therein). Bernfeld et al. [8] proved the existence of PPF (past, present and future) dependent fixed points in the Razumikhin class of functions for mappings that have different domains and ranges. Recently, Dhage [9] extended the existence of PPF dependent fixed points to PPF common dependent fixed points for mappings satisfying the weaker contractive conditions. In this paper, the PPF dependent fixed point theorems and the PPF dependent coincidence points for a pair of mappings are proven in terms of more general contractive conditions in Banach spaces.

Suppose that E is a Banach space with the norm E and I is a closed interval [ a , b ] in . Let E 0 = C ( I , E ) be the set of all continuous E-valued functions on I equipped with the supremum norm E 0 defined by
ϕ E 0 = sup t I ϕ ( t ) E
(1.1)
for all ϕ E 0 . For a fixed element c I , the Razumikhin class of functions in E 0 is defined by
R c = { ϕ E 0 : ϕ E 0 = ϕ ( c ) E } .
(1.2)

Recall that a point ϕ E 0 is said to be a PPF dependent fixed point or a fixed point with PPF dependence of T : E 0 E if T ϕ = ϕ ( c ) for some c I .

Definition 1.1 Let A be a subset of E. Then:
  1. (i)

    A is said to be topologically closed with respect to the norm topology if for each sequence { x n } in A with x n x as n implies x A .

     
  2. (ii)

    A is said to be algebraically closed with respect to the difference if x y A for all x , y A .

     

Definition 1.2 Let S , T : E 0 E be two mappings. A point ϕ E 0 is said to be a PPF dependent common fixed point or a common fixed point with PPF dependence of S and T if S ϕ = ϕ ( c ) = T ϕ for some c I .

Recently, Dhage [9] proved the existence of PPF common fixed points for mappings satisfying the condition of Cirić type generalized contraction in a Razumikhin class.

Definition 1.3 Two mappings S , T : E 0 E are said to satisfy the condition of Cirić type generalized contraction if there exists a real number λ [ 0 , 1 ) satisfying
S ϕ T α λ max { ϕ α E 0 , ϕ ( c ) S ϕ E , α ( c ) T α E , 1 2 [ ϕ ( c ) T α E + α ( c ) S ϕ E ] }
(1.3)

for all ϕ , α E 0 and for some c I .

Theorem 1.4 (Dhage [9])

Suppose that S , T : E 0 E satisfy the condition of Cirić type generalized contraction. Assume that R c is topologically closed with respect to the norm topology and is algebraically closed with respect to the difference, then S and T have a unique PPF dependent common fixed point in R c .

Definition 1.5 Let A : E 0 E and S : E 0 E 0 . A point ϕ E 0 is said to be a PPF dependent coincidence point or a coincident point with PPF dependence of A and S if A ϕ = S ϕ ( c ) for some c I .

Dhage [9] also assured the existence of PPF dependent coincidence points for mappings satisfying the condition of Cirić type generalized contraction (C) in a Razumikhin class.

Definition 1.6 A : E 0 E and S : E 0 E 0 are said to satisfy the condition of Cirić type generalized contraction (C) if there exists a real number λ [ 0 , 1 ) satisfying
A ϕ A α E λ max { S ϕ S α E 0 , S ϕ ( c ) A ϕ E , S α ( c ) A α E , 1 2 [ S ϕ ( c ) A α E + S α ( c ) A ϕ E ] }
(1.4)

for all ϕ , α E 0 and for some c I .

Theorem 1.7 (Dhage [9])

Let A : E 0 E and S : E 0 E 0 be two mappings satisfying the condition of Cirić type generalized contraction (C). Suppose that
  1. (a)

    A ( E 0 ) S ( E 0 ) ( c ) ;

     
  2. (b)

    S ( E 0 ) is complete;

     
  3. (c)

    S is continuous.

     

If R c is topologically closed with respect to the norm topology and is algebraically closed with respect to the difference, then A and S have a PPF dependent coincidence point in R c .

In this work, we prove the existence of PPF dependent common fixed point theorems for mappings satisfying the contractive condition which is weaker than the condition of Cirić type generalized contraction. Furthermore, we also prove the PPF dependent coincidence points for mappings satisfying the weaker contractive condition mentioned in [9] without being topologically closed with respect to the norm topology of a Razumikhin class. Our results extend Theorem 1.4 and partially generalize Theorem 1.7.

