Research Article New Results in the Startpoint Theory for...

6
Research Article New Results in the Startpoint Theory for Quasipseudometric Spaces Yaé Ulrich Gaba Department of Mathematics and Applied Mathematics, University of Cape Town, Rondebosch, Cape Town 7701, South Africa Correspondence should be addressed to Ya´ e Ulrich Gaba; [email protected] Received 22 July 2014; Revised 24 October 2014; Accepted 27 October 2014; Published 17 November 2014 Academic Editor: Aref Jeribi Copyright © 2014 Ya´ e Ulrich Gaba. is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. We give two generalizations of eorem 35 proved by Gaba (2014). More precisely, we change the structure of the contractive condition; namely, we introduce a function Φ instead of a simple constant . Dedicated to Professor Guy A. Degla for his mentorship 1. Introduction and Preliminaries In [1], we introduced the concept of startpoint and end- point for set-valued mappings defined on quasipseudometric spaces. As mentioned there, the purpose of this theory is to study fixed point like related properties. In the present, we give more results from the theory. More precisely, we generalize eorem 35 of [1] by changing the structure of the contractive condition; namely, we introduce a function Φ instead of a simple constant (as it appears in the original statement). is new condition is interesting in the sense that it allows us to have a condition involving a functional of the variables and not just the variables themselves. For the convenience of the reader, we will recall some necessary definitions but for a detailed expos´ e of the definition and examples, the interested reader is referred to [1]. Definition 1. Let be a nonempty set. A function → [0, ∞) is called quasipseudometric on if (i) (, ) = 0 ∀ ∈ ; (ii) (, ) ≤ (, ) + (, ) ∀, , ∈ . Moreover, if (, ) = 0 = (, ) ⇒ = , then is said to be a 0 -quasipseudometric. e latter condition is referred to as the 0 -condition. Remark 2. (i) Let be quasipseudometric on ; then the map −1 defined by −1 (, ) = (, ) whenever , ∈ is also a quasipseudometric on , called the conjugate of . In the literature, −1 is also denoted as or . (ii) It is easy to verify that the function defined by := −1 , that is, (, ) = max{(, ), (, )}, defines a metric on whenever is a 0 -quasipseudometric on . e quasipseudemetric induces a topology () on . Definition 3. Let (, ) be a quasipseudometric space. e -convergence of a sequence ( ) to a point , also called leſt-convergence and denoted by , is defined in the following way: → ⇐⇒ (, ) → 0. (1) Similarly, we define the −1 -convergence of a sequence ( ) to a point or right convergence and denote it by −1 , in the following way: −1 → ⇐⇒ ( ,) → 0. (2) Finally, in a quasipseudometric space (, ), we will say that a sequence ( ) -converges to if it is both leſt and right Hindawi Publishing Corporation Journal of Operators Volume 2014, Article ID 741818, 5 pages http://dx.doi.org/10.1155/2014/741818

Transcript of Research Article New Results in the Startpoint Theory for...

Page 1: Research Article New Results in the Startpoint Theory for ...downloads.hindawi.com/archive/2014/741818.pdf · Research Article New Results in the Startpoint Theory for Quasipseudometric

Research ArticleNew Results in the Startpoint Theory forQuasipseudometric Spaces

Yaeacute Ulrich Gaba

Department of Mathematics and Applied Mathematics University of Cape Town Rondebosch Cape Town 7701 South Africa

Correspondence should be addressed to Yae Ulrich Gaba gabayae2gmailcom

Received 22 July 2014 Revised 24 October 2014 Accepted 27 October 2014 Published 17 November 2014

Academic Editor Aref Jeribi

Copyright copy 2014 Yae Ulrich Gaba This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

We give two generalizations of Theorem 35 proved by Gaba (2014) More precisely we change the structure of the contractivecondition namely we introduce a functionΦ instead of a simple constant 119888

Dedicated to Professor Guy A Degla for his mentorship

1 Introduction and Preliminaries

In [1] we introduced the concept of startpoint and end-point for set-valued mappings defined on quasipseudometricspaces As mentioned there the purpose of this theory isto study fixed point like related properties In the presentwe give more results from the theory More precisely wegeneralize Theorem 35 of [1] by changing the structure ofthe contractive condition namely we introduce a functionΦinstead of a simple constant 119888 (as it appears in the originalstatement) This new condition is interesting in the sensethat it allows us to have a condition involving a functionalof the variables and not just the variables themselves Forthe convenience of the reader we will recall some necessarydefinitions but for a detailed expose of the definition andexamples the interested reader is referred to [1]

