PROCEEDINGS OF THE

AMERICAN MATHEMATICAL SOCIETY

Volume 146, Number 4, April 2018, Pages 1585–1598
http://dx.doi.org/10.1090/proc/13808
Article electronically published on December 28, 2017

SOLVING EXISTENCE PROBLEMS VIA F -CONTRACTIONS

DARIUSZ WARDOWSKI

(Communicated by Wenxian Shen)

Abstract. The main results of the paper concern the existence of fixed points
of nonlinear F -contraction and the sum of this type of mapping with a compact
operator. The results of Krasnosel’skii type are obtained with a usage of
the Hausdorff measure of noncompactness and condensing mappings. The
presented new tools give the possibility to verify the existence problems of the
solutions for some classes of integral equations.

1. Introduction and preliminaries

Krasnosel’skii’s fixed point theorem about the sum of contraction and compact
mappings is an important result merging two fundamental results in the fixed point
theory, Banach contraction principle and Schauder fixed point theorem. The origi-
nal statement of this theorem can be found in [15] and we may read it as follows:
If M is a nonempty closed convex and bounded subset of the Banach space X,
A : M → X is a strict contraction, B : M → X is a compact mapping, i.e. continu-
ous and maps a subset of M into a compact subset of X, and A(M ) + B(M ) ⊂ M ,
then A + B admits a fixed point. This result can be applied in many interesting
contexts, especially in the theory of differential and integral equations. In the liter-
ature we can find many important contributions in this direction; see e.g. Burton’s
results in [2], where Krasnosel’skii’s thesis is obtained for large contractions; in [3]
Burton and Colleen Kirk combined the contractions with Schaefer’s theorem [24].
In [17] Liu and Li investigated the case where the mappings are multivalued and the
operator I − A need not be injective. All these improvements have made it easier
to apply the obtained more general results. However, a significant breakthrough
was achieved when it started to involve the measure of noncompactness. Then it
was possible to replace the difficulty of verifying the condition A(M ) + B(M ) ⊂ M
onto the weaker
(1.1) (A + B)(M ) ⊂ M.
We recall the notion of the Hausdorff measure of noncompactness; i.e. for any
bounded subset C ⊂ X, there is assigned a nonnegative number β(C) by the
formula

N
β(C) := inf{r > 0 : C ⊂ B(xi , r), xi ∈ X, i = 1, . . . , N },
i=1

Received by the editors February 2, 2017 and, in revised form, May 4, 2017.
2010 Mathematics Subject Classification. Primary 47H10, 47H09, 47H08, 47N20.
Key words and phrases. Fixed point, nonlinear F -contraction, compact operator, condensing
mapping, integral equation.

2017
c American Mathematical Society

1585

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 1586 DARIUSZ WARDOWSKI

where B(xi , r) denotes the closed ball centred at xi with a radius r. Some of
the basic properties of the Hausdorff measure of noncompactness are the following:
β(C) = 0 if and only if C is relatively compact (i.e. clC is compact), β(C) = β(clC);
if we assume that D ⊂ X is bounded, then we also have C ⊂ D implies β(C) ≤
β(D) and β(C + D) ≤ β(C) + β(D). More information about the measure of
noncompactness and its properties can be found e.g. in [1]. The classical fixed
point results where measure of noncompactness is applied are due to Darbo [5] and
Sadovskii [23]. If M is a nonempty bounded closed convex subset of a Banach space
X, T : M → M is a continuous mapping such that one of the following conditions
holds:
(a) There exists k ∈ [0, 1) such that for any set C ⊂ M ,
(Darbo) β(T (C)) ≤ kβ(C).
(b) For any set C ⊂ M with positive measure of noncompactness
(Sadovskii) β(T (C)) < β(C).
Then T has a fixed point.
The mappings satisfying the contraction condition in Darbo’s and Sadovskii’s re-
sults are called k-set contraction and β-condensing respectively. Using these results
we can obtain the existence of a fixed point of A + B, provided that A + B is β-
condensing or a k-set contraction. In this way Przeradzki in his work [21] showed
that the sum of a generalized contraction (sometimes called Krasnosel’skii’s con-
traction) and compact mapping satisfying (1.1) has a fixed point. In the literature,
one can find also other significant contributions inspired by Krasnosel’skii’s result.
For example Kryszewski and Mederski in their work [16] for the set-valued oper-
ators of Krasnosel’skii type (the sum A + B was replaced by a general nonlinear
operator) obtained fixed point results on the complete absolute neighbourhood re-
tracts. In [6] Garcia-Falset et al. proved a Krasnosel’skii-Schaefer type theorem,
where the investigated operators need not be weakly continuous.
As was mentioned, some authors made an effort to improve Krasnosel’skii’s the-
orem by extending the family of mappings satisfying the appropriate new, more
general contraction condition. There are many generalizations of Banach’s contrac-
tion. One can mention here e.g. the recent paper by Wlodarczyk [36], where the
author proves fixed point results in a very general setting in the so-called quasi-
triangular spaces. In [36] there is also included a comprehensive list of references of
the articles where other known contraction conditions have been investigated. In
Kirk’s handbook [13] we can also find out about the recent developments in metric
fixed point theory. However many new defined contractions are special cases of the
others. The information on this topic can be found for example in Rhoades’ article
[22] or in Jachymski’s papers [11] and [12]. The inspiration for our investigations
is the aforementioned Przeradzki article, where there were taken into consideration
generalized contractions. If X is a nonempty set and (X, d) is a metric space with
a metric function d : X × X → R, then a mapping T : X → X is called a generalized
contraction if there exists a function Γ : X × X → [0, ∞) such that:
(Γ1) supa≤d(x,y)≤b Γ(x, y) < 1 for all 0 < a ≤ b,
(Γ2) d(T x, T y) ≤ Γ(x, y)d(x, y) for all x, y ∈ X.
Many known contractions are equivalent to generalized contractions; for details see
[12]. Przeradzki showed that generalized contractions are a proper generalization

