1 Introduction and preliminaries

The theoretical framework of metric fixed point theory has been an active research field over the last nine decades. Of course, the Banach contraction principle [1] is the first important result on fixed points for contractive-type mappings. So far, there have been a lot of fixed point results dealing with mappings satisfying various types of contractive inequalities. In particular, the concepts of K-contraction and C-contraction were introduced by Kannan [2], respectively, Chatterjea [3] as follows.

Definition 1 Let (X,d) be a metric space and f:XX.

  1. 1.

    ([2]) The mapping f is said to be a K-contraction if there exists α(0, 1 2 ) such that for all x,yX the following inequality holds:

    d(fx,fy)α ( d ( x , f x ) + d ( y , f y ) ) .
  2. 2.

    ([3]) The mapping f is said to be a C-contraction if there exists α(0, 1 2 ) such that for all x,yX the following inequality holds:

    d(fx,fy)α ( d ( x , f y ) + d ( y , f x ) ) .

In 1968, Kannan [2] proved that if (X,d) is a complete metric space, then every K-contraction on X has a unique fixed point. In 1972, Chatterjea [3] established a fixed point theorem for C-contractions.

Definition 2 Let (X,d) be a metric space, f:XX and φ: [ 0 , ) 2 [0,) be a continuous function such that φ(x,y)=0 if and only if x=y=0.

  1. 1.

    ([4]) f is said to be weakly C-contractive (or a weak C-contraction) if for all x,yX,

    d(fx,fy) 1 2 ( d ( x , f y ) + d ( y , f x ) ) φ ( d ( x , f y ) , d ( y , f x ) ) .
  2. 2.

    ([5]) f is said to be weakly K-contractive (or a weak K-contraction) if for all x,yX,

    d(fx,fy) 1 2 ( d ( x , f x ) + d ( y , f y ) ) φ ( d ( x , f x ) , d ( y , f y ) ) .

In 2009, Choudhury [4] proved the following theorem.

Theorem 1 ([[4], Theorem 2.1])

Every weak C-contraction on a complete metric space has a unique fixed point.

For more details of weakly C-contractive mappings we refer to [6] and [7].

Definition 3 Let (X,d) be a metric space and T,f:XX be two mappings.

  1. 1.

    ([8]) f:XX is said to be a T-Kannan-contraction if there exists α(0, 1 2 ) such that for all x,yX the following inequality holds:

    d(Tfx,Tfy)α ( d ( T x , T f x ) + d ( T y , T f y ) ) .
  2. 2.

    ([5]) f:XX is said to be a T-Chatterjea-contraction if there exists α(0, 1 2 ) such that for all x,yX the following inequality holds:

    d(Tfx,Tfy)α ( d ( T x , T f y ) + d ( T y , T f x ) ) .

T-Kannan-contractions (in short, T-K-contractions) and T-Chatterjea-contractions (in short, T-C-contractions) are special cases of T-Hardy-Rogers contractions [9]. Recently, existence and uniqueness of fixed points for these types of contractions in cone metric spaces have been investigated in [9] and [10].

Definition 4 ([11])

Let (X,d) be a metric space. A mapping T:XX is said to be sequentially convergent (respectively, subsequentially convergent) if, for a sequence { x n } in X for which {T x n } is convergent, { x n } is also convergent (respectively, { x n } has a convergent subsequence).

In [8], Moradi has extended Kannan’s theorem [2] as follows.

Theorem 2 (Extended Kannan’s theorem [8])

Let (X,d) be a complete metric space and T,f:XX be mappings such that T is continuous, one-to-one and subsequentially convergent. If f is a T-K-contraction then f has a unique fixed point. Moreover, if T is sequentially convergent then, for every x 0 X, the sequence of iterates { f n x 0 } converges to this fixed point.

The notion of an altering distance function was introduced by Khan et al. as follows.

Definition 5 ([12])

The function ψ:[0,)[0,) is called an altering distance function, if the following properties are satisfied:

  1. 1.

    ψ is continuous and strictly increasing.

  2. 2.

    ψ(0)=0.

In the following definitions and theorems, ψ is an altering distance function and φ: [ 0 , ) 2 [0,) is a continuous function such that φ(x,y)=0 if and only if x=y=0.

