We recall the following lemma.
Lemma 8 [8]
Let T:Y\to CB(X) and f:Y\to X be fweak compatible. If \{fw\}=Tw for some w\in Y and H(Tx,Ty)\le h(a\cdot L(x,y)+(1a)\cdot N(x,y)) for all x, y in Y, where 0<h<1, 0\le a\le 1, then fTw=Tfw.
We remark that the abovementioned lemma has been used in [8, 9] and [10] to prove the existence of fixed points of hybrid mappings. However, we have noticed some typos in its original statement which have been rectified in the above statement without altering the proof.
Next, we prove a fixed point result for hybrid mappings under a general integraltype contractivity condition. In contrast to [20], we avoid the complete character of the base space X, and we introduce hybrid mappings. With respect to the study in [21], we consider here occasionally coincidentally idempotent mappings.
Theorem 9 Let Y be an arbitrary nonempty set, (X,d) be a metric space, f:Y\to X and T:Y\to CB(X) be such that
that is, {\bigcup}_{y\in Y}T(y)\subseteq f(Y),
\begin{array}{c}\mathit{\text{there exists}}\phantom{\rule{0.1em}{0ex}}q\in (0,1)\phantom{\rule{0.1em}{0ex}}\mathit{\text{such that}}\hfill \\ {\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q{\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\phantom{\rule{1em}{0ex}}\mathit{\text{for all}}\phantom{\rule{0.1em}{0ex}}x,y\phantom{\rule{0.1em}{0ex}}\mathit{\text{in}}\phantom{\rule{0.1em}{0ex}}Y,\hfill \end{array}
(2)
f(Y)\phantom{\rule{0.1em}{0ex}}\mathit{\text{is complete}},
(3)
\psi :{\mathbb{R}}_{+}\to {\mathbb{R}}_{+} is a Lebesgue measurable mapping which is nonnegative, summable on each compact interval and such that
\psi (x)>0,\phantom{\rule{1em}{0ex}}\mathrm{\forall}x>0,
(4)
which trivially implies that
{\int}_{0}^{\u03f5}\psi (t)\phantom{\rule{0.2em}{0ex}}dt>0\phantom{\rule{1em}{0ex}}\mathit{\text{for each}}\phantom{\rule{0.1em}{0ex}}\u03f5>0
(5)
and
{\int}_{0}^{\u03f5}\psi (t)\phantom{\rule{0.2em}{0ex}}dt<{\int}_{0}^{\tilde{\u03f5}}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\phantom{\rule{1em}{0ex}}\mathit{\text{for each}}\phantom{\rule{0.1em}{0ex}}0<\u03f5<\tilde{\u03f5}.
(6)
Suppose also that
{\int}_{0}^{\mu \u03f5}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \gamma (\mu ){\int}_{0}^{\u03f5}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\phantom{\rule{1em}{0ex}}\mathit{\text{for each}}\phantom{\rule{0.1em}{0ex}}\mu >1\phantom{\rule{0.1em}{0ex}}\mathit{\text{and}}\phantom{\rule{0.1em}{0ex}}\u03f5>0,
(7)
where
\gamma :(1,+\mathrm{\infty})\u27f6{\mathbb{R}}_{+}
is such that
0<\gamma \left({q}^{1/2}\right)\cdot q<1
(8)
and
\gamma \left({q}^{1/2}\right)\cdot q\cdot \gamma \left(\frac{1}{\gamma ({q}^{1/2})\cdot q}\right)\le 1.
(9)
Then T and f have a coincidence point. Further, if f and T are occasionally coincidentally idempotent, then f and T have a common fixed point.
Proof In view of (1) and Nadler’s remark in [24], given the point {x}_{0}\in Y, we can construct two sequences \{{x}_{n}\} in Y and \{{y}_{n}\} in X such that, for each n\in \mathbb{N},
{y}_{n}=f{x}_{n}\in T{x}_{n1}\phantom{\rule{1em}{0ex}}\text{and}\phantom{\rule{1em}{0ex}}d({y}_{n},{y}_{n+1})\le {q}^{1/2}\cdot H(T{x}_{n1},T{x}_{n}).
Indeed, since T{x}_{0}\subseteq f(Y), there exists {x}_{1}\in Y such that f{x}_{1}={y}_{1}\in T{x}_{0}. Besides, given {y}_{1}\in T{x}_{0}, by Nadler’s remark in [24] and using that {q}^{1/2}>1, we can choose {y}_{2}\in T{x}_{1}\subseteq f(Y) such that d({y}_{1},{y}_{2})\le {q}^{1/2}\cdot H(T{x}_{0},T{x}_{1}) and {y}_{2}=f{x}_{2} for a certain {x}_{2}\in Y. The continuation of this process allows to construct the two abovementioned sequences \{{x}_{n}\} and \{{y}_{n}\} inductively.
We claim that \{{y}_{n}\} is a Cauchy sequence. Using the inequality in (2) and also property (7), which is trivially valid for \u03f5=0, it follows, for n\ge 2, that
\begin{array}{rcl}{\int}_{0}^{d(f{x}_{n1},f{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& \le & {\int}_{0}^{{q}^{1/2}H(T{x}_{n2},T{x}_{n1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & \gamma \left({q}^{1/2}\right){\int}_{0}^{H(T{x}_{n2},T{x}_{n1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & \gamma \left({q}^{1/2}\right)\cdot q{\int}_{0}^{L({x}_{n2},{x}_{n1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,\end{array}
where
\begin{array}{rcl}L({x}_{n2},{x}_{n1})& =& max\{d(f{x}_{n2},f{x}_{n1}),d(f{x}_{n2},T{x}_{n2}),d(f{x}_{n1},T{x}_{n1}),\\ \frac{1}{2}(d(f{x}_{n2},T{x}_{n1})+d(f{x}_{n1},T{x}_{n2}))\}\\ \le & max\{d(f{x}_{n2},f{x}_{n1}),d(f{x}_{n2},f{x}_{n1}),d(f{x}_{n1},f{x}_{n}),\frac{1}{2}d(f{x}_{n2},f{x}_{n})\}\\ \le & max\{d(f{x}_{n2},f{x}_{n1}),d(f{x}_{n1},f{x}_{n}),\frac{1}{2}(d(f{x}_{n2},f{x}_{n1})+d(f{x}_{n1},f{x}_{n}))\}\\ =& max\{d(f{x}_{n2},f{x}_{n1}),d(f{x}_{n1},f{x}_{n})\}.\end{array}
Suppose that
d(f{x}_{n1},f{x}_{n})>\lambda \cdot d(f{x}_{n2},f{x}_{n1})\phantom{\rule{1em}{0ex}}\text{for some}n\in \mathbb{N}\text{with}n\ge 2,
where \lambda =\gamma ({q}^{1/2})\cdot q\in (0,1), hence d(f{x}_{n1},f{x}_{n})>0 and
0<max\{d(f{x}_{n2},f{x}_{n1}),d(f{x}_{n1},f{x}_{n})\}<\frac{1}{\lambda}d(f{x}_{n1},f{x}_{n}),
so that
\begin{array}{rcl}{\int}_{0}^{d(f{x}_{n1},f{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& \le & \gamma \left({q}^{1/2}\right)\cdot q{\int}_{0}^{max\{d(f{x}_{n2},f{x}_{n1}),d(f{x}_{n1},f{x}_{n})\}}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ <& \gamma \left({q}^{1/2}\right)\cdot q{\int}_{0}^{\frac{1}{\lambda}d(f{x}_{n1},f{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & \gamma \left({q}^{1/2}\right)\cdot q\cdot \gamma \left(\frac{1}{\lambda}\right){\int}_{0}^{d(f{x}_{n1},f{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\int}_{0}^{d(f{x}_{n1},f{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,\end{array}
where we have also used (6) (a consequence of (4)), (7), (8) and (9). The previous inequalities imply that
{\int}_{0}^{d(f{x}_{n1},f{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt<{\int}_{0}^{d(f{x}_{n1},f{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,
which is a contradiction. In consequence,
d(f{x}_{n1},f{x}_{n})\le \lambda \cdot d(f{x}_{n2},f{x}_{n1}),\phantom{\rule{1em}{0ex}}\text{for every}n\in \mathbb{N},n\ge 2,
where \lambda =\gamma ({q}^{1/2})\cdot q\in (0,1), by hypothesis, and hence \{f{x}_{n}\} is a Cauchy sequence in f(Y). This is clear from the following inequality, valid for n,m\in \mathbb{N}, n>m,
\begin{array}{rcl}d(f{x}_{n},f{x}_{m})& \le & \sum _{j=m+1}^{n}d(f{x}_{j},f{x}_{j1})\le \sum _{j=m+1}^{n}{\lambda}^{j1}d(f{x}_{1},f{x}_{0})\\ =& \frac{{\lambda}^{m}{\lambda}^{n}}{1\lambda}d(f{x}_{1},f{x}_{0})\le \frac{{\lambda}^{m}}{1\lambda}d(f{x}_{1},f{x}_{0}),\end{array}
which tends to zero as m\to +\mathrm{\infty}.