Definition 1 Let 119883 be a nonempty set A function 119889 119883 times119883 rarr [0infin) is called quasipseudometric on119883 if

(i) 119889(119909 119909) = 0 forall119909 isin 119883(ii) 119889(119909 119911) le 119889(119909 119910) + 119889(119910 119911) forall119909 119910 119911 isin 119883

Moreover if 119889(119909 119910) = 0 = 119889(119910 119909) rArr 119909 = 119910 then 119889 is saidto be a 119879

0-quasipseudometric The latter condition is referred

to as the 1198790-condition

Remark 2 (i) Let 119889 be quasipseudometric on119883 then themap119889minus1 defined by 119889minus1(119909 119910) = 119889(119910 119909) whenever 119909 119910 isin 119883 is also

a quasipseudometric on 119883 called the conjugate of 119889 In theliterature 119889minus1 is also denoted as 119889119905 or 119889

(ii) It is easy to verify that the function 119889119904 defined by 119889119904 =119889 or 119889

minus1 that is 119889119904(119909 119910) = max119889(119909 119910) 119889(119910 119909) defines ametric on119883 whenever 119889 is a 119879

0-quasipseudometric on119883

The quasipseudemetric 119889 induces a topology 120591(119889) on119883

Definition 3 Let (119883 119889) be a quasipseudometric space The119889-convergence of a sequence (119909

119899) to a point 119909 also called

left-convergence and denoted by 119909119899

119889

997888rarr 119909 is defined in thefollowing way

119909119899

119889

997888997888rarr 119909 lArrrArr 119889 (119909 119909119899) 997888rarr 0 (1)

Similarly we define the 119889minus1-convergence of a sequence

(119909119899) to a point 119909 or right convergence and denote it by 119909

119899

119889minus1

997888997888997888rarr

119909 in the following way

119909119899

119889minus1

997888997888997888rarr 119909 lArrrArr 119889 (119909119899 119909) 997888rarr 0 (2)

Finally in a quasipseudometric space (119883 119889) we will saythat a sequence (119909

119899) 119889119904-converges to119909 if it is both left and right

Hindawi Publishing CorporationJournal of OperatorsVolume 2014 Article ID 741818 5 pageshttpdxdoiorg1011552014741818

2 Journal of Operators

convergent to119909 andwedenote it as119909119899

119889119904

997888rarr 119909 or119909119899rarr 119909 when

there is no confusion Hence

119909119899

119889119904

997888997888rarr 119909 lArrrArr 119909119899

119889

997888997888rarr 119909 119909119899

119889minus1

997888997888rarr 119909 (3)

Definition 4 A sequence (119909119899) in quasipseudometric (119883 119889) is

called

(a) left 119889-Cauchy if for every 120598 gt 0 there exist 119909 isin 119883 and1198990isin N such that

forall119899 ge 1198990119889 (119909 119909

119899) lt 120598 (4)

(b) left 119870-Cauchy if for every 120598 gt 0 there exists 1198990isin N

such that

forall119899 119896 1198990le 119896 le 119899 119889 (119909

119896 119909119899) lt 120598 (5)

(c) 119889119904-Cauchy if for every 120598 gt 0 there exists 1198990isin N such

that

forall119899 119896 ge 1198990119889 (119909119899 119909119896) lt 120598 (6)

Dually we define right 119889-Cauchy and right 119870-Cauchysequences

Definition 5 A quasipseudometric space (119883 119889) is called

(i) left 119870-complete provided that any left 119870-Cauchysequence is 119889-convergent

(ii) left Smyth sequentially complete if any left 119870-Cauchysequence is 119889119904-convergent

Definition 6 A 1198790-quasipseudometric space (119883 119889) is called

bicomplete provided that the metric 119889119904 on119883 is complete

As usual a subset 119860 of a quasipseudometric space (119883 119889)will be called bounded provided that there exists a positivereal constant119872 such that 119889(119909 119910) lt 119872 whenever 119909 119910 isin 119860

Let (119883 119889) be a quasipseudometric spaceWe setP0(119883) =

2119883 0 where 2119883 denotes the power set of119883 For 119909 isin 119883 and