Definition 6 ([5])

Let (X,d) be a metric space and let T,f:XX be two mappings.

  1. 1.

    f is said to be a generalized weak T-C-contraction if, for all x,yX,

    ψ ( d ( T f x , T f y ) ) ψ ( d ( T x , T f y ) + d ( T y , T f x ) 2 ) φ ( d ( T x , T f y ) , d ( T y , T f x ) ) .
  2. 2.

    f is said to be a generalized weak T-K-contraction if, for all x,yX,

    ψ ( d ( T f x , T f y ) ) ψ ( d ( T x , T f x ) + d ( T y , T f y ) 2 ) φ ( d ( T x , T f x ) , d ( T y , T f y ) ) .

Putting ψ(t)=t in the above definition, we obtain the concepts of weak T-C-contraction and weak T-K-contraction.

The following are the main results of [5].

Theorem 3 [5]

Let (X,d) be a complete metric space and let T,f:XX be two mappings such that T is one-to-one and continuous. Suppose that:

  1. 1.

    f is a generalized weak T-C-contraction, or

  2. 2.

    f is a generalized weak T-K-contraction.

Then we have the following.

  1. (i)

    For every x 0 X the sequence {T f n x 0 } is convergent.

  2. (ii)

    If T is subsequentially convergent then f has a unique fixed point.

  3. (iii)

    If T is sequentially convergent then for each x 0 X the sequence { f n x 0 } converges to the fixed point of f.

The aim of this article is to extend the stated results to the framework of b-metric spaces, introduced in 1993 by Czerwik [13]. These form a nontrivial generalization of metric spaces and several fixed point results for single and multivalued mappings in such spaces have been obtained since then (see, e.g., [1417] and the references cited therein). We recall the following.

Definition 7 ([13])

Let X be a (nonempty) set and s1 be a given real number. A function d:X×X[0,) is a b-metric if, for all x,y,zX, the following conditions are satisfied:

( b 1 ) d(x,y)=0 iff x=y,

( b 2 ) d(x,y)=d(y,x),

( b 3 ) d(x,z)s[d(x,y)+d(y,z)].

The pair (X,d) is called a b-metric space.

It should be noted that the class of b-metric spaces is effectively larger than that of metric spaces, since a b-metric is a metric if (and only if) s=1. We present an easy example to show that in general a b-metric need not be a metric.

Example 1 Let (X,ρ) be a metric space, and d(x,y)= ( ρ ( x , y ) ) p , where p1 is a real number. Then d is a b-metric with s= 2 p 1 .

However, (X,d) is not necessarily a metric space. For example, if X=R is the set of real numbers and ρ(x,y)=|xy| is the usual Euclidean metric, then d(x,y)= ( x y ) 2 is a b-metric on ℝ with s=2, but it is not a metric on ℝ.

Recently, Hussain et al. [15] have presented an example of a b-metric which is not continuous (see [[15], Example 2]). Thus, while working in b-metric spaces, the following lemma is useful.

Lemma 1 ([14])

Let (X,d) be a b-metric space with s1, and suppose that the sequences { x n } and { y n } are b-convergent to x, y, respectively. Then we have

1 s 2 d(x,y) lim inf n d( x n , y n ) lim sup n d( x n , y n ) s 2 d(x,y).

In particular, if x=y, then we have lim n d( x n , y n )=0. Moreover, for each zX, we have,

1 s d(x,z) lim inf n d( x n ,z) lim sup n d( x n ,z)sd(x,z).

2 Fixed points of weakly T-Chatterjea contractions

From now on, we assume:

Ψ= { ψ : [ 0 , ) [ 0 , ) ψ  is an altering distance function }

and

Φ = { φ : [ 0 , ) 2 [ 0 , ) | φ ( x , y ) = 0 x = y = 0  and φ ( lim inf n a n , lim inf n b n ) lim inf n φ ( a n , b n ) } .

Our first result is the following.

Theorem 4 Let (X,d) be a complete b-metric space with parameter s1, T,f:XX be such that, for some ψΨ, φΦ and all x,yX,

ψ ( s d ( T f x , T f y ) ) ψ ( d ( T x , T f y ) + d ( T y , T f x ) s + 1 ) φ ( d ( T x , T f y ) , d ( T y , T f x ) ) ,
(2.1)

and let T be one-to-one and continuous. Then we have the following.