Since f(Y) is complete, then the sequence \{f{x}_{n}\} has a limit in f(Y), say u. Let w\in {f}^{1}(u) and prove that fw\in Tw.
Suppose that fw\notin Tw, then, by (2), we have
{\int}_{0}^{d(f{x}_{n+1},Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\int}_{0}^{H(T{x}_{n},Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q{\int}_{0}^{L({x}_{n},w)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,
where
\begin{array}{rcl}L({x}_{n},w)& =& max\{d(f{x}_{n},fw),d(f{x}_{n},T{x}_{n}),d(fw,Tw),\frac{1}{2}(d(f{x}_{n},Tw)+d(fw,T{x}_{n}))\}\\ =& d(fw,Tw)\phantom{\rule{1em}{0ex}}\text{for}n\text{large}.\end{array}
Here, we have used that d(f{x}_{n},fw)=d(f{x}_{n},u)\to 0, as n\to +\mathrm{\infty}, d(f{x}_{n},T{x}_{n})\le d(f{x}_{n},f{x}_{n+1})\to 0, as n\to +\mathrm{\infty}, d(fw,Tw)>0 due to fw\notin Tw and Tw closed, and
\begin{array}{rcl}\frac{1}{2}(d(f{x}_{n},Tw)+d(fw,T{x}_{n}))& \le & \frac{1}{2}(2d(f{x}_{n},fw)+d(fw,Tw)+d(f{x}_{n},T{x}_{n}))\\ \to & \frac{1}{2}d(fw,Tw),\phantom{\rule{1em}{0ex}}\text{as}n\to +\mathrm{\infty}.\end{array}
Hence, for n large enough, we have
{\int}_{0}^{d(f{x}_{n+1},Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q{\int}_{0}^{d(fw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.
Making n tend to +∞ in the previous inequality, we have
{\int}_{0}^{d(fw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q{\int}_{0}^{d(fw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt
and, therefore, since q<1 and d(fw,Tw)>0, we get {\int}_{0}^{d(fw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt<{\int}_{0}^{d(fw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt, which is a contradiction. Hence fw\in Tw, that is, w is a coincidence point for T and f.
Although this fact is not relevant to the proof, we note that H(T{x}_{n1},T{x}_{n})\to 0 since
\underset{n\to \mathrm{\infty}}{lim}d({y}_{n1},{y}_{n})=0.
Indeed,
{\int}_{0}^{H(T{x}_{n1},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q{\int}_{0}^{L({x}_{n1},{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,
where
\begin{array}{rcl}L({x}_{n1},{x}_{n})& \le & max\{d(f{x}_{n1},f{x}_{n}),d(f{x}_{n},f{x}_{n+1})\}\\ \le & max\{d(f{x}_{n1},f{x}_{n}),\lambda d(f{x}_{n1},f{x}_{n})\}=d(f{x}_{n1},f{x}_{n}),\end{array}
therefore
{\int}_{0}^{H(T{x}_{n1},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q{\int}_{0}^{d(f{x}_{n1},f{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.
Then {lim}_{n\to +\mathrm{\infty}}{\int}_{0}^{H(T{x}_{n1},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt=0 and, by the properties of ψ, we get H(T{x}_{n1},T{x}_{n})\to 0 as n\to +\mathrm{\infty}. From the definition of \{{y}_{n}\}, we deduce that d(f{x}_{n},T{x}_{n})\le H(T{x}_{n1},T{x}_{n}) for every n and, therefore, {lim}_{n\to \mathrm{\infty}}d(f{x}_{n},T{x}_{n})=0, so that \{{x}_{n}\} is asymptotically Tregular with respect to f. However, this property can be deduced directly from the fact that
0\le d(f{x}_{n},T{x}_{n})\le d(f{x}_{n},f{x}_{n+1})\to 0\phantom{\rule{1em}{0ex}}\text{as}n\to +\mathrm{\infty}.
Now, if f and T are occasionally coincidentally idempotent, then ffw=fw for some w\in C(T,f). Then we have
{\int}_{0}^{H(Tfw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q{\int}_{0}^{L(fw,w)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,
(10)
where
\begin{array}{rcl}L(fw,w)& =& max\{d(ffw,fw),d(ffw,Tfw),d(fw,Tw),\frac{1}{2}(d(ffw,Tw)+d(fw,Tfw))\}\\ =& max\{d(fw,fw),d(fw,Tfw),d(fw,Tw),\frac{1}{2}(d(fw,Tw)+d(fw,Tfw))\}\\ =& d(fw,Tfw)\le H(Tw,Tfw).\end{array}
If Tfw\ne Tw, then from inequality (10) and using (5) (which is guaranteed by (4)), we have that
{\int}_{0}^{H(Tfw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q{\int}_{0}^{H(Tfw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt<{\int}_{0}^{H(Tfw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,
which is a contradiction. Hence Tfw=Tw. Thus we have fw=ffw and fw\in Tw=Tfw, i.e., fw is a common fixed point of f and T. □
Let Φ denote the family of maps ϕ from the set {\mathbb{R}}_{+} of nonnegative real numbers to itself such that
\varphi (t)\le qt\phantom{\rule{1em}{0ex}}\text{for all}t\ge 0\text{and for some}q\in (0,1).
(11)
Corollary 10 Let Y be an arbitrary nonempty set, (X,d) be a metric space, f:Y\to X and T:Y\to CB(X) be such that T(Y)\subseteq f(Y),
{\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \varphi ({\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt)
(12)
for all x, y in Y, where \varphi \in \mathrm{\Phi} (satisfying (11) for a certain q\in (0,1)),
f(Y)\phantom{\rule{0.1em}{0ex}}\mathit{\text{is complete}},
\psi :{\mathbb{R}}_{+}\to {\mathbb{R}}_{+} is a Lebesgue measurable mapping which is nonnegative, summable on each compact interval and such that (4) holds. Suppose also that (7), (8) and (9) hold for a certain \gamma :(1,+\mathrm{\infty})\u27f6{\mathbb{R}}_{+} and q determined by (11). Then T and f have a coincidence point. Further, if f and T are occasionally coincidentally idempotent, then f and T have a common fixed point.