119860 119861 isin P0(119883) we define

119889 (119909 119860) = inf 119889 (119909 119886) 119886 isin 119860

119889 (119860 119909) = inf 119889 (119886 119909) 119886 isin 119860 (7)

and119867(119860 119861) by

119867(119860 119861) = maxsup119886isin119860

119889 (119886 119861) sup119887isin119861

119889 (119860 119887) (8)

Then 119867 is an extended quasipseudometric on P0(119883)

Moreover we know from [2] that the restriction of 119867 to119878119888119897(119883) = 119860 sube 119883 119860 = (119888119897

120591(119889)119860) cap (119888119897

120591(119889minus1

)119860) is an

extended 1198790-quasipseudometric We will denote by 119862119861(119883)

the collection of all nonempty bounded and 119889119904-closed subsetsof119883

Definition 7 (compare [1]) Let 119865 119883 rarr 2119883 be a set-valued

map An element 119909 isin 119883 is said to be

(i) a fixed point of 119865 if 119909 isin 119865119909(ii) a startpoint of 119865 if119867(119909 119865119909) = 0(iii) an endpoint of 119865 if119867(119865119909 119909) = 0

We complete this section by recalling the followinglemma

Lemma 8 (compare [1]) Let (119883 119889) be a quasipseudometricspace For every fixed 119909 isin 119883 the mapping 119910 997891rarr 119889(119909 119910) is 120591(119889)upper semicontinuous (120591(119889)-119906119904119888 in short) and 120591(119889minus1) lowersemicontinuous (120591(119889minus1)-lsc in short) For every fixed 119910 isin 119883the mapping 119909 997891rarr 119889(119909 119910) is 120591(119889)-lsc and 120591(119889minus1)-usc

2 Main Results

We commence this section with the main result of this paper

Theorem9 Let (119883 119889) be a left119870-complete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and define119891 119883 rarr R as119891(119909) = 119867(119909 119879119909) LetΦ [0infin) rarr [0 1) bea function such that lim sup

119903rarr 119905+Φ(119903) lt 1 for each 119905 isin [0infin)

Moreover assume that for any 119909 isin 119883 there exists 119910 isin 119879119909satisfying

119867(119910 119879119910) le Φ (119867 (119909 119910))119867 (119909 119910) (9)

and then 119879 has a startpoint

Proof First observe that since Φ(119867(119909 119910)) lt 1 for any119909 119910 isin 119883 it follows that 2 minus Φ(119867(119909 119910)) gt 1 for any 119909 119910 isin119883 and hence

119867(119909 119910) le [2 minus Φ (119867 (119909 119910))]119867 (119909 119879119909) (10)

for any 119909 isin 119883 and 119910 isin 119879119909For any initial 119909

0isin 119883 there exists (for all actually) 119909

1isin

1198791199090sube 119883 such that

119867(1199090 1199091) le [2 minus Φ (119867 (119909

0 1199091))]119867 (119909

0 1198791199090)

(11)

From (9) we get

119867(1199091 1198791199091) le Φ (119867 (119909

0 1199091))119867 (119909

0 1199091)

le Φ (119867 (1199090 1199091))

times [2 minus Φ (119867 (1199090 1199091))]119867 (119909

0 1198791199090)

= Ψ (119867 (1199090 1199091))119867 (119909

0 1198791199090)

(12)

where Ψ [0infin) rarr [0 1) is defined by

Ψ (119905) = Φ (119905) (2 minus Φ (119905)) (13)

Now choosing 1199092isin 1198651199091sube 119883 we have that

119867(1199091 1199092) le [2 minus Φ (119867 (119909

1 1199092))]119867 (119909

1 1198791199091)

(14)

Journal of Operators 3

and from (9) we get

119867(1199092 1198791199092) le Ψ (119867 (119909

1 1199092))119867 (119909

1 1198791199091) (15)

Continuing this process we obtain a sequence (119909119899)where

119909119899+1isin 119879119909119899sube 119883 with

119867(119909119899 119909119899+1)

le [2 minus Φ (119867 (119909119899 119909119899+1))]119867 (119909

119899 119879119909119899)

(16)

119867(119909119899+1 119879119909119899+1) le Ψ (119867 (119909

119899 119909119899+1))119867 (119909

119899 119879119909119899)

119899 = 1 2

(17)