2 PPF dependent common fixed points

Let Ψ be the set of all functions ψ where ψ : [ 0 , + ) [ 0 , + ) is a continuous nondecreasing function with ψ ( t ) < t for all t ( 0 , + ) and ψ ( 0 ) = 0 . If ψ Ψ , then ψ is called a Ψ-map. We now introduce the definition of the condition of Cirić type generalized ψ-contraction and prove our first result.

Definition 2.1 Let S , T : E 0 E . We say that S and T satisfy the condition of Cirić type generalized ψ-contraction if
S ϕ T α E ψ ( max { ϕ α E 0 , ϕ ( c ) S ϕ E , α ( c ) T α E , 1 2 [ ϕ ( c ) T α E + α ( c ) S ϕ E ] } )
(2.1)

for all ϕ , α E 0 and for some c I .

Theorem 2.2 Suppose that S , T : E 0 E satisfy the condition of Cirić type generalized ψ-contraction. Assume that R c is topologically closed with respect to norm topology and is algebraically closed with respect to the difference, then S and T have a unique PPF dependent common fixed point in R c .

Proof Let ϕ 0 R c . Since S ϕ 0 E , there exists x 1 E such that S ϕ 0 = x 1 . Choose ϕ 1 R c such that
x 1 = ϕ 1 ( c ) and ϕ 1 ϕ 0 E 0 = ϕ 1 ( c ) ϕ 0 ( c ) E .
Since ϕ 1 R c and by assumption, we have T ϕ 1 E . This implies that there exists x 2 E such that T ϕ 1 = x 2 . Therefore, we can choose ϕ 2 R c such that
x 2 = ϕ 2 ( c ) and ϕ 2 ϕ 1 E 0 = ϕ 2 ( c ) ϕ 1 ( c ) E .
By continuing the process as before, we can construct the sequence { ϕ n } such that
S ϕ 2 n = ϕ 2 n + 1 ( c ) , T ϕ 2 n + 1 = ϕ 2 n + 2 ( c )
and
ϕ n ϕ n + 1 E 0 = ϕ n ( c ) ϕ n + 1 ( c ) E
for all n N { 0 } . We will show that { ϕ n } is a Cauchy sequence in E 0 . Assume that ϕ N 1 = ϕ N for some N N . If N is even, then we have N = 2 m for some m N . Therefore
ϕ 2 m ϕ 2 m + 1 E 0 = ϕ 2 m ( c ) ϕ 2 m + 1 ( c ) E = S ϕ 2 m T ϕ 2 m 1 E ψ ( max { ϕ 2 m ϕ 2 m 1 E 0 , ϕ 2 m ( c ) S ϕ 2 m E , ϕ 2 m 1 ( c ) T ϕ 2 m 1 E , 1 2 [ ϕ 2 m ( c ) T ϕ 2 m 1 E + ϕ 2 m 1 ( c ) S ϕ 2 m E ] } ) ψ ( max { ϕ 2 m ϕ 2 m 1 E 0 , ϕ 2 m ( c ) ϕ 2 m + 1 ( c ) E , ϕ 2 m 1 ( c ) ϕ 2 m ( c ) E , 1 2 [ ϕ 2 m ( c ) ϕ 2 m ( c ) E + ϕ 2 m 1 ( c ) ϕ 2 m + 1 ( c ) E ] } ) ψ ( max { ϕ 2 m ϕ 2 m 1 E 0 , ϕ 2 m ϕ 2 m + 1 E 0 , 1 2 ϕ 2 m 1 ϕ 2 m + 1 E 0 } ) ψ ( max { ϕ 2 m ϕ 2 m 1 E 0 , ϕ 2 m ϕ 2 m + 1 E 0 , 1 2 ϕ 2 m 1 ϕ 2 m E 0 + ϕ 2 m ϕ 2 m + 1 E 0 } ) ψ ( max { ϕ 2 m ϕ 2 m 1 E 0 , ϕ 2 m ϕ 2 m + 1 E 0 } ) ψ ( max { ϕ 2 m ϕ 2 m + 1 E 0 } ) .
This implies that ϕ 2 m ϕ 2 m + 1 E 0 = 0 and so ϕ 2 m = ϕ 2 m + 1 . Similarly, we can prove that ϕ 2 m + 1 = ϕ 2 m + 2 . Therefore ϕ N = ϕ N + 1 . By mathematical induction, we can conclude that ϕ N 1 = ϕ N + k for all k 0 . If N is odd, then by the same argument we also obtain that ϕ N 1 = ϕ N + k for all k 0 . Therefore { ϕ n } is a constant sequence for all n N 1 . This implies that { ϕ n } is a Cauchy sequence in E 0 . Suppose that ϕ n 1 ϕ n for all n N . For each n N , we obtain that