  1. (1)

    For every x 0 X the sequence {T f n x 0 } is convergent.

  2. (2)

    If T is subsequentially convergent, then f has a unique fixed point.

  3. (3)

    If T is sequentially convergent, then for each x 0 X the sequence { f n x 0 } converges to the fixed point of f.

Proof Let x 0 X be arbitrary. Consider the sequence { x n } n = 0 given by x n + 1 =f x n = f n + 1 x 0 , n=0,1,2, . We will complete the proof in three steps.

Step I. We will prove that lim n d(T x n ,T x n + 1 )=0.

Using condition (2.1), we obtain

ψ ( s d ( T x n + 1 , T x n ) ) = ψ ( s d ( T f x n , T f x n 1 ) ) ψ ( d ( T x n , T f x n 1 ) + d ( T x n 1 , T f x n ) s + 1 ) φ ( d ( T x n , T f x n 1 ) , d ( T x n 1 , T f x n ) ) = ψ ( d ( T x n , T x n ) + d ( T x n 1 , T x n + 1 ) s + 1 ) φ ( d ( T x n , T x n ) , d ( T x n 1 , T x n + 1 ) ) .
(2.2)

Therefore, by the triangular inequality and since φ is nonnegative and ψ is an increasing function,

ψ ( s d ( T x n + 1 , T x n ) ) ψ ( d ( T x n 1 , T x n + 1 ) s + 1 ) ψ ( s s + 1 ( d ( T x n 1 , T x n ) + d ( T x n , T x n + 1 ) ) ) .

Again, since ψ is increasing, we have

d(T x n + 1 ,T x n ) 1 s + 1 ( d ( T x n 1 , T x n ) + d ( T x n , T x n + 1 ) ) ,

wherefrom

d(T x n + 1 ,T x n ) 1 s d(T x n ,T x n 1 )d(T x n ,T x n 1 ).

Thus, {d(T x n + 1 ,T x n )} is a decreasing sequence of nonnegative real numbers and hence it is convergent.

Assume that lim n d(T x n + 1 ,T x n )=r0. From the above argument we have

s d ( T x n + 1 , T x n ) 1 s + 1 d ( T x n 1 , T x n + 1 ) s s + 1 ( d ( T x n 1 , T x n ) + d ( T x n , T x n + 1 ) ) s 2 ( d ( T x n 1 , T x n ) + d ( T x n , T x n + 1 ) ) .

Passing to the limit when n, we obtain

lim n d(T x n 1 ,T x n + 1 )=s(s+1)r.

We have proved in (2.2) that

ψ ( s d ( T x n + 1 , T x n ) ) ψ ( 0 + d ( T x n 1 , T x n + 1 ) s + 1 ) φ ( 0 , d ( T x n 1 , T x n + 1 ) ) .

Now, letting n and using the continuity of ψ and the properties of φ we obtain

ψ(sr)ψ(sr)φ ( 0 , s ( s + 1 ) r ) ,

and consequently, φ(0,s(s+1)r)=0. This yields

r= lim n d(T x n ,T x n + 1 )=0,
(2.3)

by our assumptions about φ.

Step II. {T x n } is a b-Cauchy sequence.

Suppose that {T x n } is not a b-Cauchy sequence. Then there exists ε>0 for which we can find subsequences {T x m ( k ) } and {T x n ( k ) } of {T x n } such that n(k) is the smallest index for which n(k)>m(k)>k and

d(T x m ( k ) ,T x n ( k ) )ε.
(2.4)

This means that

d(T x m ( k ) ,T x n ( k ) 1 )<ε.
(2.5)

From (2.4), (2.5) and the triangular inequality,

ε d ( T x m ( k ) , T x n ( k ) ) s [ d ( T x m ( k ) , T x n ( k ) 1 ) + d ( T x n ( k ) 1 , T x n ( k ) ) ] < s ε + s d ( T x n ( k ) 1 , T x n ( k ) ) .