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 SOLVING EXISTENCE PROBLEMS VIA F -CONTRACTIONS 1587

of the so-called large contractions (introduced by Burton in [2]), i.e. the mappings
T : X → X such that for every ε > 0 there exists δ > 0 such that for all x, y ∈ X,
d(x, y) ≥ ε implies d(T x, T y) < ε.
Przeradzki in [21] also showed that generalized contractions are β-condensing, which
together with Sadovskii’s result and condition (1.1) derived a significant improve-
ment of Krasnosel’skii’s result. On the other hand, in [35] there was introduced a
new type of contraction condition, called F -contraction, i.e. a mapping T : X → X
satisfying
τ + F (d(T x, T y)) ≤ F (d(x, y)) for all x, y ∈ X with T x = T y,
where τ > 0 and F : (0, ∞) → R satisfies the following conditions:
(F1) for all t1 , t2 > 0, t1 > t2 implies F (t1 ) > F (t2 );
(F2) for any sequence (tn ) ⊂ (0, ∞), tn → 0 if and only if F (tn ) → −∞;
(F3) there exists k ∈ (0, 1) satisfying limt→0+ tk F (t) = 0.
It was also proved that every F -contraction defined on a complete metric space has
a unique fixed point. Using concrete forms of F it is possible to obtain other known
types of contractions; e.g. for F (t) = ln(t), t > 0, we get a Banach contraction (for
details, see [35]). Many articles concerning F -contractions and their extensions
have appeared so far; see e.g. [7–10, 18–20, 25–34]. Turinici in [32] observed that
the condition (F2) can be relaxed to the form
(F2’) limt→0+ F (t) = −∞.
Then the implication
(F2”)] F (tn ) → −∞ ⇒ tn → 0
can be derived from (F1). In the present paper, we are going to reconsider this
type of mapping in a broader setting, i.e. with τ taken as a function. A mapping
T : X → X is said to be a (ϕ, F )-contraction (or nonlinear F -contraction) if there
exist the functions F : (0, ∞) → R and ϕ : (0, ∞) → (0, ∞) satisfying
(H1) F satisfies (F1) and (F2’);
(H2) lim inf s→t+ ϕ(s) > 0 for all t ≥ 0;
(H3) ϕ(d(x, y)) + F (d(T x, T y)) ≤ F (d(x, y)) for all x, y ∈ X such that T x = T y.
Nonlinear F -contractions have been studied e.g. in [14], however observe that in
the present investigations we omit the condition (F3). The proposed type of the
contraction condition is not a special case of a generalized contraction. In the
following, there is presented the example of a (ϕ, F )-contraction which is not a
generalized contraction.
Example 1.1. Let (xn ) be a sequence given by the formula
xn := n − 2 + 2−n+1 , n ≥ 1.
The set X = {xn : n ∈ N} together with the metric d(x, y) = |x − y|, x, y ∈ X, is a
complete metric space. Consider the mapping T : X → X defined by the formula

xn−1 , for n ≥ 2,
T xn :=
x1 , for n = 1.
First, observe that T is not a generalized contraction. Indeed, taking any n ∈ N we
have
|xn+1 − xn | = 1 − 2−n

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 1588 DARIUSZ WARDOWSKI

and hence
1
≤ |xn+1 − xn | < 1 for all n ∈ N.
2
If there existed a function Γ satisfying (Γ2), then for all n ≥ 2 we would get

|T xn+1 − T xn | |xn − xn−1 | 2n − 2

Γ(xn+1 , xn ) ≥ = = n .
|xn+1 − xn | |xn+1 − xn | 2 −1

In consequence, by (Γ1), we would obtain

2n − 2
1> sup Γ(x, y) ≥ Γ(xn+1 , xn ) ≥ .
2n − 1
2 ≤d(x,y)≤1
1

Tending with n → ∞ we get a contradiction. In order to show that T is a nonlinear

F -contraction let us consider the mapping ϕ : (0, ∞) → (0, ∞) by the formula

−t + 1, for 0 < t < 1,
ϕ(t) :=
−t + n, for n − 1 ≤ t < n, n ≥ 2.

Obviously lim inf s→t+ ϕ(s) > 0 for any t ≥ 0, and for any m, n ∈ N, 2 ≤ m < n we
have

|xm − xn | = n − m + 2−n+1 − 2−m+1 and n − m − 1 < |xm − xn | < n − m.

Hence we get
|T xm − T xn | |T xm −T xn |−|xm −xn |
e < exn−1 −xm−1 −(xn −xm )
|xm − xn |
−n+1
−2−m+1
= e2 = e−ϕ(|xm −xn |) .