ϕ 2 n ϕ 2 n + 1 E 0 = ϕ 2 n ( c ) ϕ 2 n + 1 ( c ) E = S ϕ 2 n T ϕ 2 n 1 E ψ ( max { ϕ 2 n ϕ 2 n 1 E 0 , ϕ 2 n ( c ) S ϕ 2 n E , ϕ 2 n 1 ( c ) T ϕ 2 n 1 E , 1 2 [ ϕ 2 n ( c ) T ϕ 2 n 1 E + ϕ 2 n 1 ( c ) S ϕ 2 n E ] } ) ψ ( max { ϕ 2 n ϕ 2 n 1 E 0 , ϕ 2 n ( c ) ϕ 2 n + 1 ( c ) E , ϕ 2 n 1 ( c ) ϕ 2 n ( c ) E , 1 2 [ ϕ 2 n ( c ) ϕ 2 n ( c ) E + ϕ 2 n 1 ( c ) ϕ 2 n + 1 ( c ) E ] } ) ψ ( max { ϕ 2 n ϕ 2 n 1 E 0 , ϕ 2 n ϕ 2 n + 1 E 0 , 1 2 ϕ 2 n 1 ϕ 2 n + 1 E 0 } ) ψ ( max { ϕ 2 n ϕ 2 n 1 E 0 , ϕ 2 n ϕ 2 n + 1 E 0 , 1 2 ϕ 2 n 1 ϕ 2 n E 0 + ϕ 2 n ϕ 2 n + 1 E 0 } ) ψ ( max { ϕ 2 n ϕ 2 n 1 E 0 , ϕ 2 n ϕ 2 n + 1 E 0 } ) .
If max { ϕ 2 n ϕ 2 n 1 E 0 , ϕ 2 n ϕ 2 n + 1 E 0 } = ϕ 2 n ϕ 2 n + 1 E 0 , then
ϕ 2 n ϕ 2 n + 1 E 0 ψ ( ϕ 2 n ϕ 2 n + 1 E 0 ) < ϕ 2 n ϕ 2 n + 1 E 0 .
This leads to a contradiction. Therefore
ϕ 2 n ϕ 2 n + 1 E 0 ψ ( ϕ 2 n ϕ 2 n 1 E 0 ) < ϕ 2 n ϕ 2 n 1 E 0 .
Thus ϕ 2 n ϕ 2 n + 1 E 0 ϕ 2 n ϕ 2 n 1 E 0 . Similarly, we can prove that
ϕ 2 n + 1 ϕ 2 n + 2 E 0 ϕ 2 n ϕ 2 n + 1 E 0 .
It follows that ϕ n ϕ n + 1 E 0 ϕ n 1 ϕ n E 0 for all n N . Since the sequence { ϕ n ϕ n + 1 E 0 } is a nonincreasing sequence of nonnegative real numbers, we obtain that it is a convergent sequence. Suppose that
lim n ϕ n ϕ n + 1 E 0 = α
for some nonnegative real number α. We will prove that α = 0 . Suppose that α > 0 . Since
ϕ 2 n ϕ 2 n + 1 E 0 ψ ( ϕ 2 n ϕ 2 n 1 E 0 )
for all n N and the continuity of ψ, we have
α ψ ( α ) < α ,
which leads to a contradiction. This implies that α = 0 . We next prove that the sequence { ϕ n } is a Cauchy sequence in E 0 . It suffices to prove that the sequence { ϕ 2 n } is a Cauchy sequence in E 0 . Assume that { ϕ 2 n } is not a Cauchy sequence. It follows that there exist ε > 0 and two sequences of even positive integers { 2 m k } and { 2 n k } satisfying 2 m k > 2 n k > k for each k N and
ϕ 2 m k ϕ 2 n k E 0 ε .
(2.2)
Let { 2 m k } be the sequence of the least positive integers exceeding { 2 n k } which satisfies (2.2) and
ϕ 2 m k 2 ϕ 2 n k E 0 < ε .
(2.3)
We will prove that lim k ϕ 2 m k ϕ 2 n k E 0 = ε . Since ϕ 2 m k ϕ 2 n k E 0 ε for all k N , we have
lim k ϕ 2 m k ϕ 2 n k E 0 ε .
For each k N , we obtain that
ϕ 2 m k ϕ 2 n k E 0 ϕ 2 m k ϕ 2 m k 1 E 0 + ϕ 2 m k 1 ϕ 2 m k 2 E 0 + ϕ 2 m k 2 ϕ 2 n k E 0 ϕ 2 m k ϕ 2 m k 1 E 0 + ϕ 2 m k 1 ϕ 2 m k 2 E 0 + ε .
This implies that lim k ϕ 2 m k ϕ 2 n k E 0 ε . Therefore
lim k ϕ 2 m k ϕ 2 n k E 0 = ε .