Letting k, and taking into account (2.3), we can conclude that

ε lim sup k d(T x m ( k ) ,T x n ( k ) )sε.
(2.6)

Further, from

d(T x m ( k ) ,T x n ( k ) )s [ d ( T x m ( k ) , T x n ( k ) 1 ) + d ( T x n ( k ) 1 , T x n ( k ) ) ]

and (2.5), and using (2.3), we get

ε s lim sup k d(T x n ( k ) 1 ,T x m ( k ) )ε.
(2.7)

Moreover, from

d(T x m ( k ) ,T x n ( k ) )s [ d ( T x m ( k ) , T x m ( k ) 1 ) + d ( T x m ( k ) 1 , T x n ( k ) ) ]

and

d(T x m ( k ) 1 ,T x n ( k ) )s [ d ( T x m ( k ) 1 , T x m ( k ) ) + d ( T x m ( k ) , T x n ( k ) ) ] ,

and using (2.3) and (2.6), we get

ε s lim sup k d(T x m ( k ) 1 ,T x n ( k ) ) s 2 ε.
(2.8)

Similarly, we can show that

ε s lim inf k d(T x n ( k ) 1 ,T x m ( k ) )ε
(2.9)

and

ε s lim inf k d(T x m ( k ) 1 ,T x n ( k ) ) s 2 ε.
(2.10)

Using (2.1) and (2.7)-(2.10) we have

ψ ( s ε ) ψ ( s lim sup k d ( T x m ( k ) , T x n ( k ) ) ) = ψ ( s lim sup k d ( T f x m ( k ) 1 , T f x n ( k ) 1 ) ) lim sup k ψ ( d ( T x m ( k ) 1 , T f x n ( k ) 1 ) + d ( T x n ( k ) 1 , T f x m ( k ) 1 ) s + 1 ) lim inf k φ ( d ( T x m ( k ) 1 , T f x n ( k ) 1 ) , d ( T x n ( k ) 1 , T f x m ( k ) 1 ) ) ψ ( lim sup k d ( T x m ( k ) 1 , T x n ( k ) ) + d ( T x n ( k ) 1 , T x m ( k ) ) s + 1 ) φ ( lim inf k d ( T x m ( k ) 1 , T x n ( k ) ) , lim inf k d ( T x n ( k ) 1 , T x m ( k ) ) ) ψ ( s 2 ε + ε s + 1 ) φ ( lim inf k d ( T x m ( k ) 1 , T x n ( k ) ) , lim inf k d ( T x n ( k ) 1 , T x m ( k ) ) ) ψ ( s ε ) φ ( lim inf k d ( T x m ( k ) 1 , T x n ( k ) ) , lim inf k d ( T x n ( k ) 1 , T x m ( k ) ) )

since s 2 + 1 s + 1 s. Hence, we have

φ ( lim inf k d ( T x m ( k ) 1 , T x n ( k ) ) , lim inf k d ( T x n ( k ) 1 , T x m ( k ) ) ) 0.

By our assumption about φ, we have

lim inf k d(T x m ( k ) 1 ,T x n ( k ) )= lim inf k d(T x n ( k ) 1 ,T x m ( k ) )=0,

which contradicts (2.9) and (2.10).

Since (X,d) is b-complete and {T x n }={T f n x 0 } is a b-Cauchy sequence, there exists vX such that

lim n T f n x 0 =v.
(2.11)

Step III. f has a unique fixed point, assuming that T is subsequentially convergent.

As T is subsequentially convergent, { f n x 0 } has a b-convergent subsequence. Hence, there exist uX and a subsequence { n i } such that

lim i f n i x 0 =u.
(2.12)

Since T is continuous, by (2.12) we obtain

lim i T f n i x 0 =Tu,
(2.13)

and by (2.11) and (2.13) we conclude that Tu=v.

From Lemma 1 and (2.1) we have

ψ ( s 1 s d ( T f u , T u ) ) ψ ( lim sup n s d ( T f u , T f n + 1 x 0 ) ) = ψ ( lim sup n s d ( T f u , T f x n ) ) ψ ( lim sup n d ( T u , T f x n ) + d ( T x n , T f u ) s + 1 ) lim inf n φ ( d ( T u , T f x n ) , d ( T x n , T f u ) ) ψ ( s d ( T u , T u ) + s d ( T u , T f u ) s + 1 ) φ ( lim inf n d ( T u , T f x n ) , lim inf n d ( T x n , T f u ) ) ψ ( d ( T u , T f u ) ) φ ( 0 , lim inf n d ( T x n , T f u ) ) ,

since ψ is increasing. By the properties of φΦ, it follows that lim inf n d(T x n ,Tfu)=0. By the triangular inequality we have

d(Tfu,Tu)s [ d ( T f u , T x n ) + d ( T x n , T u ) ] .