Next, for any m ≥ 3 we obtain

|x1 − xm | = m − 2 + 2−m+1 and m − 2 < |x1 − xm | < m − 1,

which gives
|T x1 − T xm | |T x1 −T xm |−|x1 −xm |
e < exm−1 −xm
|x1 − xm |
−m+1
−1
= e2 = e−ϕ(|x1 −xm |) .

Now, after some simple calculations we can observe that T satisfies (H3) for F (t) =
ln t + t, t > 0.

In the present paper first we prove a fixed point theorem for nonlinear F -
contractions; next we show that some of these contractions are β-condensing. Fi-
nally, using Sadovskii’s result, we get that the sum of compact mapping with a
(ϕ, F )-contraction has a fixed point. In the last section we apply our results to
certain classes of integral equations.
Within the article we will use the following notation: N denotes the set of all
positive integers, Q the set of all rational numbers, R and R+ the set of all real
numbers and all nonnegative real numbers respectively.

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 SOLVING EXISTENCE PROBLEMS VIA F -CONTRACTIONS 1589

2. The results
First we enunciate a fixed point result concerning nonlinear F -contractions. Ob-
serve that in the proof of this fact we will not use the condition (F3).
Theorem 2.1. Let (X, d) be a complete metric space and let T : X → X be a
(ϕ, F )-contraction. Then T has a unique fixed point.
Proof. T has at most one fixed point, which is an immediate consequence of (H3)
and the fact that ϕ > 0.
Take any x0 ∈ X and define the sequence xn = T n x0 , n = 1, 2, . . . . Denote the
sequence γn = d(xn−1 , xn ), n ∈ N. Without loss of generality we can assume that
γn > 0 for all n ∈ N. From (H3) we have
F (γn+1 ) ≤ F (γn ) − ϕ(γn ) < F (γn ) for all n ∈ N.
From the above and from (F1) we get that (γn ) is decreasing, and hence, γn t,
t ≥ 0. From (H2) there exists c > 0 and n0 ∈ N such that ϕ(γn ) > c for all n ≥ n0 .
In consequence, we have

n−1
F (γn ) ≤ F (γn−1 ) − ϕ(γn−1 ) ≤ · · · ≤ F (γ1 ) − ϕ(γi )
i=1
0 −1
n 
n−1
= F (γ1 ) − ϕ(γi ) − ϕ(γi ) < F (γ1 ) − (n − n0 )c, n > n0 .
i=1 i=n0

Tending with n → ∞ we get F (γn ) → −∞ and, by (F2”), γn → 0.

To show that (xn ) is the Cauchy sequence, we will use the technique due to
Turinici [32]. Suppose on the contrary that (xn ) is not Cauchy. From (F1) the set
Δ of all discontinuity points of F is at most countable. There exists η > 0, η ∈ Δ
such that for every k ≥ 0 one can find mk , nk ∈ N satisfying
(2.2) k ≤ mk < nk and d(xmk , xnk ) > η.
Denote by mk the least of mk satisfying (2.2) and by nk the least of nk such that
mk < nk and d(xmk , xnk ) > η. Naturally
(2.3) d(xmk , xnk ) > η for all k ≥ 0.
Observe that taking k0 ∈ N such that γk < η for all k ≥ k0 , we have
η < d(xmk , xnk ) ≤ d(xmk , xnk −1 ) + d(xnk −1 , xnk ) ≤ η + γnk , for all k ≥ k0 .
Tending with k → ∞ we obtain
(2.4) d(xmk , xnk ) → η.
Also observe that for all k ≥ 0 we have
d(xmk , xnk ) − γmk +1 − γnk +1 ≤ d(xmk +1 , xnk +1 ) ≤ γmk +1 + d(xmk , xnk ) + γnk +1 .
Again, tending with k → ∞ we have
(2.5) d(xmk +1 , xnk +1 ) → η.
Finally observe that from (H3) we get
ϕ(d(xmk , xnk )) ≤ F (d(xmk , xnk )) − F (d(xmk +1 , xnk +1 )), k ≥ 0.

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 1590 DARIUSZ WARDOWSKI

Now, from the above inequality, using (2.3)-(2.5) and the fact that F is continuous
at η one gets
lim inf
+
ϕ(s) ≤ lim inf ϕ(d(xmk , xnk )) ≤ lim (F (d(xmk , xnk )) − F (d(xmk +1 , xnk +1 )))
s→η k→∞ k→∞
= 0,
which contradicts (H2). Therefore (xn ) is Cauchy.
The completeness of X and the continuity of T end the proof. 