Similarly, we can prove that
lim k ϕ 2 m k + 1 ϕ 2 n k E 0 = ε , lim k ϕ 2 m k ϕ 2 n k 1 E 0 = ε
and
lim k ϕ 2 m k + 1 ϕ 2 n k 1 E 0 = ε .
For each k N , we obtain that
ϕ 2 n k ϕ 2 m k + 1 E 0 ϕ 2 n k ( c ) ϕ 2 m k + 1 ( c ) E S ϕ 2 m k T ϕ 2 n k 1 E ψ ( max { ϕ 2 m k ϕ 2 n k 1 E 0 , ϕ 2 m k ( c ) S ϕ 2 m k E , ϕ 2 n k 1 ( c ) T ϕ 2 n k 1 E , 1 2 [ ϕ 2 m k ( c ) T ϕ 2 n k 1 E + ϕ 2 n k 1 ( c ) S ϕ 2 n k E ] } ) ψ ( max { ϕ 2 m k ϕ 2 n k 1 E 0 , ϕ 2 m k ( c ) ϕ 2 m k + 1 ( c ) E , ϕ 2 n k 1 ( c ) ϕ 2 n k ( c ) E , 1 2 [ ϕ 2 m k ( c ) ϕ 2 n k ( c ) E + ϕ 2 n k 1 ( c ) ϕ 2 n k + 1 ( c ) E ] } ) ψ ( max { ϕ 2 m k ϕ 2 n k 1 E 0 , ϕ 2 m k ϕ 2 m k + 1 E 0 , ϕ 2 n k 1 ϕ 2 n k E 0 , 1 2 [ ϕ 2 m k ϕ 2 n k E 0 + ϕ 2 n k 1 ϕ 2 n k + 1 E 0 ] } ) ψ ( max { ϕ 2 m k ϕ 2 n k 1 E 0 , ϕ 2 m k ϕ 2 m k + 1 E 0 , ϕ 2 n k 1 ϕ 2 n k E 0 , 1 2 [ ϕ 2 m k ϕ 2 n k E 0 + ϕ 2 n k 1 ϕ 2 n k E 0 + ϕ 2 n k ϕ 2 n k + 1 E 0 ] } ) .
By taking the limit of both sides, we have
ε ψ ( ε ) < ε ,
which leads to a contradiction. It follows that the sequence { ϕ 2 n } is a Cauchy sequence and so { ϕ n } is a Cauchy sequence. By the completeness of E 0 , we have { ϕ n } is a convergent sequence. Suppose that lim n ϕ n = ϕ for some ϕ E 0 . Since R c is algebraically closed with respect to the norm topology, we have ϕ R c . Moreover, we also obtain that
lim n ϕ 2 n + 1 = ϕ = lim n ϕ 2 n + 2 .
We will prove that ϕ is a PPF dependent fixed point of S. By using (2.1), we obtain that
S ϕ ϕ ( c ) E S ϕ ϕ 2 n + 2 ( c ) E + ϕ 2 n + 2 ( c ) ϕ ( c ) E S ϕ T ϕ 2 n + 1 E + ϕ 2 n + 2 ϕ E 0 ψ ( max { ϕ ϕ 2 n + 1 E 0 , ϕ ( c ) S ϕ E , ϕ 2 n + 1 ( c ) T ϕ 2 n + 1 E , 1 2 [ ϕ ( c ) T ϕ 2 n + 1 E + ϕ 2 n + 1 ( c ) S ϕ E ] } ) + ϕ 2 n + 2 ϕ E 0 ψ ( max { ϕ ϕ 2 n + 1 E 0 , ϕ ( c ) S ϕ E , ϕ 2 n + 1 ( c ) ϕ 2 n + 2 ( c ) E , 1 2 [ ϕ ( c ) ϕ 2 n + 2 ( c ) E + ϕ 2 n + 1 ( c ) S ϕ E ] } ) + ϕ 2 n + 2 ϕ E 0 ψ ( max { ϕ ϕ 2 n + 1 E 0 , ϕ ( c ) S ϕ E , ϕ 2 n + 1 ϕ 2 n + 2 E 0 , 1 2 [ ϕ ϕ 2 n + 2 E 0 + ϕ 2 n + 1 ( c ) S ϕ E ] } ) + ϕ 2 n + 2 ϕ E 0 .
Letting n , we obtain that S ϕ ϕ ( c ) E = 0 . Therefore S ϕ = ϕ ( c ) . Similarly, we can prove that T ϕ = ϕ ( c ) . This implies that ϕ is a PPF dependent common fixed point of S and T. We finally prove the uniqueness of the PPF dependent fixed point of S and T in  R c . Let α R c be any PPF dependent common fixed point of S and T. Therefore
ϕ α E 0 = ϕ ( c ) α ( c ) E S ϕ T α E ψ ( max { ϕ α E 0 , ϕ ( c ) S ϕ E , α ( c ) T α E , 1 2 [ ϕ ( c ) T α E + α ( c ) S ϕ E ] } ) ψ ( max { ϕ α E 0 , 1 2 [ ϕ ( c ) α ( c ) E + α ( c ) ϕ ( c ) E ] } ) ψ ( max { ϕ α E 0 , ϕ α E 0 } ) = ψ ( ϕ α E 0 ) .