Proof It is a consequence of Theorem 9 since (11) and (12) imply that
{\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \varphi ({\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt)\le q{\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt
for all x, y in Y and q\in (0,1). □
Remark 11 The condition
{\int}_{0}^{\mu \u03f5}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \mu {\int}_{0}^{\u03f5}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\phantom{\rule{1em}{0ex}}\text{for each}\mu 1\text{and}\u03f50
(13)
implies the validity of hypothesis (7) in Theorem 9 for the particular case of γ the identity mapping. Moreover, for 0<q<1, hypotheses (8) and (9) are trivially satisfied for this choice of γ. Indeed, using that 0<q<1, we get
0<\gamma \left({q}^{1/2}\right)\cdot q={q}^{1/2}\cdot q={q}^{1/2}<1
and
\gamma \left({q}^{1/2}\right)\cdot q\cdot \gamma \left(\frac{1}{\gamma ({q}^{1/2})\cdot q}\right)={q}^{1/2}q\frac{1}{{q}^{1/2}q}=1.
Remark 12 Assuming (8), condition (9) is trivially valid if \lambda \cdot \gamma (\frac{1}{\lambda})\le 1 for every \lambda \in (0,1) or, equivalently, \gamma (\frac{1}{\lambda})\le \frac{1}{\lambda} for every \lambda \in (0,1), that is, \gamma (z)\le z for every z>1. Note that this last condition is trivially valid for γ the identity mapping. Moreover, if \gamma (z)\le z for every z>1, then \gamma (z)<{z}^{2} for every z>1 and, therefore, if q\in (0,1), then \gamma ({q}^{1/2})<{q}^{1}, obtaining (8) if \gamma ({q}^{1/2})>0.
Remark 13 According to Remark 12, for q\in (0,1) fixed and ψ satisfying (4), an admissible function γ can be obtained by taking
\gamma (z)\ge \underset{\u03f5>0}{sup}\frac{{\int}_{0}^{z\u03f5}\psi (t)\phantom{\rule{0.2em}{0ex}}dt}{{\int}_{0}^{\u03f5}\psi (t)\phantom{\rule{0.2em}{0ex}}dt},\phantom{\rule{1em}{0ex}}z>1,
provided that \gamma ({q}^{1/2})>0 and \gamma (z)\le z for every z>1.
Example 14 Taking ψ as the constant function \psi (t)=K>0, t>0, in the statement of Theorem 9, condition (7) is reduced to
K\mu \u03f5\le \gamma (\mu )K\u03f5\phantom{\rule{1em}{0ex}}\text{for each}\mu 1\text{and}\u03f50,
so that we must choose γ as a nonnegative function satisfying that \gamma (z)=z for z>1 (obviously, \gamma ({q}^{1/2})>0 since q\in (0,1)) in order to guarantee conditions (7), (8) and (9).
Example 15 A simple calculation provides that, for the function \psi (t)=t, t>0, condition (7) is written as \gamma (z)\ge {z}^{2} for z>1 and, therefore, in this case condition (8) is never fulfilled. If we take \psi (t)=K{t}^{m}, t>0, for K>0 and m>0 fixed, then (7) implies that \gamma (z)\ge {z}^{m+1}>z for z>1.
Example 16 Now, we choose \psi (t)=K{t}^{m}, t>0, where K>0 and 1<m<0 are fixed. Note that the case m=0 has already been studied in Example 14. In this case 1<m<0, condition (7) is reduced to
K\frac{{(\mu \u03f5)}^{m+1}}{m+1}\le \gamma (\mu )K\frac{{\u03f5}^{m+1}}{m+1}\phantom{\rule{1em}{0ex}}\text{for}\mu 1\text{and}\u03f50,
which is equivalent to \gamma (z)\ge {z}^{m+1} for z>1. Note that this inequality implies, for 0<q<1, that \gamma ({q}^{1/2})>0. If we add the hypothesis \gamma (z)\le z for z>1, then we guarantee the validity of conditions (8) and (9) due to Remark 12. Hence, we can take any nonnegative function γ satisfying that
{z}^{m+1}\le \gamma (z)\le z\phantom{\rule{1em}{0ex}}\text{for}z1.
Of course, \gamma (z)=z and \gamma (z)={z}^{m+1} are valid choices.
Example 17 Take \psi (t)={e}^{t}, t>0. Condition (7) is equivalent to
1{e}^{\mu \u03f5}\le \gamma (\mu )(1{e}^{\u03f5})\phantom{\rule{1em}{0ex}}\text{for}\mu 1\text{and}\u03f50,
that is,
\gamma (\mu )\ge \frac{1{e}^{\mu \u03f5}}{1{e}^{\u03f5}}\phantom{\rule{1em}{0ex}}\text{for}\mu 1\text{and}\u03f50.
Now, for each z>1 fixed, we calculate {sup}_{\u03f5>0}\frac{1{e}^{z\u03f5}}{1{e}^{\u03f5}}, which is obviously positive, and we check that its value is equal to z.
It is easy to prove that for z>1 fixed, the function \u03f5\in (0,+\mathrm{\infty})\u27f6{\mathcal{R}}_{z}(\u03f5)=\frac{1{e}^{z\u03f5}}{1{e}^{\u03f5}} is decreasing on (0,+\mathrm{\infty}). Indeed, the sign of its derivative coincides with the sign of the function \nu (\u03f5)=z{e}^{z\u03f5}(1{e}^{\u03f5})(1{e}^{z\u03f5}){e}^{\u03f5} and also with the sign of \tau (\u03f5)=z{e}^{z\u03f5}({e}^{\u03f5}1)+{e}^{z\u03f5}1 for \u03f5\in (0,+\mathrm{\infty}). Now, the function τ is strictly negative on (0,+\mathrm{\infty}) since \tau (0)=\tau ({0}^{+})=0 and {\tau}^{\prime}(\u03f5)=z(1z){e}^{z\u03f5}({e}^{\u03f5}1)<0 for \u03f5>0.
Moreover, {lim}_{\u03f5\to {0}^{+}}{\mathcal{R}}_{z}(\u03f5)=z for each z>1; in consequence, {sup}_{\u03f5>0}{\mathcal{R}}_{z}(\u03f5)=z for every z>1. Therefore, if \gamma (z)\ge z for every z>1, then (7) follows. Note also that if q\in (0,1), then \gamma ({q}^{1/2})>0. Finally, for q\in (0,1), if we take \gamma :(1,+\mathrm{\infty})\u27f6{\mathbb{R}}_{+} such that \gamma (z)=z for z>1, we deduce the validity of (7), (8) and (9).
The following example shows that Theorem 9 is a proper generalization of the fixed point results in [7–10].