Before we prove that some class of nonlinear F -contractions consists of β-

condensing mappings we introduce a notion which will enable us to measure how
much a given function is discontinuous. It will help us to prove that (ϕ, F )-
contractions can be condensing even for discontinuous F .
Let U be a nonempty subset of R, x0 ∈ U , f : U → R a function and η ≥ 0.
Definition 2.1. We say that f is continuous at x0 with accuracy η if for any ε > η
there exists δ > 0 such that for all x ∈ U , |f (x) − f (x0 )| < ε whenever |x − x0 | < δ.
The set of all η ≥ 0 such that f is continuous at x0 with accuracy η will be denoted
by disc(f, x0 ).
Remark 2.1. Note that the set disc(f, x0 ) is closed. Indeed, if we assume that
disc(f, x0 ) is nonempty and consider a sequence (ηn ) ⊂ disc(f, x0 ), ηn → η ∈ R,
then taking any ε > η one can find N ∈ N such that ε > ηN . Since f is continuous
with accuracy ηN , there exists δ > 0 such that |x−x0 | < δ implies |f (x)−(x0 )| < ε.
In consequence f is continuous at x0 with accuracy η, i.e. η ∈ disc(f, x0 ).
In light of the above remark we can establish the following definition.
Definition 2.2. If disc(f, x0 ) = ∅, then the measure of discontinuity of f at x0 ,
denoted by σ (f, x0 ), will be called the nonnegative number given by the formula
(2.6) σ (f, x0 ) := min disc(f, x0 ).
In case disc(f, x0 ) = ∅ we put σ(f, x0 ) := ∞.
Lemma 2.1. If σ(f, x0 ) < ∞, then for any sequence (xn ) ⊂ U such that xn → x0 ,
the following inequalities hold:
(2.7) lim sup f (xn ) − σ (f, x0 ) ≤ f (x0 ) ≤ lim inf f (xn ) + σ (f, x0 ) .
n→∞ n→∞

Proof. Suppose that lim supn→∞ f (xn ) − σ(f, x0 ) > f (x0 ). There exists c > f (x0 )
and a subsequence (xnk ) of (xn ) satisfying
f (xnk ) − σ(f, x0 ) ≥ c for all k ∈ N.
Letting ε = σ(f, x0 ) + c − f (x0 ) and taking any δ > 0 one can find N ∈ N such
that |xnN − x0 | < δ and |f (xnN ) − f (x0 )| ≥ f (xnN ) − f (x0 ) ≥ c + σ(f, x0 ) − f (x0 ) = ε,
which gives that f is not continuous at x0 with accuracy σ(f, x0 ), which is impos-
sible. By an analogous argument we prove the second inequality. 

Observe that from the definition of the measure of discontinuity, the equality
σ (f, x0 ) = 0 easily implies the continuity of the function f at x0 , and then inequality
(2.7) takes the form limn→∞ f (xn ) = f (x0 ).
A straightforward consequence of Lemma 2.1 is the following.

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 SOLVING EXISTENCE PROBLEMS VIA F -CONTRACTIONS 1591

Remark 2.2. If σ(f, x0 ) < ∞, then for any sequence (xn ) ⊂ U such that xn → x0
the following holds:
(2.8) lim sup f (xn ) − lim inf f (xn ) ≤ 2σ (f, x0 ) .
n→∞ n→∞

Example 2.1.
1) For the Dirichlet function IQ : R → {0, 1}, where IQ (x) = 1 for x ∈ Q and
IQ (x) = 0 otherwise, σ (IQ , x0 ) = 1 for every x0 ∈ R.
2) Consider the function f : (−∞, 0) → (−∞, 0) of the form
⎧
⎨ t − 1, if t < −1;
f (t) = 
⎩ t− 1 if t ∈ − n1 , − n+1
1
, n ∈ N.
n+1

Then σ f, − n1 = n(n+1) 1
and σ (f, x) = 0 for x = − n1 , n ∈ N.
3) If f (t) = 1/t, t > 0 and f (0) = 0, then σ(f, 0) = ∞.
Now, we are in a position to state our main results.
Theorem 2.2. Let (X, d) be a metric space and let T : X → X be a (ϕ, F )-
contraction. If
(2.9) lim inf ϕ(s) > 2σ(F, t)
s→t
for all t > 0, then T is β-condensing.
Proof. First, observe that by (F1) we have σ(F, t) < ∞ and thus inequality (2.9) is
valid. Take any C ⊂ X with positive Hausdorff measure of noncompactness. Since
2σ(F, β(C)) < lim inf s→β(C) ϕ(s), there exists ε > 0 such that
(2.10) 2σ(F, β(C)) < ϕ(s) for all s ∈ [β(C) − ε, β(C) + ε].
Applying to F Remark 2.2 and using (F1) we have
(2.11) lim F (t) − lim F (t) ≤ 2σ(F, β(C)) < ϕ(s)
t→β(C)+ t→β(C)−

for all s ∈ [β(C) − ε, β(C) + ε]. Taking in (2.10) appropriately small enough ε, we
can deduce from (2.11) the inequality
(2.12) F (r) − F (t) < ϕ(s)
for all r ∈ (β(C), β(C) + ε], t ∈ [β(C) − ε, β(C)), s ∈ [β(C) − ε, β(C) + ε]. Consider
R = β(C) + ε and take a finite R-net of C, i.e.

k
(2.13) C⊂ B(xi , R), x1 , . . . , xk ∈ X.
i=1

Denote R = β(C)−ε. We will show that the open balls B(T xi , R ), 1 ≤ i ≤ k, cover
T (C). Let y ∈ T (C) and let x ∈ C be such that T x = y. From (2.13) there exists
i ∈ {1, . . . , k} such that d(x, xi ) < R. If T x = T xi , then obviously d(y, T xi ) < R .
Suppose that T x = T xi and consider two cases. If 0 < d(x, xi ) < R , then, since T
is a (ϕ, F )-contraction, we have
F (d(T x, T xi )) ≤ F (d(x, xi )) − ϕ(d(x, xi )) ≤ F (R ) − ϕ(d(x, xi )) < F (R ).
If R ≤ d(x, xi ) < R, then due to (2.12) we obtain the inequalities
F (d(T x, T xi )) ≤ F (d(x, xi )) − ϕ(d(x, xi )) < F (R) − ϕ(d(x, xi )) < F (R ).