It follows that ϕ = α . Hence S and T have a unique PPF dependent common fixed point in  R c . □

Remark 2.3 From the proof of Theorem 2.2, we obtain the following observations:
  1. (1)
    We assume that the Razumikhin class R c is algebraically closed with respect to difference, that is, ϕ α R c for all ϕ , α R c , in order to construct the sequence { ϕ n } satisfying
    ϕ α E 0 = ϕ ( c ) α ( c ) E .
     
  2. (2)

    If the Razumikhin class R c is not assumed to be topologically closed, then the limit of the sequence { ϕ n } may be outside of R c .

     

By applying Theorem 2.2, we obtain the following corollary which is Theorem 3.3 in [9].

Corollary 2.4 Suppose that S , T : E 0 E satisfy the condition of Cirić type generalized contraction. Assume that R c is topologically and algebraically closed with respect to the difference, then S and T have a unique PPF dependent fixed point in R c .

Proof Define a function ψ : [ 0 , + ) [ 0 , + ) by ψ ( t ) = λ t for all t [ 0 , + ) . Therefore ψ is a continuous nondecreasing function and
ψ ( t ) < t for all  t ( 0 , + )  and  ψ ( 0 ) = 0 .

This implies that all assumptions in Theorem 2.2 are satisfied. Hence the proof is complete. □

3 PPF dependent coincidence points

Definition 3.1 Let A : E 0 E and S : E 0 E 0 . A point ϕ E 0 is said to be a PPF dependent coincidence point or a coincidence point with PPF dependence of A and S if A ϕ = S ϕ ( c ) for some c I .

Definition 3.2 Let A : E 0 E and S : E 0 E 0 . We say that A and S satisfy the condition of Cirić type generalized contraction (C) if there exists a real number λ [ 0 , 1 ) satisfying
A ϕ A α E λ max { S ϕ S α E 0 , S ϕ ( c ) A ϕ E , S α ( c ) A α E , 1 2 [ S ϕ ( c ) A α E + S α ( c ) A ϕ E ] }
(3.1)

for all ϕ , α E 0 and for some c I .

We introduce the following contractive condition which is weaker than the condition of Cirić type generalized contraction (C).

Definition 3.3 Let A : E 0 E and S : E 0 E 0 . We say that A and S satisfy the condition of Cirić type generalized ψ-contraction (C) if
A ϕ A α E ψ ( max { S ϕ S α E 0 , S ϕ ( c ) A ϕ E , S α ( c ) A α E , 1 2 [ S ϕ ( c ) A α E + S α ( c ) A ϕ E ] } )
(3.2)

for all ϕ , α E 0 and for some c I .

We now prove the existence of PPF dependent coincidence points for mappings satisfying the weaker contractive condition assumed in [9] without being topologically closed of the Razumikhin class.

Theorem 3.4 Let A : E 0 E and S : E 0 E 0 be two mappings satisfying the condition of Cirić type generalized ψ-contraction (C). Suppose that
  1. (a)

    A ( E 0 ) S ( E 0 ) ( c ) ;

     
  2. (b)

    S ( R c ) is complete;

     
  3. (c)

    S is continuous.

     

If R c is algebraically closed with respect to the difference, then A and S have a PPF dependent coincidence point in R c .