Example 18 Let X={\mathbb{R}}_{+} be endowed with the Euclidean metric, let f:X\to X and T:X\to CB(X) be defined by fx=4({x}^{2}+x) and Tx=[0,{x}^{2}+7]. Let \varphi :{\mathbb{R}}_{+}\to {\mathbb{R}}_{+} be defined by \varphi (t)=\frac{1}{4}t for all t\in {\mathbb{R}}_{+}. Then mappings f and T are not commuting and also do not satisfy any of its generalizations, viz weakly commuting, compatibility, weak compatibility. Also the mappings f and T are not coincidentally commuting. Note that f1\in T1, but ff1\ne f1 and so f and T are not coincidentally idempotent, but f0\in T0 and ff0=f0 thus f and T are occasionally coincidentally idempotent. For all x and y in X, we have
\begin{array}{rcl}{\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& =& {\int}_{0}^{{x}^{2}{y}^{2}}\psi (t)\phantom{\rule{0.2em}{0ex}}dt={\int}_{0}^{(\frac{x+y}{4})\cdot \frac{1}{(x+y+1)}\cdot (4xy\cdot (x+y+1))}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ =& {\int}_{0}^{(\frac{x+y}{4})\cdot \frac{1}{(x+y+1)}\cdot (4{x}^{2}{y}^{2}+xy)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\int}_{0}^{\frac{1}{4}d(fx,fy)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & \frac{1}{4}{\int}_{0}^{d(fx,fy)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \frac{1}{4}{\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt=\varphi ({\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt).\end{array}
Note that these inequalities are valid if
{\int}_{0}^{\frac{1}{4}d(fx,fy)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \frac{1}{4}{\int}_{0}^{d(fx,fy)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,
which is satisfied taking, for instance, the constant function \psi \equiv 1. On the other hand, γ is chosen as the identity map and it satisfies (8) and (9).
Note that 0 is a common fixed point of f and T. We remark that the results of [7–9] and [10] cannot be applied to these mappings f and T.
Theorem 19 In Theorem 9, we can assume, instead of condition (2), one of the inequalities
{\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q(a{\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+b{\int}_{0}^{N(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt)\phantom{\rule{1em}{0ex}}\mathit{\text{for all}}\phantom{\rule{0.1em}{0ex}}x,y\phantom{\rule{0.1em}{0ex}}\mathit{\text{in}}\phantom{\rule{0.1em}{0ex}}Y,
(14)
or
{\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q{\int}_{0}^{aL(x,y)+bN(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\phantom{\rule{1em}{0ex}}\mathit{\text{for all}}\phantom{\rule{0.1em}{0ex}}x,y\phantom{\rule{0.1em}{0ex}}\mathit{\text{in}}\phantom{\rule{0.1em}{0ex}}Y,
(15)
where a,b\ge 0, a+b\le 1 and q\in (0,1).
Similarly, in Corollary 10, we can consider one of the contractivity conditions
{\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \varphi (a{\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+b{\int}_{0}^{N(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt)\phantom{\rule{1em}{0ex}}\mathit{\text{for all}}\phantom{\rule{0.1em}{0ex}}x,y\phantom{\rule{0.1em}{0ex}}\mathit{\text{in}}\phantom{\rule{0.1em}{0ex}}Y,
(16)
or
{\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \varphi ({\int}_{0}^{aL(x,y)+bN(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt)\phantom{\rule{1em}{0ex}}\mathit{\text{for all}}\phantom{\rule{0.1em}{0ex}}x,y\phantom{\rule{0.1em}{0ex}}\mathit{\text{in}}\phantom{\rule{0.1em}{0ex}}Y,
(17)
where a,b\ge 0, a+b\le 1 and \varphi \in \mathrm{\Phi} (satisfying (11) for a certain q\in (0,1)) and the conclusion follows.
Proof It follows from the inequality
N(x,y)\le L(x,y)\phantom{\rule{1em}{0ex}}\text{for every}x,y
and the nonnegative character of a, b and ψ. Indeed, {d}^{2}(fx,fy)\le {[L(x,y)]}^{2},
\begin{array}{c}d(fx,Tx)\cdot d(fy,Ty)\le {[L(x,y)]}^{2},\hfill \\ d(fx,Ty)\cdot d(fy,Tx)\le \frac{1}{4}{(d(fx,Ty)+d(fy,Tx))}^{2}\le {[L(x,y)]}^{2},\hfill \\ \frac{1}{2}d(fx,Tx)\cdot d(fy,Tx)\le {[max\{d(fx,Tx),\frac{1}{2}(d(fx,Ty)+d(fy,Tx))\}]}^{2}\le {[L(x,y)]}^{2},\hfill \\ \frac{1}{2}d(fx,Ty)\cdot d(fy,Ty)\le {[max\{d(fy,Ty),\frac{1}{2}(d(fx,Ty)+d(fy,Tx))\}]}^{2}\le {[L(x,y)]}^{2},\hfill \end{array}
hence, for instance,
a{\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+b{\int}_{0}^{N(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le (a+b){\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\int}_{0}^{L(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.
Note that, in cases (16) and (17), it is not necessary to assume the nondecreasing character of the function ϕ since, using that \varphi \in \mathrm{\Phi}, we deduce (14) and (15), respectively. □
Of course, the function \varphi \equiv 0 is admissible in the results of this paper.
Note that, taking a=1 and b=0 in the inequalities of Theorem 19, we obtain the corresponding contractivity conditions of Theorem 9 and Corollary 10. On the other hand, taking a=0 and b=1 in Theorem 19, we have the following results, which are also corollaries of Theorem 9.
Corollary 20 Let Y be an arbitrary nonempty set, (X,d) be a metric space, f:Y\to X and T:Y\to CB(X) be such that conditions (1), (3) hold and
{\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le q{\int}_{0}^{N(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\phantom{\rule{1em}{0ex}}\mathit{\text{for all}}\phantom{\rule{0.1em}{0ex}}x,y\in Y,
(18)
where 0<q<1 and \psi :{\mathbb{R}}_{+}\to {\mathbb{R}}_{+} is a Lebesgue measurable mapping which is nonnegative, summable on each compact interval and such that (4) holds. Assume also that (7), (8) and (9) are fulfilled for a certain \gamma :(1,+\mathrm{\infty})\u27f6{\mathbb{R}}_{+}. Then f and T have a coincidence point. Further, if f and T are occasionally coincidentally idempotent, then f and T have a common fixed point.
Corollary 21 Let Y be an arbitrary nonempty set, (X,d) be a metric space, f:Y\to X and T:Y\to CB(X) be such that conditions (1), (3) hold and
{\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \varphi ({\int}_{0}^{N(x,y)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt)\phantom{\rule{1em}{0ex}}\mathit{\text{for all}}\phantom{\rule{0.1em}{0ex}}x,y\in Y,
(19)
where \varphi \in \mathrm{\Phi} (satisfying (11) for q\in (0,1)) and \psi :{\mathbb{R}}_{+}\to {\mathbb{R}}_{+} is a Lebesgue measurable mapping which is nonnegative, summable on each compact interval and such that (4) holds. Assume also that (7), (8) and (9) are fulfilled for a certain \gamma :(1,+\mathrm{\infty})\u27f6{\mathbb{R}}_{+}. Then f and T have a coincidence point. Further, if f and T are occasionally coincidentally idempotent, then f and T have a common fixed point.
Let \eta :[0,\mathrm{\infty})\to [0,1) be a function having the following property (see, for instance, [6, 25]):
(\mathcal{P}) For t\ge 0, there exist \delta (t)>0, s(t)<1 such that 0\le rt<\delta (t) implies \eta (r)\le s(t).
This property obviously holds if η is continuous since η attains its maximum (less than 1) on each compact [t,t+\delta (t)].
Definition 22 A sequence \{{x}_{n}\} is said to be asymptotically Tregular with respect to f if {lim}_{n\to \mathrm{\infty}}d(f{x}_{n},T{x}_{n})=0.
The following theorem is related to the main results of Hu [[25], Theorem 2], Jungck [14], Kaneko [26], Nadler [[24], Theorem 5] and Beg and Azam [[6], Theorem 5.4 and Corollary 5.5].