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 1592 DARIUSZ WARDOWSKI

In the above both cases (F1) implies d(T x, T xi ) < R , which gives
β(T (C)) ≤ R < β(C).


The immediate consequence of Theorem 2.2 is the following result.

Corollary 2.1. Every (ϕ, F )-contraction with continuous F satisfying
lim inf s→t ϕ(s) > 0 for all t > 0 is β-condensing.
The consequence of Sadovskii’s result is the following theorem.
Theorem 2.3. Let C be a closed bounded and convex subset of a Banach space X.
If A : C → X is a (ϕ, F )-contraction with lim inf s→t ϕ(s) > 2ρ(F, t) for all t > 0,
B : C → X is a compact operator and (A + B)(C) ⊂ C, then A + B has a fixed
point.
Proof. Take any K ⊂ C which is not relatively compact. Using the properties of the
Hausdorff measure of noncompactness, Theorem 2.2 and the fact that β(B(C)) = 0
we obtain
β((A + B)(C)) ≤ β(A(C) + B(C)) ≤ β(A(C)) + β(B(C)) < β(C),
which gives that A+B is a β-condensing mapping; thus by Sadovskii’s result, A+B
has a fixed point. 

Corollary 2.1 and Theorem 2.3 simply yield the following result, which will be
applied in the next section.
Corollary 2.2. Let X be a Banach space, C a closed bounded and convex sub-
set of X. If A : C → X is a (ϕ, F )-contraction with continuous F satisfying
lim inf s→t ϕ(s) > 0 for all t > 0, B : C → X a compact operator and (A + B)(C) ⊂
C, then A + B has a fixed point.

3. The applications of nonlinear F -contractions

In this section we show the applicability of the obtained results.

3.1. Integral equation of Volterra type. First we present the application of the
existence of fixed point for (ϕ, F )-contractions to the following equation of Volterra
type:
t
(3.1) x(t) = K(t, s, x(s))ds + h(t), t ∈ I,
0

where T > 0, I = [0, T ], K : I × I × R → R, h : I → R.

In order to obtain our claims, we will need the following assumptions:
(C1) The functions h, K are continuous.
(C2) There exists a strictly increasing sequence (αn )n∈N∪{0} satisfying α0 = 0,
αn ≥ 1, αn − αn−1 ≤ 1 for all n ∈ N, αn → ∞ such that for any n ∈ N,
αn
(3.2) |K(t, s, u) − K(t, s, v)| ≤ e−tαn |u − v|
1 + αn (αn − αn−1 )
for all s, t ∈ I and u, v ∈ R such that |u − v| < αn eT .

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 SOLVING EXISTENCE PROBLEMS VIA F -CONTRACTIONS 1593

Consider the Banach space C(I) of all continuous functions x : I → R equipped

with Bielecki’s norm:

x = sup e−t |x(t)|.

t∈I

Now, we are ready to enunciate our first existence result.

Theorem 3.1. If (C1) and (C2) are satisfied, then the nonlinear problem (3.1) has
a unique solution in C(I).

Proof. Consider the operator L : C(I) → C(I) as follows:

t
(Lx)(t) = K(t, s, x(s))ds + h(t), x ∈ C(I).
0

A fixed point of the operator L will be a solution of the equation (3.1). In order
to fulfil all the assumptions of Theorem 2.1 let us consider a function F (t) = −1/t,
t > 0, and ϕ : (0, ∞) → (0, ∞) of the form

−t + α1 , 0 < t < α1 ,
ϕ(t) =
−t + αn , αn−1 ≤ t < αn , n ≥ 2.

In this case one can calculate that the contraction condition (H3) takes the following
form:

x−y
(3.3) Tx − Ty ≤ ,
1 + x − y αn − x − y

for all x, y ∈ C(I) satisfying αn−1 ≤ x−y < αn when n ≥ 2 and 0 < x−y < α1
for n = 1.
We will show that L satisfies (3.3). Fix n ≥ 2 and take any x, y ∈ C(I) such
that αn−1 ≤ x − y < αn . Observe that for each s ∈ I we have

|x(s) − y(s)| ≤ es sup e−s |x(s) − y(s)| < es αn ≤ eT αn .

s∈I

Therefore, due to (C2), we obtain

t
|(Lx)(t) − (Ly)(t)| ≤ |K(t, s, x(s)) − K(t, s, y(s))|ds
0
t
αn
≤ e−tαn |x(s) − y(s)|ds, t ∈ I.
1 + αn (αn − αn−1 ) 0

Next, we see that

1 + x − y αn − x − y < 1 + αn (αn − αn−1 ),

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 1594 DARIUSZ WARDOWSKI

and, since αn+1 > 1, −sαn+1 ≤ −s for all s ∈ I. In consequence, the following
holds:
t
αn
|(Lx)(t) − (Ly)(t)| ≤ e−tαn |x(s) − y(s)|ds
1 + x − y αn − x − y 0
t
αn
= e−tαn |x(s) − y(s)|e−sαn+1 esαn+1 ds
1 + x − y αn − x − y 0