Proof Let ϕ 0 R c . Since A ϕ 0 E , there exists x 1 E such that A ϕ 0 = x 1 . Because A ( E 0 ) S ( E 0 ) ( c ) , we can choose ϕ 1 R c such that
x 1 = S ϕ 1 ( c ) = α 1 ( c ) and α 1 α 0 E 0 = α 1 ( c ) α 0 ( c ) E .
Since A ϕ 1 E 0 and by assumption, we have A ϕ 1 = x 2 for some x 2 E . Because A ( E 0 ) S ( E 0 ) ( c ) , we can choose ϕ 2 R c such that
x 2 = S ϕ 2 ( c ) = α 2 ( c ) and α 2 α 1 E 0 = α 2 ( c ) α 1 ( c ) E .
By continuing the process as before, we can construct the sequence { α n } such that
A ϕ n = S ϕ n + 1 ( c ) , S ϕ n + 1 = α n + 1
and
α n α n + 1 E 0 = α n ( c ) α n + 1 ( c ) E
for all n N { 0 } . We will show that { α n } is a Cauchy sequence in E 0 . If α N = α N + 1 for some N N , then we have
α N + 1 α N + 2 E 0 = α N + 1 ( c ) α N + 2 ( c ) E = A ϕ N A ϕ N + 1 E ψ ( max { S ϕ N S ϕ N + 1 E 0 , S ϕ N ( c ) A ϕ N E , S ϕ N + 1 ( c ) A ϕ N + 1 E , 1 2 [ S ϕ N ( c ) A ϕ N + 1 E + S ϕ N + 1 ( c ) A ϕ N E ] } ) ψ ( max { α N α N + 1 E 0 , α N ( c ) α N + 1 ( c ) E , α N + 1 ( c ) α N + 2 ( c ) E , 1 2 [ α N ( c ) α N + 2 ( c ) E + α N + 1 ( c ) α N + 1 ( c ) E ] } ) ψ ( max { α N α N + 1 E 0 , α N α N + 1 E 0 , α N + 1 α N + 2 E 0 , 1 2 α N α N + 2 E 0 } ) ψ ( max { α N α N + 1 E 0 , α N + 1 α N + 2 E 0 , 1 2 [ α N α N + 2 E 0 ] } ) ψ ( max { α N α N + 1 E 0 , α N + 1 α N + 2 E 0 , 1 2 [ α N α N + 1 E 0 + α N + 1 α N + 2 E 0 ] } ) ψ ( max { α N α N + 1 E 0 , α N + 1 α N + 2 E 0 } ) ψ ( α N + 1 α N + 2 E 0 ) .
Therefore α N + 1 = α N + 2 . By mathematical induction, we obtain that
α N = α N + k for all  k N .
This implies that { α n } is a constant sequence for n N . Thus { α n } is a Cauchy sequence in E 0 . Suppose that α n α n + 1 for all n N . For each n N , we have
α n + 1 α n + 2 E 0 = α n + 1 ( c ) α n + 2 ( c ) E = A ϕ n A ϕ n + 1 E ψ ( max { S ϕ n S ϕ n + 1 E 0 , S ϕ n ( c ) A ϕ n E , S ϕ n + 1 ( c ) A ϕ n + 1 E , 1 2 [ S ϕ n ( c ) A ϕ n + 1 E + S ϕ n + 1 ( c ) A ϕ n E ] } ) ψ ( max { α n α n + 1 E 0 , α n ( c ) α n + 1 ( c ) E , α n + 1 ( c ) α n + 2 ( c ) E , 1 2 [ α n ( c ) α n + 2 ( c ) E + α n + 1 ( c ) α n + 1 ( c ) E ] } ) ψ ( max { α n α n + 1 E 0 , α n α n + 1 E 0 , α n + 1 α n + 2 E 0 , 1 2 α n α n + 2 E 0 } ) ψ ( max { α n α n + 1 E 0 , α n + 1 α n + 2 E 0 , 1 2 [ α n α n + 2 E 0 ] } ) ψ ( max { α n α n + 1 E 0 , α n + 1 α n + 2 E 0 , 1 2 [ α n α n + 1 E 0 + α n + 1 α n + 2 E 0 ] } ) ψ ( max { α n α n + 1 E 0 , α n + 1 α n + 2 E 0 } ) .
If max { α n α n + 1 E 0 , α n + 1 α n + 2 E 0 } = α n + 1 α n + 2 E 0 , then
α n + 1 α n + 2 E 0 ψ ( α n + 1 α n + 2 E 0 ) < α n + 1 α n + 2 E 0 .
This leads to a contradiction. Therefore
α n + 1 α n + 2 E 0 ψ ( α n α n + 1 E 0 ) < α n α n + 1 E 0 .
It follows that α n + 1 α n + 2 E 0 α n α n + 1 E 0 for all n N . Since the sequence { α n α n + 1 E 0 } is a nonincreasing sequence of real numbers, we obtain that it is a convergent sequence. Suppose that