Theorem 23 Let Y be an arbitrary nonempty set, (X,d) be a metric space, f:Y\to X and T:Y\to CL(X) be such that condition (1) holds and
{\int}_{0}^{H(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt<\eta (d(fx,fy)){\int}_{0}^{d(fx,fy)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt
(20)
for all x,y\in Y, where \eta :[0,\mathrm{\infty})\to [0,1) satisfies (\mathcal{P}) and \psi \ge 0 is nonincreasing.
Suppose also that Tx is a compact set for every x\in Y.
If f(Y) is complete, then

(i)
there exists an asymptotically Tregular sequence \{{x}_{n}\} with respect to f in Y,

(ii)
f and T have a coincidence point.
Further, if f and T are occasionally coincidentally idempotent, then f and T have a common fixed point.
Proof For some {x}_{0} in Y, let {y}_{0}=f{x}_{0} and choose {x}_{1} in Y such that {y}_{1}=f{x}_{1}\in T{x}_{0}. Then, by (20), we have
{\int}_{0}^{H(T{x}_{0},T{x}_{1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt<\eta (d(f{x}_{0},f{x}_{1})){\int}_{0}^{d(f{x}_{0},f{x}_{1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.
Using (1), we can choose {x}_{2}\in Y such that {y}_{2}=f{x}_{2}\in T{x}_{1} and satisfying that
d({y}_{1},{y}_{2})=d(f{x}_{1},{y}_{2})=d(f{x}_{1},T{x}_{1})\le H(T{x}_{0},T{x}_{1}),
hence
\begin{array}{rcl}{\int}_{0}^{d({y}_{1},{y}_{2})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& =& {\int}_{0}^{d(f{x}_{1},f{x}_{2})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt<\eta (d(f{x}_{0},f{x}_{1})){\int}_{0}^{d(f{x}_{0},f{x}_{1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & {\int}_{0}^{d(f{x}_{0},f{x}_{1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\end{array}
Note that, in the previous inequalities, we have used that d(f{x}_{0},f{x}_{1})>0. If d(f{x}_{0},f{x}_{1})=0, then f{x}_{0}=f{x}_{1}\in T{x}_{0} and \{{x}_{n}\} is asymptotically Tregular with respect to f.
By induction, we construct a sequence \{{x}_{n}\} in Y and \{{y}_{n}\} in f(Y) such that, for every n,
d(f{x}_{n1},{y}_{n})=d(f{x}_{n1},T{x}_{n1})=\underset{y\in T{x}_{n1}}{min}d(f{x}_{n1},y)\le H(T{x}_{n2},T{x}_{n1}),
and {y}_{n}=f{x}_{n}\in T{x}_{n1}.
Also, we have
\begin{array}{c}{\int}_{0}^{d({y}_{n+1},{y}_{n+2})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}={\int}_{0}^{d(f{x}_{n+1},f{x}_{n+2})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}={\int}_{0}^{d(f{x}_{n+1},T{x}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\int}_{0}^{H(T{x}_{n},T{x}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le \eta (d(f{x}_{n},f{x}_{n+1})){\int}_{0}^{d(f{x}_{n},f{x}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}<{\int}_{0}^{d(f{x}_{n},f{x}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt={\int}_{0}^{d({y}_{n},{y}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\hfill \end{array}
It follows that the sequence \{d({y}_{n},{y}_{n+1})\} is decreasing and converges to its greatest lower bound, say t. Clearly t\ge 0. If t>0, then by the property (\mathcal{P}) of η, there will exist \delta (t)>0 and s(t)<1 such that
0\le rt<\delta (t)\phantom{\rule{1em}{0ex}}\text{implies}\phantom{\rule{1em}{0ex}}\eta (r)\le s(t).
For this \delta (t)>0, there exists N\in \mathbb{N} such that 0\le d({y}_{n},{y}_{n+1})t<\delta (t), whenever n\ge N. Hence \eta (d({y}_{n},{y}_{n+1}))\le s(t), whenever n\ge N. Let K=max\{\eta (d({y}_{0},{y}_{1})),\eta (d({y}_{1},{y}_{2})),\dots ,\eta (d({y}_{N1},{y}_{N})),s(t)\}. Then for n=1,2,3,\dots , we have
\begin{array}{rcl}{\int}_{0}^{d({y}_{n},{y}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& <& \eta (d({y}_{n1},{y}_{n})){\int}_{0}^{d({y}_{n1},{y}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & K{\int}_{0}^{d({y}_{n1},{y}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & {K}^{n}{\int}_{0}^{d({y}_{0},{y}_{1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\to 0\phantom{\rule{1em}{0ex}}\text{as}n\to \mathrm{\infty},\end{array}
which contradicts the assumption that t>0. Thus {lim}_{n\to \mathrm{\infty}}d({y}_{n},{y}_{n+1})=0; i.e., d(f{x}_{n},T{x}_{n})\to 0 as n\to +\mathrm{\infty}. Hence the sequence \{{x}_{n}\} is asymptotically Tregular with respect to f.
We claim that \{f{x}_{n}\} is a Cauchy sequence. Let n,m\in \mathbb{N} with n<m, then, by the nonincreasing character of ψ, we get
\begin{array}{c}{\int}_{0}^{d({y}_{n},{y}_{m})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le {\int}_{0}^{d({y}_{n},{y}_{n+1})+d({y}_{n+1},{y}_{n+2})+\cdots +d({y}_{m1},{y}_{m})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt={\int}_{0}^{d({y}_{n},{y}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\phantom{\rule{1em}{0ex}}+{\int}_{d({y}_{n},{y}_{n+1})}^{d({y}_{n},{y}_{n+1})+d({y}_{n+1},{y}_{n+2})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+\cdots +{\int}_{d({y}_{n},{y}_{n+1})+d({y}_{n+1},{y}_{n+2})+\cdots +d({y}_{m2},{y}_{m1})}^{d({y}_{n},{y}_{n+1})+d({y}_{n+1},{y}_{n+2})+\cdots +d({y}_{m1},{y}_{m})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le {\int}_{0}^{d({y}_{n},{y}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\int}_{0}^{d({y}_{n+1},{y}_{n+2})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+\cdots +{\int}_{0}^{d({y}_{m1},{y}_{m})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}=\sum _{i=n}^{m1}{\int}_{0}^{d({y}_{i},{y}_{i+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\hfill \end{array}
Now, we recall that
{\int}_{0}^{d({y}_{n+1},{y}_{n+2})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \eta (d({y}_{n},{y}_{n+1})){\int}_{0}^{d({y}_{n},{y}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt
for every n, which implies that
\begin{array}{rcl}{\int}_{0}^{d({y}_{n+2},{y}_{n+3})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& \le & \eta (d({y}_{n+1},{y}_{n+2})){\int}_{0}^{d({y}_{n+1},{y}_{n+2})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & \eta (d({y}_{n+1},{y}_{n+2}))\eta (d({y}_{n},{y}_{n+1})){\int}_{0}^{d({y}_{n},{y}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\end{array}
Following this procedure, we prove that
{\int}_{0}^{d({y}_{j},{y}_{j+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \prod _{i=n}^{j1}\eta (d({y}_{i},{y}_{i+1})){\int}_{0}^{d({y}_{n},{y}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,\phantom{\rule{1em}{0ex}}\text{for every}j=n+1,\dots ,m1.