For simplicity denote 119889119899= 119867(119909

119899 119909119899+1) and 119863

119899=

119867(119909119899 119879119909119899) for all 119899 ge 0 So from (17) we can write

119863119899+1le Ψ (119889

119899)119863119899le 119863119899 (18)

for all 119899 ge 0 Hence (119863119899) is a strictly decreasing sequence and

hence there exists 120575 ge 0 such that

lim119899rarrinfin

119863119899= 120575 (19)

From (16) it is easy to see that

119889119899lt 2119863119899 (20)

Thus the sequence (119889119899) is bounded and so there is 119889 ge 0

such that lim sup119899rarrinfin

119889119899= 119889 and hence a subsequence (119889

119899119896

)

of (119889119899) such that lim

119896rarrinfin119889119899119896

= 119889+ From (17)we have119863

119899119896

+1le

Ψ(119889119899119896

)119863119899119896

and thus

120575 = lim sup119896rarrinfin

119863119899119896

+1

le (lim sup119896rarrinfin

Ψ(119889119899119896

))(lim sup119896rarrinfin

119863119899119896

)

le lim sup119889119899

119896

rarr119889+

Ψ(119889119899119896

) 120575

(21)

This together with the fact lim sup119903rarr 119905+Φ(119903) lt 1 for each

119905 isin [0infin) implies that 120575 = 0 Then from (19) and (20) wederive that lim

119899rarrinfin119889119899= 0

Claim 1 (119909119899) is a left 119870-Cauchy sequence

Now let 120572 = lim sup119889119899

rarr0+Ψ(119889119899) and 119902 such that 120572 lt 119902 lt

1 This choice of 119902 is always possible since 120572 lt 1 Then thereis 1198990such that Ψ(119889

119899) lt 119902 for all 119899 ge 119899

0 So from (18) we have

119863119899+1le 119902119863

119899for all 119899 ge 119899

0 Then by induction we get 119863

119899le

119902119899minus11989901198631198990

for all 119899 ge 1198990+ 1 Combining this and inequality

(20) we get

119898

sum

119896=1198990

119889 (119909119896 119909119896+1) le 2

119898

sum

119896=1198990

119902119896minus11989901198631198990

le 21

1 minus 1199021198631198990

(22)

for all119898 gt 119899 ge 1198990+1 Hence (119909

119899) is a left119870-Cauchy sequence

According to the left 119870-completeness of (119883 119889) thereexists 119909lowast isin 119883 such that 119909

119899

119889

997888rarr 119909lowast

Claim 2 119909lowast is a startpoint of 119879Observe that the sequence 119863

119899= (119891119909

119899) = (119867(119909

119899 119879119909119899))

converges to 0 Since 119891 is 120591(119889)-lower semicontinuous (assupremumof 120591(119889)-lower semicontinuous functions) we have

0 le 119891 (119909lowast) le lim inf119899rarrinfin

119891 (119909119899) = 0 (23)

Hence 119891(119909lowast) = 0 that is119867(119909lowast 119879119909lowast) = 0This completes the proof

We give below an example to illustrate the theorem

Example 10 Let 119883 = [0 1] and 119889 119883 times 119883 rarr R be themapping defined by 119889(119886 119887) = max119886 minus 119887 0 Then 119889 is a1198790-quasipseudometric on119883 Observe that any left119870-Cauchy

sequence in (119883 119889) is 119889-convergent to 0 Indeed if (119909119899) is a left

119870-Cauchy sequence for every 120598 gt 0 there exists 1198990isin N such

that

forall119899 119896 1198990le 119896 le 119899 119889 (119909

119896 119909119899) lt 120598 (24)

This entails that forall119899 1198990lt 119899

119889 (0 119909119899) le 119889 (0 119909

119899minus1) + 119889 (119909

119899minus1 119909119899) = 0 + 119889 (119909

119899minus1 119909119899) lt 120598

(25)

Hence 119889(0 119909119899) rarr 0 that is 119909

119899

119889

997888rarr 0 Therefore (119883 119889) isleft 119870-complete Let 119879 119883 rarr 119862119861(119883) be such that

119879 (119909) =

1

21199092 for 119909 isin [0 15

32) cup (

15

32 1]

17

961

4 for 119909 = 15

32

(26)

Let Φ [0infin) rarr [0 34) sub [0 1) be defined by

Φ (119903) =

3

2119903 for 119903 isin [0 7

24) cup (

7

241

2)