Letting n we can conclude that d(Tfu,Tu)=0. Hence, Tfu=Tu. As T is one-to-one, fu=u. Consequently, f has a fixed point.

If we assume that w is another fixed point of f, because of (2.1), we have

ψ ( s d ( T u , T w ) ) = ψ ( s d ( T f u , T f w ) ) ψ ( d ( T u , T f w ) + d ( T w , T f u ) s + 1 ) φ ( d ( T u , T f w ) , d ( T w , T f u ) ) = ψ ( d ( T u , T w ) + d ( T w , T u ) s + 1 ) φ ( d ( T u , T w ) , d ( T w , T u ) ) ψ ( s d ( T u , T w ) ) φ ( d ( T u , T w ) , d ( T w , T u ) ) ,

since 2 s + 1 s and ψ is increasing. Hence, d(Tu,Tw)=0. Since T is one-to-one, it follows that u=w. Consequently, f has a unique fixed point.

Finally, if T is sequentially convergent, replacing {n} with { n i } we conclude that lim n f n x 0 =u. □

Taking ψ(t)=t and φ(t,u)=( 1 s + 1 α)(t+u), where α[0, 1 s + 1 ) in Theorem 4, the extended Chatterjea’s theorem in the setting of b-metric spaces is obtained.

Corollary 1 Let (X,d) be a complete b-metric space and T,f:XX be mappings such that T is continuous, one-to-one and subsequentially convergent. If α[0, 1 s + 1 ) and

d(Tfx,Tfy) α s ( d ( T x , T f y ) + d ( T y , T f x ) ) ,

for all x,yX, then f has a unique fixed point. Moreover, if T is sequentially convergent, then for every x 0 X the sequence of iterates f n x 0 converges to this fixed point.

Remark 1 In the case when Tx=x, this corollary reduces to [[18], Corollary 3.8.3] (the case g=f), which is Chatterjea’s theorem [3] in the framework of b-metric spaces.

By taking Tx=x and ψ(t)=t in Theorem 4, we derive an extension of Choudhury’s theorem (Theorem 1) to the setup of b-metric spaces.

If s=1, Theorem 4 reduces to Theorem 3 (case (1)).

We demonstrate the use of the obtained results by the following.

Example 2 (Inspired by [8])

Let X={0}{1/nnN}, and let d(x,y)= ( x y ) 2 for x,yX. Then d is a b-metric with the parameter s=2 and (X,d) is a complete b-metric space. Consider the mappings f,T:XX given by

f(0)=0,f ( 1 n ) = 1 n + 1 ,T(0)=0,T ( 1 n ) = 1 n n ,nN.

We will show that the mappings f, T satisfy the conditions of Corollary 1 with α= 2 9 < 1 3 = 1 s + 1 . Indeed, for m,nN, m>n, we have

d ( T f 1 n , T f 1 m ) = [ 1 ( n + 1 ) n + 1 1 ( m + 1 ) m + 1 ] 2 < [ 1 ( n + 1 ) n + 1 ] 2 .

It is easy to prove that, for nN,

1 ( n + 1 ) n + 1 < 1 3 [ 1 n n 1 ( n + 2 ) n + 2 ] .

It follows that

d ( T f 1 n , T f 1 m ) < 1 9 [ 1 n n 1 ( n + 2 ) n + 2 ] 2 .

Now, m>n implies that mn+1 and n+2m+1. It follows that 1/ ( n + 2 ) n + 2 1/ ( m + 1 ) m + 1 , and hence

d ( T f 1 n , T f 1 m ) < 1 9 [ 1 n n 1 ( m + 1 ) m + 1 ] 2 α s [ d ( T 1 n , T f 1 m ) + d ( T 1 m , T F 1 n ) ] .

If one of the points is equal to 0, the proof is even simpler.

By Corollary 1, it follows that f has a unique fixed point (which is u=0).

3 Fixed points of weakly T-Kannan contractions

Our second main result is the following.