αn x−y t
≤ e−tαn esαn+1 ds
1+ x−y αn − x − y 0
αn x−y 1 tαn+1
< e−tαn e
1+ x−y αn − x − y αn+1
αn x − y
=  et(αn+1 −αn ) , t ∈ I.
αn+1 1 + x − y αn − x − y

Using again the properties of the sequence (αn ) we obtain

x−y
e−t |(Lx)(t) − (Ly)(t)| ≤ , t ∈ I.
1 + x − y αn − x − y
Taking the supremum with respect to t in the above inequality we obtain (3.3).
The analogous calculations are for n = 1. Theorem 2.1 ends the proof. 

3.2. Implicit integral equations. In this section we apply our result of Kras-
nosel’skii type to the general class of integral equations

 t 
(3.4) V (t, x(t)) = S t, H(t, s, x(s))ds ,
0

where V, S : [−α, α] × [−α, α] → R and H : [−α, α] × [−α, α] × [−α, α] → R are

continuous, α > 0. The equations of this type were studied e.g. in [2] and [4]. We
will look for the solution of (3.4) in a subset C of the Banach space X of continuous
functions ζ : [−β, β] → R, 0 < β < α, equipped with the supremum norm of the
form
C := {ζ ∈ X : ζ(0) = 0, ζ ≤ α}.
Theorem 3.2. If S(0, 0) = V (t, 0) = 0 for all t ∈ [−α, α] and the operator
(Aζ)(t) = ζ(t) − V (t, ζ(t)) is a (ϕ, F )-contraction on C with continuous F and
nonincreasing ϕ, then (3.4) has a solution in C.
Proof. Denote θ(t) = 0 for all t ∈ [−α, α]. We have (Aθ)(t) = θ(t) − V (t, θ(t)) = 0,
t ∈ [−α, α]. Taking any ζ ∈ C with Aζ = θ, by (H3), we have
F ( Aζ ) = F ( Aζ − Aθ ) ≤ F ( ζ ) − ϕ( ζ ) < F ( ζ ),
which, by (F1), gives
(3.5) Aζ < ζ .
We show that for any ζ ∈ C,
(3.6) Aζ ≤ γ < α for some γ > 0.

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 SOLVING EXISTENCE PROBLEMS VIA F -CONTRACTIONS 1595

In another case there exists a sequence (ζn ) ⊂ C such that ( Aζn ) is increasing
and limn→∞ Aζn = α. Since ζn ∈ C, (3.5) implies Aζn ≤ ζn ≤ α and thus
limn→∞ ζn = α. In consequence, we obtain
ϕ(α) ≤ ϕ( ζn ) ≤ F ( ζn ) − F ( Aζn ).
Tending with n → ∞ and using a continuity of F we get ϕ(α) ≤ 0, which is
impossible.
The rest of the proof goes as in Burton’s result, Theorem 3 in [2]. Defining a
t
mapping B : C → C by (Bζ)(t) = S(t, 0 H(t, s, ζ(s))ds) and using a continuity of
H on its compact domain, continuity of S and the fact that S(0, 0) = 0 we show
that there exists 0 < β < α such that ζ ∈ C and t < β imply |(Bζ)(t)| ≤ α − γ,
which together with (3.6) gives Aζ +Bζ ≤ α. A simple fact (Aζ)(0)+(Bζ)(0) = 0
finally gives (A+B)(ζ) ∈ C. The compactness of B is received by showing that BC
is an equicontinuous set (for details, see [2]). Finally, note that lim inf s→t+ ϕ(s) > 0
for all t ≥ 0, since ϕ is decreasing. Thus, all the assumptions of Corollary 2.2 are
satisfied. 
In many practical situations a common difficulty which may occur is when we
want to verify the contractivity of the operator A. However, when we reduce
our investigations to nonnegative solutions, we may obtain more comfortable and
applicable tools. Applying the analogous methods as in the proof of the above
result we can prove the following theorem.
Theorem 3.3. If S ≥ 0, S(0, 0) = V (t, 0) = 0 for all t ∈ [−α, α] and the operator
(Aζ)(t) = ζ(t) − V (t, ζ(t)) is a (ϕ, F )-contraction on C + of the form
C + := {ζ ∈ C : ζ ≥ 0}
with continuous F , nonincreasing ϕ and 0 ≤ V (t, ζ(t)) ≤ ζ(t) for every ζ ∈ C + ,
t ∈ [−α, α], then (3.4) has a solution in C + .
Proof. In light of the proof of Theorem 3.2 it is enough to observe that Aζ ∈ C +
and (A + B)(ζ) ≥ 0 for each ζ ∈ C + . 
We show the applicability of the above result by the following example.
Example 3.1. Consider the differential equation of the form
(3.7) 2xx (x + 1) = 2t + G(t, x) (2x + 1)2 ,
where G : [−α, α] × [−α, α] → R+ is continuous, α > 0. We will verify the existence
of a solution in C + . The equation (3.7) can be transformed into the form
2xx (x + 1)
− 2t = G(t, x),
(2x + 1)2
where x is evaluated at each t. Next, we obtain
t
2x(s)x (s)(x(s) + 1)
t x(t)
2u(u + 1)
t2 + G(s, x(s))ds = 2
ds = du
0 0 (2x(s) + 1) 0 (2u + 1)2
x(t)  
1 1 x(t) 2x(t)
= − 2
du = −
0 2 2(2u + 1) 2 4(2x(t) + 1)
x2 (t)
= .
2x(t) + 1