lim n α n α n + 1 E 0 = α
for some nonnegative real number α. We will prove that α = 0 . Assume that α > 0 . Since
α n + 1 α n + 2 E 0 ψ ( α n α n + 1 E 0 )
for all n N and the continuity of ψ, we have
α ψ ( α ) < α ,
which leads to a contradiction. This implies that α = 0 . We will prove that { α n } is a Cauchy sequence in E 0 . It suffices to prove that the sequence { α 2 n } is a Cauchy sequence in E 0 . Assume that { α 2 n } is not a Cauchy. It follows that there exist ε > 0 and two sequences of even positive integers { 2 m k } and { 2 n k } satisfying 2 m k > 2 n k > k for each k N and
α 2 m k α 2 n k E 0 ε .
(3.3)
Let { 2 m k } be the sequence of the least positive integers exceeding { 2 n k } which satisfies (3.3) and
α 2 m k 2 α 2 n k E 0 < ε .
(3.4)
We will prove that lim k α 2 m k α 2 n k E 0 = ε . Since α 2 m k α 2 n k E 0 ε for all k N , we have
lim k α 2 m k α 2 n k E 0 ε .
For each k N , we obtain that
α 2 m k α 2 n k E 0 α 2 m k α 2 m k 1 E 0 + α 2 m k 1 α 2 m k 2 E 0 + α 2 m k 2 α 2 n k E 0 α 2 m k α 2 m k 1 E 0 + α 2 m k 1 α 2 m k 2 E 0 + ε .
This implies that lim k α 2 m k α 2 n k E 0 ε . Therefore
lim k α 2 m k α 2 n k E 0 = ε .
Similarly, we can prove that
lim k α 2 m k + 1 α 2 n k E 0 = ε , lim k α 2 m k α 2 n k 1 E 0 = ε
and
lim k α 2 m k + 1 α 2 n k 1 E 0 = ε .
Since
α 2 n k α 2 m k + 1 E 0 α 2 n k ( c ) α 2 m k + 1 ( c ) E = A ϕ 2 n k 1 A ϕ 2 m k E ψ ( max { S ϕ 2 n k 1 S ϕ 2 m k E 0 , S ϕ 2 n k 1 ( c ) A ϕ 2 n k 1 E , S ϕ 2 m k ( c ) A ϕ 2 m k E , 1 2 [ S ϕ 2 n k 1 ( c ) A ϕ 2 m k E + S ϕ 2 m k ( c ) A ϕ 2 n k 1 E ] } ) ψ ( max { α 2 n k 1 α 2 m k E 0 , α 2 n k 1 ( c ) α 2 n k ( c ) E , α 2 m k ( c ) α 2 m k + 1 ( c ) E , 1 2 [ α 2 n k 1 ( c ) α 2 m k + 1 ( c ) E + α 2 m k ( c ) α 2 n k ( c ) E ] } ) ψ ( max { α 2 n k 1 α 2 m k E 0 , α 2 n k 1 α 2 n k E 0 , α 2 m k α 2 m k + 1 E 0 , 1 2 α 2 n k 1 α 2 m k + 1 E 0 + α 2 m k α 2 n k E 0 } ) ,
by taking the limit of both sides, we have
ε ψ ( ε ) < ε ,
which leads to a contradiction. It follows that the sequence { α 2 n } is a Cauchy sequence and so { α n } is a Cauchy sequence in E 0 . Therefore { S ϕ n } is a Cauchy sequence in S ( R c ) . By the completeness of S ( R c ) , we have { S ϕ n } is a convergent sequence. Suppose that lim n S ϕ n = ϕ for some ϕ S ( R c ) . Therefore ϕ = S ϕ for some ϕ R c . Moreover, we also have
lim n A ϕ n = lim n S ϕ n + 1 ( c ) = S ϕ ( c ) .
We will prove that ϕ is a PPF dependent coincidence fixed point of A and S. By using (3.2), we obtain that
A ϕ S ϕ ( c ) E A ϕ A ϕ n E + A ϕ n S ϕ ( c ) E A ϕ A ϕ n E + S ϕ n + 1 ( c ) S ϕ ( c ) E ψ ( max { S ϕ S ϕ n E 0 , S ϕ ( c ) A ϕ E , S ϕ n ( c ) A ϕ n E , 1 2 [ S ϕ ( c ) A ϕ n E + S ϕ n ( c ) A ϕ E ] } ) + S ϕ n + 1 ( c ) S ϕ ( c ) E .