Therefore,
\begin{array}{rcl}{\int}_{0}^{d({y}_{n},{y}_{m})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& \le & \sum _{i=n}^{m1}{\int}_{0}^{d({y}_{i},{y}_{i+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ =& {\int}_{0}^{d({y}_{n},{y}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+\sum _{i=n+1}^{m1}{\int}_{0}^{d({y}_{i},{y}_{i+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & [1+\sum _{i=n+1}^{m1}\prod _{l=n}^{i1}\eta (d({y}_{l},{y}_{l+1}))]{\int}_{0}^{d({y}_{n},{y}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\end{array}
We check that the righthand side in the last inequality tends to 0 as n,m\to +\mathrm{\infty}. Since {\int}_{0}^{d({y}_{n},{y}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\to 0 as n\to +\mathrm{\infty}, it suffices to show that {\sum}_{i=n+1}^{m1}{\prod}_{l=n}^{i1}\eta (d({y}_{l},{y}_{l+1})) is bounded (uniformly on n, m). Indeed, we check that {\sum}_{i=n+1}^{m1}{\prod}_{l=n}^{i1}\eta ({z}_{l}) is bounded for any sequence \{{z}_{l}\} with nonnegative terms and tending to 0 as l\to +\mathrm{\infty}, using the property (\mathcal{P}) of the function η. Given t=0, by (\mathcal{P}), there exist \delta (0)>0, {s}_{0}<1 such that 0\le r<\delta (0) implies \eta (r)\le {s}_{0}. Since \{{z}_{l}\}\to 0, given \delta (0)>0, there exists {l}_{0}\in \mathbb{N} such that, for every l\ge {l}_{0}, we have 0\le {z}_{l}<\delta (0). This implies that \eta ({z}_{l})\le {s}_{0} for every l\ge {l}_{0}.
In consequence, for n\ge {l}_{0}, we get
\begin{array}{rcl}0& \le & \sum _{i=n+1}^{m1}\prod _{l=n}^{i1}\eta ({z}_{l})\le \sum _{i=n+1}^{m1}\prod _{l=n}^{i1}{s}_{0}=\sum _{i=n+1}^{m1}{({s}_{0})}^{in}\\ =& \frac{{s}_{0}{({s}_{0})}^{mn}}{1{s}_{0}}<\frac{{s}_{0}}{1{s}_{0}},\end{array}
and this expression is bounded independently of m, n.
Hence \{f{x}_{n}\} is a Cauchy sequence in f(Y). Since f(Y) is complete, \{f{x}_{n}\} converges to some p in f(Y). Let z\in {f}^{1}(p). Then fz=p. Next, we have
\begin{array}{rcl}{\int}_{0}^{d(fz,Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& \le & {\int}_{0}^{d(f{x}_{n+1},fz)+d(f{x}_{n+1},Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ =& {\int}_{0}^{d(f{x}_{n+1},Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\int}_{d(f{x}_{n+1},Tz)}^{d(f{x}_{n+1},Tz)+d(fz,f{x}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & {\int}_{0}^{d(f{x}_{n+1},Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\int}_{0}^{d(fz,f{x}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & {\int}_{0}^{H(T{x}_{n},Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\int}_{0}^{d(fz,f{x}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & \eta (d(f{x}_{n},fz)){\int}_{0}^{d(f{x}_{n},fz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\int}_{0}^{d(fz,f{x}_{n+1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\end{array}
Letting n\to \mathrm{\infty}, we get {\int}_{0}^{d(fz,Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le 0. Thus we have d(fz,Tz)=0. Hence fz\in Tz.
Now, if f and T are occasionally coincidentally idempotent, then ffw=fw for some w\in C(T,f). Then we have
{\int}_{0}^{H(Tfw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le \eta (d(ffw,fw)){\int}_{0}^{d(ffw,fw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt=0.
Thus, Tfw=Tw. It follows that ffw=fw\in Tw=Tfw. Hence, fw is a common fixed point of T and f. □
Now we state some fixed point theorems for Kannantype multivalued mappings which extend and generalize the corresponding results of Shiau et al. [10] and Beg and Azam [6, 27]. A proper blend of the proof of Theorem 9 and those of [[10], Th. 6, Th. 7, Th. 8 respectively] and [[9], Theorems 3.1, 3.2, 3.3] will complete the proof.
Theorem 24 Let Y be an arbitrary nonempty set, (X,d) be a metric space, f:Y\to X and T:Y\to CB(X) be such that (1) holds and
\begin{array}{c}{\int}_{0}^{{H}^{r}(Tx,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le {\alpha}_{1}(d(fx,Tx)){\int}_{0}^{{d}^{r}(fx,Tx)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\alpha}_{2}(d(fy,Ty)){\int}_{0}^{{d}^{r}(fy,Ty)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \end{array}
(21)
for all x,y\in Y, where {\alpha}_{i}:{\mathbb{R}}_{+}\to [0,1) (i=1,2) are bounded on bounded sets, r is some fixed positive real number and \psi :{\mathbb{R}}_{+}\to {\mathbb{R}}_{+} is a Lebesgue measurable mapping which is summable on each compact interval and {\int}_{0}^{\u03f5}\psi (t)\phantom{\rule{0.2em}{0ex}}dt>0 for each \u03f5>0. Suppose that there exists an asymptotically Tregular sequence \{{x}_{n}\} with respect to f in Y. If T(Y) is complete or
\mathit{\text{there exists}}\phantom{\rule{0.1em}{0ex}}k\in \mathbb{N}\phantom{\rule{0.1em}{0ex}}\mathit{\text{such that}}\phantom{\rule{0.1em}{0ex}}f{x}_{n+k}\in T{x}_{n}\phantom{\rule{1em}{0ex}}\mathit{\text{for every}}\phantom{\rule{0.1em}{0ex}}n\in \mathbb{N},\phantom{\rule{0.1em}{0ex}}\mathit{\text{and}}\phantom{\rule{0.1em}{0ex}}f(Y)\phantom{\rule{0.1em}{0ex}}\mathit{\text{is complete}},
(22)
then f and T have a coincidence point. Further, if f and T are occasionally coincidentally idempotent, then f and T have a common fixed point.
Proof By hypotheses,
\begin{array}{c}{\int}_{0}^{{H}^{r}(T{x}_{n},T{x}_{m})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le {\alpha}_{1}(d(f{x}_{n},T{x}_{n})){\int}_{0}^{{d}^{r}(f{x}_{n},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\alpha}_{2}(d(f{x}_{m},T{x}_{m})){\int}_{0}^{{d}^{r}(f{x}_{m},T{x}_{m})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\hfill \end{array}
Since \{{x}_{n}\} is asymptotically Tregular with respect to f in Y, then {\{{\alpha}_{1}(d(f{x}_{n},T{x}_{n}))\}}_{n} and {\{{\alpha}_{2}(d(f{x}_{m},T{x}_{m}))\}}_{m} are bounded sequences and {\int}_{0}^{{d}^{r}(f{x}_{n},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\to 0, {\int}_{0}^{{d}^{r}(f{x}_{m},T{x}_{m})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\to 0, as n,m\to +\mathrm{\infty}. This provides the property H(T{x}_{n},T{x}_{m})\to 0 as n,m\to +\mathrm{\infty}, so that \{T{x}_{n}\} is a Cauchy sequence in (CB(X),H).