Theorem 5 Let (X,d) be a complete b-metric space with the parameter s1, T,f:XX be such that for some ψΨ, φΦ and all x,yX,

ψ ( d ( T f x , T f y ) ) ψ ( d ( T x , T f x ) + d ( T y , T f y ) s + 1 ) φ ( d ( T x , T f x ) , d ( T y , T f y ) ) .
(3.1)

and let T be one-to-one and continuous. Then:

  1. (1)

    For every x 0 X the sequence {T f n x 0 } is convergent.

  2. (2)

    If T is subsequentially convergent, then f has a unique fixed point.

  3. (3)

    If T is sequentially convergent then, for each x 0 X, the sequence { f n x 0 } converges to the fixed point of f.

Proof Let x 0 X be arbitrary. Consider the sequence { x n } n = 0 given by x n + 1 =f x n = f n + 1 x 0 , n=0,1,2, . At first, we will prove that

lim n d(T x n ,T x n + 1 )=0.

Using condition (3.1), we obtain

ψ ( d ( T x n + 1 , T x n ) ) = ψ ( d ( T f x n , T f x n 1 ) ) ψ ( d ( T x n , T f x n ) + d ( T x n 1 , T f x n 1 ) s + 1 ) φ ( d ( T x n , T f x n ) , d ( T x n 1 , T f x n 1 ) ) = ψ ( d ( T x n , T x n + 1 ) + d ( T x n 1 , T x n ) s + 1 ) φ ( d ( T x n , T f x n ) , d ( T x n 1 , T f x n 1 ) ) .
(3.2)

Since φ is nonnegative and ψ is increasing, it follows that

d(T x n + 1 ,T x n ) d ( T x n , T x n + 1 ) + d ( T x n 1 , T x n ) s + 1 ,

that is,

d(T x n + 1 ,T x n ) 1 s d(T x n ,T x n 1 )d(T x n ,T x n 1 ).

Thus, {d(T x n + 1 ,T x n )} is a decreasing sequence of nonnegative real numbers and hence it is convergent.

Assume that lim n d(T x n + 1 ,T x n )=r. If in (3.2) n, using the properties of ψ and φ we obtain

ψ(r)ψ ( 2 r s + 1 ) φ(r,r)ψ(r)φ(r,r),

which is possible only if

r= lim n d(T x n ,T x n + 1 )=0.

Now, we will show that {T x n } is a b-Cauchy sequence.

Suppose that this is not true. Then there exists ε>0 for which we can find subsequences {T x m ( k ) } and {T x n ( k ) } of {T x n } such that n(k) is the smallest index for which n(k)>m(k)>k and d(T x m ( k ) ,T x n ( k ) )ε. This means that

d(T x m ( k ) ,T x n ( k ) 1 )<ε.

Again, as in Step II of Theorem 4 one can prove that

ε lim sup k d(T x m ( k ) ,T x n ( k ) )sε.
(3.3)

Using (3.1) we have

ψ ( d ( T x m ( k ) , T x n ( k ) ) ) = ψ ( d ( T f x m ( k ) 1 , T f x n ( k ) 1 ) ) ψ ( d ( T x m ( k ) 1 , T f x m ( k ) 1 ) + d ( T x n ( k ) 1 , T f x n ( k ) 1 ) s + 1 ) φ ( d ( T x m ( k ) 1 , T f x m ( k ) 1 ) , d ( T x n ( k ) 1 , T f x n ( k ) 1 ) ) = ψ ( d ( T x m ( k ) 1 , T x m ( k ) ) + d ( T x n ( k ) 1 , T x n ( k ) ) s + 1 ) φ ( d ( T x m ( k ) 1 , T x m ( k ) ) , d ( T x n ( k ) 1 , T x n ( k ) ) ) .

Passing to the upper limit as k in the above inequality and taking into account (3.3), we have

ψ(ε)ψ(0)φ(0,0)=0,

and so ψ(ε)=0. By our assumptions about ψ, we have ε=0, which is a contradiction.

Since (X,d) is b-complete and {T x n }={T f n x 0 } is a b-Cauchy sequence, there exists vX such that

lim n T f n x 0 =v.
(3.4)

Now, if T is subsequentially convergent, then { f n x 0 } has a convergent subsequence. Hence, there exist a point uX and a sequence { n i } such that

lim i f n i x 0 =u.
(3.5)

Since T is continuous, by (3.5) we obtain

lim i T f n i x 0 =Tu,
(3.6)

and by (3.4) and (3.6) we conclude that Tu=v.