By taking n , we obtain that A ϕ = S ϕ ( c ) . Hence ϕ is a PPF dependent coincidence point of A and S. □

By applying Theorem 3.4, we obtain the following corollary.

Corollary 3.5 Let A : E 0 E and S : E 0 E 0 be two mappings satisfying the condition of Cirić type generalized contraction (C). Suppose that
  1. (a)

    A ( E 0 ) S ( E 0 ) ( c ) ;

     
  2. (b)

    S ( R c ) is complete;

     
  3. (c)

    S is continuous.

     

If R c is algebraically closed with respect to the difference, then A and S have a PPF dependent coincidence point in R c .

Proof Define a function ψ : [ 0 , + ) [ 0 , + ) by ψ ( t ) = λ t for all t [ 0 , + ) . Therefore A : E 0 E and S : E 0 E 0 satisfy the condition of Cirić type generalized ψ-contraction (C). This implies that all assumptions in Theorem 3.4 are now satisfied. Hence the proof is complete. □

Questions
  1. (i)

    Are the results in Theorem 2.2 and Theorem 3.4 still true when the norm closedness for R c is replaced by weak closedness or weak closedness (for dual Banach spaces)?

     
  2. (ii)

    Is there some way to improve the results to more than two mappings or a family of mappings as in the case of nonexpansive mappings (see, for example, [10] and references contained therein)?

     

Declarations

Acknowledgements

This research is supported by the Centre of Excellence in Mathematics, the Commission of Higher Education, Thailand. I would like to express my deep thanks to the referees for their suggestions and comments which led to the improvement of the manuscript.

Authors’ Affiliations

(1)
Department of Mathematics, Faculty of Science, Naresuan University, Phitsanulok, Thailand
(2)
Centre of Excellence in Mathematics, Bangkok, Thailand

References

  1. Beg I, Abbas M: Coincidence point and invariant approximation for mappings satisfying generalized weak contractive condition. Fixed Point Theory Appl. 2006. doi:10.1155/FPTA/2006/74503Google Scholar
  2. Jungck G: Compatible mappings and common fixed points. Int. J. Math. Math. Sci. 1968, 9: 771–779.MathSciNetView ArticleGoogle Scholar
  3. Jungck G: Common fixed points for commuting and compatible maps on compacta. Proc. Am. Math. Soc. 1988, 103: 977–983. 10.1090/S0002-9939-1988-0947693-2MathSciNetView ArticleGoogle Scholar
  4. Jungck G: Common fixed points for noncontinuous nonself maps on nonmetric spaces. Far East J. Math. Sci. 1996, 4: 199–215.MathSciNetGoogle Scholar
  5. Jungck G, Hussain N: Compatible maps and invariant approximations. J. Math. Model. Algorithms 2007, 325: 1003–1012.MathSciNetGoogle Scholar
  6. Pant RP: Common fixed points of noncommuting mappings. J. Math. Anal. Appl. 1994, 188: 436–440. 10.1006/jmaa.1994.1437MathSciNetView ArticleGoogle Scholar
  7. Sintunavarat W, Kumam P:Common fixed point theorems for generalized JH -operators classed and invariant approximations. J. Inequal. Appl. 2011. doi:10.1186/1029–242X-2011–67Google Scholar
  8. Bernfeld SR, Lakshmikantham V, Reddy YM: Fixed point theorems of operators with PPF dependence in Banach spaces. Appl. Anal. 1977, 6: 271–280. 10.1080/00036817708839165MathSciNetView ArticleGoogle Scholar
  9. Dhage BC: On some common fixed point theorems with PPF dependence in Banach spaces. J. Nonlinear Sci. Appl. 2012, 5: 220–232.MathSciNetGoogle Scholar
  10. Lau AT-M, Zhang Y: Fixed point properties for semigroups of nonlinear mappings and amenability. J. Funct. Anal. 2012, 263: 2949–2977. 10.1016/j.jfa.2012.07.013MathSciNetView ArticleGoogle Scholar

Copyright

© Kaewcharoen; licensee Springer 2013

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.