If T(Y) is complete, there exists {K}^{\ast}\in T(Y)\subseteq f(Y) such that H(T{x}_{n},{K}^{\ast})\to 0 as n\to +\mathrm{\infty}. Let u\in Y be such that f(u)\in {K}^{\ast}. Then
\begin{array}{rcl}{\int}_{0}^{{d}^{r}(fu,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& \le & {\int}_{0}^{{H}^{r}({K}^{\ast},Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\int}_{0}^{{(H({K}^{\ast},T{x}_{n})+H(T{x}_{n},Tu))}^{r}}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ =& {\int}_{0}^{{H}^{r}(T{x}_{n},Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\int}_{{H}^{r}(T{x}_{n},Tu)}^{{H}^{r}(T{x}_{n},Tu)+\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{m}\mathrm{s}\phantom{\rule{0.25em}{0ex}}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{g}\phantom{\rule{0.25em}{0ex}}H({K}^{\ast},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & {\alpha}_{1}(d(f{x}_{n},T{x}_{n})){\int}_{0}^{{d}^{r}(f{x}_{n},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\alpha}_{2}(d(fu,Tu)){\int}_{0}^{{d}^{r}(fu,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ +{\int}_{{H}^{r}(T{x}_{n},Tu)}^{{H}^{r}(T{x}_{n},Tu)+\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{m}\mathrm{s}\phantom{\rule{0.25em}{0ex}}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{g}\phantom{\rule{0.25em}{0ex}}H({K}^{\ast},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,\end{array}
where the number of terms containing H({K}^{\ast},T{x}_{n}) is a finite number depending on r, and therefore fixed. Calculating the limit as n\to +\mathrm{\infty} and taking into account that the length of the intervals in the last integral tends to zero, we get
{\int}_{0}^{{d}^{r}(fu,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt(1{\alpha}_{2}(d(fu,Tu)))\le \underset{n\to +\mathrm{\infty}}{lim}{\alpha}_{1}(d(f{x}_{n},T{x}_{n})){\int}_{0}^{{d}^{r}(f{x}_{n},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt=0.
Therefore,
{\int}_{0}^{{d}^{r}(fu,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le 0
and, by the properties of ψ, we get {d}^{r}(fu,Tu)=0, which implies that fu\in Tu and u is a coincidence point.
Now, suppose that f(Y) is complete. Note that T{x}_{n} is closed and bounded for every n\in \mathbb{N}. Take k>1 fixed. By the results in [24], we can affirm that for every {y}_{1}\in T{x}_{n}, there exists {y}_{2}\in T{x}_{m} such that d({y}_{1},{y}_{2})\le kH(T{x}_{n},T{x}_{m}).
Given n,m\in \mathbb{N}, we choose {y}_{1}\in T{x}_{n} and, for this {y}_{1}\in T{x}_{n} fixed, we choose {y}_{2}\in T{x}_{m} such that d({y}_{1},{y}_{2})\le kH(T{x}_{n},T{x}_{m}). Then
\begin{array}{rcl}d(f{x}_{n},f{x}_{m})& \le & d(f{x}_{n},{y}_{1})+d({y}_{1},{y}_{2})+d({y}_{2},f{x}_{m})\\ \le & d(f{x}_{n},T{x}_{n})+kH(T{x}_{n},T{x}_{m})+d(T{x}_{m},f{x}_{m}).\end{array}
By the hypothesis on \{{x}_{n}\} and the Cauchy character of \{T{x}_{n}\}, we deduce that \{f{x}_{n}\} is a Cauchy sequence. Since f(Y) is complete, there exists f(u)\in f(Y) such that \{f({x}_{n})\}\to f(u). By hypotheses, d(f{x}_{n+k},Tu)\le H(T{x}_{n},Tu) for every n, hence
\begin{array}{c}{\int}_{0}^{{d}^{r}(f{x}_{n+k},Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le {\int}_{0}^{{H}^{r}(T{x}_{n},Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le {\alpha}_{1}(d(f{x}_{n},T{x}_{n})){\int}_{0}^{{d}^{r}(f{x}_{n},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\alpha}_{2}(d(fu,Tu)){\int}_{0}^{{d}^{r}(fu,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,\hfill \end{array}
and taking the limit as n\to +\mathrm{\infty}, we get
{\int}_{0}^{{d}^{r}(fu,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\alpha}_{2}(d(fu,Tu)){\int}_{0}^{{d}^{r}(fu,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.
In this case,
(1{\alpha}_{2}(d(fu,Tu))){\int}_{0}^{{d}^{r}(fu,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le 0
and d(fu,Tu)=0, which implies that fu\in Tu. Now, if f and T are coincidentally idempotent, then ffw=fw for some w\in C(T,f). Hence
\begin{array}{c}{\int}_{0}^{{H}^{r}(Tfw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le {\alpha}_{1}(d(ffw,Tfw)){\int}_{0}^{{d}^{r}(ffw,Tfw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\alpha}_{2}(d(fw,Tw)){\int}_{0}^{{d}^{r}(fw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}={\alpha}_{1}(d(fw,Tfw)){\int}_{0}^{{d}^{r}(fw,Tfw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\hfill \end{array}
Since ffw=fw\in Tw, we get
{\int}_{0}^{{d}^{r}(fw,Tfw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\int}_{0}^{{H}^{r}(Tw,Tfw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\alpha}_{1}(d(fw,Tfw)){\int}_{0}^{{d}^{r}(fw,Tfw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.
Therefore
{\int}_{0}^{{d}^{r}(fw,Tfw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt(1{\alpha}_{1}(d(fw,Tfw)))\le 0,
obtaining d(fw,Tfw)=0 and fw\in Tfw. Since 0\le {\int}_{0}^{{H}^{r}(Tfw,Tw)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le 0, we deduce that H(Tfw,Tw)=0 and Tfw=Tw. In consequence, ffw=fw\in Tw=Tfw and fw is a common fixed point of T and f. □
Remark 25 In the statement of Theorem 24, condition (22) can be replaced by the more general one
To complete the proof with this more general hypothesis, take into account that for y\in Y, T(y) is a closed set in X and T(Y)\subseteq f(Y). Using that f(Y) is complete, we deduce that (CL(f(Y)),H) is complete. Hence \{T{x}_{n}\} is a sequence in CL(f(Y)) and it is a Cauchy sequence in (CL(f(Y)),H). Therefore, there exists {K}^{\ast}\in CL(f(Y)) such that H(T{x}_{n},{K}^{\ast})\to 0 as n\to +\mathrm{\infty}. Note also that {K}^{\ast} is a closed set in the complete space f(Y), then {K}^{\ast} is complete and, therefore, a closed set, then {K}^{\ast}\in CL(X). Once we have proved that H(T{x}_{n},{K}^{\ast})\to 0 as n\to +\mathrm{\infty} in (CL(f(Y)),H), the proof follows analogously.
Theorem 26 In addition to the hypotheses of Theorem 24, suppose that T{x}_{n} is compact for all n\in \mathbb{N}. If f(z) is a cluster point of \{f{x}_{n}\}, then z is a coincidence point of f and T.
Proof Let {y}_{n}\in T{x}_{n} be such that d(f{x}_{n},{y}_{n})=d(f{x}_{n},T{x}_{n})\to 0, this is possible since T{x}_{n} is compact. It is obvious that a cluster point of \{f{x}_{n}\} is a cluster point of \{{y}_{n}\}. Let f(z) be a cluster point of \{f{x}_{n}\} and \{{y}_{n}\}, then we check that fz\in Tu, where u is obtained in the proof of Theorem 24. Note that, for every y\in Tu,
d(fz,y)\le d(fz,f{x}_{n})+d(f{x}_{n},{y}_{n})+d({y}_{n},y)=d(fz,f{x}_{n})+d(f{x}_{n},T{x}_{n})+d({y}_{n},y),
hence
\begin{array}{rcl}d(fz,Tu)& =& \underset{y\in Tu}{inf}d(fz,y)\le d(fz,f{x}_{n})+d(f{x}_{n},T{x}_{n})+\underset{y\in Tu}{inf}d({y}_{n},y)\\ =& d(fz,f{x}_{n})+d(f{x}_{n},T{x}_{n})+d({y}_{n},Tu)\le d(fz,f{x}_{n})+d(f{x}_{n},T{x}_{n})+H(T{x}_{n},Tu).\end{array}
In consequence,
{\int}_{0}^{{d}^{r}(fz,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\int}_{0}^{{(d(fz,f{x}_{n})+d(f{x}_{n},T{x}_{n})+H(T{x}_{n},Tu))}^{r}}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.