From Lemma 1 and (3.1) we have

ψ ( 1 s d ( T f u , T u ) ) ψ ( lim sup n d ( T f u , T f n + 1 x 0 ) ) = ψ ( lim sup n d ( T f u , T f x n ) ) ψ ( lim sup n d ( T u , T f u ) + d ( T x n , T f x n ) s + 1 ) lim inf n φ ( d ( T u , T f u ) , d ( T x n , T f x n ) ) = ψ ( d ( T u , T f u ) + 0 s + 1 ) φ ( d ( T u , T f u ) , 0 ) ψ ( d ( T u , T f u ) s ) φ ( d ( T u , T f u ) , 0 ) .

By the properties of φΦ, it follows that

d(Tu,Tfu)=0.

Since T is one-to-one, we obtain fu=u. Consequently, f has a fixed point.

Uniqueness of the fixed point can be proved in the same manner as in Theorem 4.

Finally, if T is sequentially convergent, replacing {n} with { n i } we conclude that lim n f n x 0 =u. □

Taking ψ(t)=t and φ(t,u)=( 1 s + 1 α)(t+u), where α[0, 1 s + 1 ) in Theorem 5, the extended Kannan’s theorem in the setting of b-metric spaces is obtained.

Corollary 2 Let (X,d) be a complete b-metric space with the parameter s1, T,f:XX be such that for some α< 1 s + 1 and all x,yX,

d(Tfx,Tfy)α ( d ( T x , T f x ) + d ( T y , T f y ) )
(3.7)

and let T be one-to-one and continuous. Then we have the following.

  1. (1)

    For every x 0 X the sequence {T f n x 0 } is convergent.

  2. (2)

    If T is subsequentially convergent then f has a unique fixed point.

  3. (3)

    If T is sequentially convergent then, for each x 0 X, the sequence { f n x 0 } converges to the fixed point of f.

Remark 2 In the case when Tx=x, this corollary reduces to [[18], Corollary 3.8.2] (the case g=f). If s=1, Corollary 2 reduces to Theorem 2 (i.e., [[8], Theorem 2.1]). Of course, if both of these conditions are fulfilled, we get just the classical Kannan’s theorem [2].

The following example distinguishes our results from the previously known ones.

Example 3 Let X={a,b,c} and d:X×XR be defined by d(x,x)=0 for xX, d(a,b)=d(b,c)=1, d(a,c)= 9 4 , d(x,y)=d(y,x) for x,yX. It is easy to check that (X,d) is a b-metric space (with s= 9 8 >1) which is not a metric space. Consider the mapping f:XX given by

f= ( a b c a a b ) .

We first note that the b-metric version of classical weak Kannan’s theorem is not satisfied in this example. Indeed, for x=b, y=c, we have d(fx,fy)=d(a,b)=1 and d(x,fx)+d(y,fy)=d(b,a)+d(c,b)=2, hence the inequality

ψ ( d ( f x , f y ) ) ψ ( d ( x , f x ) + d ( y , f y ) s + 1 ) φ ( d ( x , f x ) , d ( y , f y ) )

cannot hold, whatever ψΨ and φΦ are chosen.

Take now T:XX defined by

T= ( a b c b c a ) .

Obviously, all the properties of T given in Corollary 2 are fulfilled. We will check that the contractive condition (3.7) holds true if α is chosen such that

4 9 <α< 8 17 = 1 s + 1 .

Only the following cases are nontrivial:

1 x=a, y=c. Then (3.7) reduces to

d(Tfa,Tfc)=d(b,c)=1= 4 9 9 4 <α ( d ( b , b ) + d ( a , c ) ) =α ( d ( T a , T f a ) + d ( T c , T f c ) ) .

2 x=b, y=c. Then (3.7) reduces to

d(Tfb,Tfc)=d(b,c)=1< 4 9 13 4 <α ( d ( c , b ) + d ( a , c ) ) =α ( d ( T b , T f b ) + d ( T c , T f c ) ) .

All the conditions of Corollary 2 are satisfied and f has a unique fixed point (u=a).