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 1596 DARIUSZ WARDOWSKI

Now, we can see that the equation (3.7) can be presented in the form (3.4). In
order to fulfil the assumptions of Theorem 3.2 we show that the operator
x2 x2 + x
Ax := x − =
2x + 1 2x + 1
is a (ϕ, F )-contraction on C + . Indeed, taking any x, y ∈ C + we have (in the
following x, y are evaluated in each t ∈ [−α, α])
|x − y|(x + y + 2xy + 1)
|Ax − Ay| = .
2x + 2y + 4xy + 1
Observe that
|x − y| ≤ x + y + 2xy,
which together with the fact that a function t → 1+t
1+2t , t ≥ 0, is decreasing gives us
|x − y|(1 + |x − y|)
|Ax − Ay| ≤ .
1 + 2|x − y|
t(1+t)
Next, using the increasing function t → 1+2t , t ≥ 0, we have
supt∈[0,α] |x(t)−y(t)| 1 + supt∈[0,α] |x(t)−y(t| 1+ x−y
Ax−Ay ≤ = x−y .
1+2 supt∈[0,α] |x(t) − y(t)| 1+2 x−y
The last inequality, if Ax = Ay, can be presented in the form
1 1 1
− ≤− .
x−y +1 Ax − Ay x−y
Now one can observe that the above contraction condition is a (ϕ, F )-contraction
for ϕ(s) = 1/(s + 1), s > 0, and F (s) = −1/s, s > 0.

References
[1] J. Banaś and K. Goebel, Measures of noncompactness in Banach spaces, Lecture Notes in
Pure and Applied Mathematics, vol. 60, Marcel Dekker, Inc., New York, 1980. MR591679
[2] T. A. Burton, Integral equations, implicit functions, and fixed points, Proc. Amer. Math. Soc.
124 (1996), no. 8, 2383–2390, DOI 10.1090/S0002-9939-96-03533-2. MR1346965
[3] T. A. Burton and C. Kirk, A fixed point theorem of Krasnoselskii-Schaefer type, Math. Nachr.
189 (1998), 23–31, DOI 10.1002/mana.19981890103. MR1492921
[4] C. Corduneanu, Integral equations and applications, Cambridge University Press, Cambridge,
1991. MR1109491
[5] G. Darbo, Punti uniti in trasformazioni a codominio non compatto (Italian), Rend. Sem.
Mat. Univ. Padova 24 (1955), 84–92. MR0070164
[6] J. Garcia-Falset, K. Latrach, E. Moreno-Gálvez, and M.-A. Taoudi, Schaefer-Krasnoselskii
fixed point theorems using a usual measure of weak noncompactness, J. Differential Equations
252 (2012), no. 5, 3436–3452, DOI 10.1016/j.jde.2011.11.012. MR2876659
[7] N. Hussain, A. Latif, and I. Iqbal, Fixed point results for generalized F -contractions in
modular metric and fuzzy metric spaces, Fixed Point Theory Appl., 2015:158, 20 pp., DOI
10.1186/s13663-015-0407-1. MR3392555
[8] N. Hussain and P. Salimi, Suzuki-Wardowski type fixed point theorems for α-GF-contractions,
Taiwanese J. Math. 18 (2014), no. 6, 1879–1895, DOI 10.11650/tjm.18.2014.4462.
MR3284036
[9] I. Iqbal and N. Hussain, Fixed point theorems for generalized multivalued nonlinear F-
contractions, J. Nonlinear Sci. Appl. 9 (2016), no. 11, 5870–5893. MR3584046
[10] I. Iqbal, N. Hussain, and N. Sultana, Fixed points of multivalued non-linear F -contractions
with application to solution of matrix equations, Filomat 31 (2017), no. 11, 3319-3333.

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 SOLVING EXISTENCE PROBLEMS VIA F -CONTRACTIONS 1597

[11] J. R. Jachymski, Equivalence of some contractivity properties over metrical structures,