Using that there exists a subsequence f{x}_{{n}_{k}} converging to fz, the properties of \{{x}_{n}\} and the inequality
\begin{array}{c}{\int}_{0}^{{H}^{r}(T{x}_{{n}_{k}},Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le {\alpha}_{1}(d(f{x}_{{n}_{k}},T{x}_{{n}_{k}})){\int}_{0}^{{d}^{r}(f{x}_{{n}_{k}},T{x}_{{n}_{k}})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\alpha}_{2}(d(fu,Tu)){\int}_{0}^{{d}^{r}(fu,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\stackrel{k\to +\mathrm{\infty}}{\u27f6}0,\hfill \end{array}
then, taking the limit when {n}_{k}\to +\mathrm{\infty}, we get {\int}_{0}^{{d}^{r}(fz,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le 0 and fz\in Tu. To prove that fz\in Tz, using that fu\in Tu, we get
\begin{array}{rcl}{\int}_{0}^{{d}^{r}(fz,Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& \le & {\int}_{0}^{{H}^{r}(Tu,Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & {\alpha}_{1}(d(fu,Tu)){\int}_{0}^{{d}^{r}(fu,Tu)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\alpha}_{2}(d(fz,Tz)){\int}_{0}^{{d}^{r}(fz,Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ =& {\alpha}_{2}(d(fz,Tz)){\int}_{0}^{{d}^{r}(fz,Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\end{array}
This implies that
(1{\alpha}_{2}(d(fz,Tz))){\int}_{0}^{{d}^{r}(fz,Tz)}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le 0
and, by the properties of {\alpha}_{2} and ψ, we deduce that d(fz,Tz)=0, which proves that z is a coincidence point of f and T. □
The following result extends [[10], Theorem 3.3].
Theorem 27 Let Y be an arbitrary nonempty set, (X,d) be a metric space, f:Y\to X and T:Y\to CB(X) be such that (1) and (21) hold, where {\alpha}_{i}:{\mathbb{R}}_{+}\to [0,1) (i=1,2) are bounded on bounded sets and such that
{\alpha}_{1}(d(fx,Tx))+{\alpha}_{2}(d(fy,Ty))\le 1\phantom{\rule{1em}{0ex}}\mathit{\text{for every}}\phantom{\rule{0.1em}{0ex}}x,y,
r is some fixed positive real number and \psi :{\mathbb{R}}_{+}\to {\mathbb{R}}_{+} is a Lebesgue measurable mapping which is summable on each compact interval, and \psi (x)>0 for each x>0. Suppose that
inf\{d(f{z}_{n},T{z}_{n}):n\in \mathbb{N}\}=0\phantom{\rule{1em}{0ex}}\mathit{\text{for every sequence}}\phantom{\rule{0.1em}{0ex}}\{{z}_{n}\}\phantom{\rule{0.1em}{0ex}}\mathit{\text{in}}\phantom{\rule{0.1em}{0ex}}Y\phantom{\rule{0.1em}{0ex}}\mathit{\text{with}}\phantom{\rule{0.1em}{0ex}}f{z}_{n}\in T{z}_{n1},\mathrm{\forall}n.
(23)
If T(Y) is complete or f(Y) is complete, then f and T have a coincidence point. Further, if f and T are occasionally coincidentally idempotent, then f and T have a common fixed point.
Proof Using Theorem 24, it suffices to prove that there exists an asymptotically Tregular sequence \{{x}_{n}\} with respect to f in Y. Let {x}_{0}\in Y and take \{{x}_{n}\} in Y such that f{x}_{n}\in T{x}_{n1} for every n\in \mathbb{N}. Then
\begin{array}{c}{\int}_{0}^{{d}^{r}(f{x}_{n},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le {\int}_{0}^{{H}^{r}(T{x}_{n1},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\hfill \\ \phantom{\rule{1em}{0ex}}\le {\alpha}_{1}(d(f{x}_{n1},T{x}_{n1})){\int}_{0}^{{d}^{r}(f{x}_{n1},T{x}_{n1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt+{\alpha}_{2}(d(f{x}_{n},T{x}_{n})){\int}_{0}^{{d}^{r}(f{x}_{n},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\hfill \end{array}
Hence,
(1{\alpha}_{2}(d(f{x}_{n},T{x}_{n}))){\int}_{0}^{{d}^{r}(f{x}_{n},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\le {\alpha}_{1}(d(f{x}_{n1},T{x}_{n1})){\int}_{0}^{{d}^{r}(f{x}_{n1},T{x}_{n1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt,
or also, using the hypothesis on {\alpha}_{1} and {\alpha}_{2},
\begin{array}{rcl}{\int}_{0}^{{d}^{r}(f{x}_{n},T{x}_{n})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt& \le & \frac{{\alpha}_{1}(d(f{x}_{n1},T{x}_{n1}))}{(1{\alpha}_{2}(d(f{x}_{n},T{x}_{n})))}{\int}_{0}^{{d}^{r}(f{x}_{n1},T{x}_{n1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt\\ \le & {\int}_{0}^{{d}^{r}(f{x}_{n1},T{x}_{n1})}\psi (t)\phantom{\rule{0.2em}{0ex}}dt.\end{array}
The properties of ψ imply that {d}^{r}(f{x}_{n},T{x}_{n})\le {d}^{r}(f{x}_{n1},T{x}_{n1}) for every n\in \mathbb{N}, and {\{d(f{x}_{n},T{x}_{n})\}}_{n\in \mathbb{N}} is nonincreasing and bounded below. Therefore it is convergent to the infimum, that is,
d(f{x}_{n},T{x}_{n})\to inf\{d(f{x}_{n},T{x}_{n}):n\in \mathbb{N}\}=0,
and \{{x}_{n}\} is asymptotically Tregular with respect to f in Y. □
Remark 28 Note that condition (23) in Theorem 27 cannot be replaced by
since the infimum taking the sequence \{{z}_{n}\} could be positive (we calculate the infimum in a smaller set).
Remark 29 In Theorem 27, condition (23) can be replaced by the following:
\begin{array}{c}inf\{H(T{z}_{n1},T{z}_{n}):n\in \mathbb{N}\}=0,\hfill \\ \phantom{\rule{1em}{0ex}}\text{for every sequence}\{{z}_{n}\}\text{in}Y\text{with}f{z}_{n}\in T{z}_{n1},\mathrm{\forall}n.\hfill \end{array}
(24)
Indeed, since
d(f{z}_{n},T{z}_{n})\le H(T{z}_{n1},T{z}_{n}),\phantom{\rule{1em}{0ex}}\mathrm{\forall}n,
then
0\le inf\{d(f{z}_{n},T{z}_{n}):n\in \mathbb{N}\}\le inf\{H(T{z}_{n1},T{z}_{n}):n\in \mathbb{N}\}=0,
and d(f{z}_{n},T{z}_{n})\to 0.
Remark 30 In Theorem 27, if we are able to obtain a sequence \{{x}_{n}\} with an infinite number of terms which are different, then we can relax condition (23) to the following:
inf\{d(fx,Tx):x\in B\}=0\phantom{\rule{1em}{0ex}}\text{for every infinite set}B\text{of}Y\text{.}
(25)