Proc. Amer. Math. Soc. 125 (1997), no. 8, 2327–2335, DOI 10.1090/S0002-9939-97-03853-7.
MR1389524
[12] J. Jachymski, Equivalent conditions for generalized contractions on (ordered) metric spaces,
Nonlinear Anal. 74 (2011), no. 3, 768–774, DOI 10.1016/j.na.2010.09.025. MR2738628
[13] W. A. Kirk, Contraction mappings and extensions, Handbook of metric fixed point theory,
Kluwer Acad. Publ., Dordrecht, 2001, pp. 1–34. MR1904272
[14] D. Klim and D. Wardowski, Fixed points of dynamic processes of set-valued F -contractions
and application to functional equations, Fixed Point Theory Appl., 2015:22, 9 pp., DOI
10.1186/s13663-015-0272-y. MR3311515
[15] M. A. Krasnoselskiı̆, Some problems of nonlinear analysis, American Mathematical Society
Translations, Ser. 2, Vol. 10, American Mathematical Society, Providence, R.I., 1958, pp. 345–
409. MR0094731
[16] W. Kryszewski and J. Mederski, Fixed point index for Krasnosel skii-type set-valued maps
on complete ANRs, Topol. Methods Nonlinear Anal. 28 (2006), no. 2, 335–384. MR2289691
[17] Y. Liu and Z. Li, Krasnoselskii type fixed point theorems and applications, Proc. Amer. Math.
Soc. 136 (2008), no. 4, 1213–1220, DOI 10.1090/S0002-9939-07-09190-3. MR2367095
[18] G. Mınak, A. Helvacı, and I. Altun, Ćirić type generalized F -contractions on com-
plete metric spaces and fixed point results, Filomat 28 (2014), no. 6, 1143–1151, DOI
10.2298/FIL1406143M. MR3360088
[19] D. Paesano and C. Vetro, Multi-valued F -contractions in 0-complete partial metric spaces
with application to Volterra type integral equation, Rev. R. Acad. Cienc. Exactas Fı́s. Nat.
Ser. A Math. RACSAM 108 (2014), no. 2, 1005–1020, DOI 10.1007/s13398-013-0157-z.
MR3249990
[20] H. Piri and P. Kumam, Some fixed point theorems concerning F -contraction in complete
metric spaces, Fixed Point Theory Appl., 2014:210, 11 pp., DOI 10.1186/1687-1812-2014-
210. MR3357360
[21] B. Przeradzki, A generalization of Krasnosel’skii fixed point theorem for sums of compact
and contractible maps with application, Cent. Eur. J. Math. 10 (2012), no. 6, 2012–2018,
DOI 10.2478/s11533-012-0102-y. MR2983143
[22] B. E. Rhoades, A comparison of various definitions of contractive mappings, Trans. Amer.
Math. Soc. 226 (1977), 257–290, DOI 10.2307/1997954. MR0433430
[23] B. N. Sadovskiı̆, Limit-compact and condensing operators (Russian), Uspehi Mat. Nauk 27
(1972), no. 1(163), 81–146. MR0428132
[24] H. Schaefer, Über die Methode der a priori-Schranken (German), Math. Ann. 129 (1955),
415–416, DOI 10.1007/BF01362380. MR0071723
[25] N. A. Secelean, Weak F -contractions and some fixed point results, Bull. Iranian Math. Soc.
42 (2016), no. 3, 779–798. MR3518218
[26] N.-A. Secelean, Generalized F -iterated function systems on product of metric spaces, J. Fixed
Point Theory Appl. 17 (2015), no. 3, 575–595, DOI 10.1007/s11784-015-0235-2. MR3411806
[27] N.-A. Secelean, Iterated function systems consisting of F -contractions, Fixed Point Theory
Appl., 2013:277, 13 pp., DOI 10.1186/1687-1812-2013-277. MR3213104
[28] N.A. Secelean and D. Wardowski, New fixed point tools in non-metrizable spaces, Results.
Math. (2017), DOI: 10.1007/s00025-017-0688-2.
[29] N.-A. Secelean and D. Wardowski, ψF -contractions: not necessarily nonexpansive Picard
operators, Results Math. 70 (2016), no. 3-4, 415–431, DOI 10.1007/s00025-016-0570-7.
MR3544869
[30] M. Sgroi and C. Vetro, Multi-valued F -contractions and the solution of certain functional
and integral equations, Filomat 27 (2013), no. 7, 1259–1268, DOI 10.2298/FIL1307259S.
MR3243997
[31] T. Suzuki, Discussion of several contractions by Jachymski’s approach, Fixed Point Theory
Appl., 2016:91, 11 pp., DOI 10.1186/s13663-016-0581-9. MR3548683
[32] M. Turinici, Wardowski implicit contractions in metric spaces, (2013) arXiv:1212.3164v2
[Math.GN].
[33] X. Udo-utun, On inclusion of F -contractions in (δ, k)-weak contractions, Fixed Point Theory
Appl., posted on 2013, 2014:65, 6 pp., DOI 10.1186/1687-1812-2014-65. MR3345452

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use
 1598 DARIUSZ WARDOWSKI

[34] D. Wardowski and N. Van Dung, Fixed points of F -weak contractions on complete met-
ric spaces, Demonstr. Math. 47 (2014), no. 1, 146–155, DOI 10.2478/dema-2014-0012.
MR3200192
[35] D. Wardowski, Fixed points of a new type of contractive mappings in complete metric spaces,
Fixed Point Theory Appl., 2012:94, 6 pp., DOI 10.1186/1687-1812-2012-94. MR2949666
[36] K. Wlodarczyk, Periodic and fixed points of the Leader-type contractions in quasi-
triangular spaces, Fixed Point Theory Appl., 2016:85, 28 pp., DOI 10.1186/s13663-016-0575-7.
MR3538658

Department of Nonlinear Analysis, Faculty of Mathematics and Computer Science,

University of L
 ódź, Banacha 22, 90-238 L
 ódź, Poland
E-mail address: wardd@math.uni.lodz.pl

Licensed to New York Univ, Courant Inst. Prepared on Fri Feb 2 05:56:29 EST 2018 for download from IP 128.122.230.132.
License or copyright restrictions may apply to redistribution; see http://www.ams.org/journal-terms-of-use