425

768 for 119903 = 7

24

1

2 for 119903 isin [1

2infin)

(27)

An explicit computation of 119891(119909) = 119867(119909 119879119909) gives

119891 (119909) =

119909 minus1

21199092 for 119909 isin [0 15

32) cup (

15

32 1]

28

96 for 119909 = 15

32

(28)

Moreover for each 119909 isin [0 1532) cup (1532 1] 119910 =

(12)1199092 and we have

119867(119910 119879119910) le Φ (119867 (119909 119910))119867 (119909 119910) (29)

Of course inequality (29) also holds in the case of119909 = 1532 and 119910 = 1796 Therefore all assumptions ofTheorem 9 are satisfied and the endpoint of 119879 is 119909 = 0

4 Journal of Operators

Remark 11 In fact every sequence in (119883 119889) 119889-converges to0

Corollary 12 Let (119883 119889) be a right 119870-complete quasipseudo-metric space Let 119879 119883 rarr 119862119861(119883) be a set-valued map anddefine 119891 119883 rarr R as 119891(119909) = 119867(119879119909 119909) Let Φ [0infin) rarr[0 1) be a function such that lim sup

119903rarr 119905+Φ(119903) lt 1 for each

119905 isin [0infin) Moreover assume that for any 119909 isin 119883 there exists119910 isin 119879119909 satisfying

119867(119879119910 119910) le Φ (119867 (119910 119909))119867 (119910 119909) (30)

and then 119879 has an endpoint

Corollary 13 Let (119883 119889) be a bicomplete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and119891 119883 rarr R defined by 119891(119909) = 119867

119904(119879119909 119909) =

max119867(119879119909 119909)119867(119909 119879119909) If there exists 119888 isin (0 1) suchthat for all 119909 isin 119883 there exists 119910 isin 119865119909 satisfying

119867119904(119910 119865119910) le min Φ (119886) 119886 Φ (119887) 119887 (31)

where 119886 = 119867(119910 119909) and 119887 = 119867(119909 119910) then 119879 has a fixedpoint

Proof We give here the main idea of the proof Observe thatinequality (31) guarantees that the sequence (119909

119899) constructed

in the proof ofTheorem 9 is a 119889119904-Cauchy sequence and hence119889119904-converges to some 119909lowast Using the fact that 119891 is 120591(119889119904)-

lower semicontinuous (as supremum of 120591(119889119904)-continuousfunctions) we have

0 le 119891 (119909lowast) le lim inf119899rarrinfin

119891 (119909119899) = 0 (32)

Hence 119891(119909lowast) = 0 that is 119867(119909lowast 119879119909lowast) = 0 =

119867(119879119909lowast 119909lowast) and we are done

The following theorem is the second generalization thatwe propose

Theorem 14 Let (119883 119889) be a left119870-complete quasipseudomet-ric space Let 119879 119883 rarr 119862119861(119883) be a set-valued map anddefine 119891 119883 rarr R as 119891(119909) = 119867(119909 119879119909) Let 119887 [0infin) rarr[119886 1) 119886 gt 0 be a nondecreasing function Let Φ [0infin) rarr[0 1) be a function such that Φ(119905) lt 119887(119905) for each 119905 isin [0infin)and lim sup

119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for each 119905 isin [0infin)

Moreover assume that for any 119909 isin 119883 there exists 119910 isin 119879119909satisfying

119867(119910 119879119910) le Φ (119867 (119909 119910))119867 (119909 119910) (33)

and then 119879 has a startpoint

Proof Observe that because 119887(119867(119909 119910)) lt 1 for all 119909 119910 isin119883

119887 (119867 (119909 119910))119867 (119909 119910) le 119867 (119909 119879119909) (34)

for any 119910 isin 119879119909 Let 1199090isin 119883 be arbitrary Then we can choose

1199091isin 1198791199090such that

119887 (119867 (1199090 1199091))119867 (119909

0 1199091) le 119867 (119909

0 1198791199090) (35)

119867(1199091 1198791199091) le Φ (119867 (119909

0 1199091))119867 (119909

0 1199091) (36)

Define the function Ψ [0infin) rarr [0 1) by Ψ(119905) =Φ(119905)119887(119905) Hence (35) and (36) together sum to

119867(1199091 1198791199091) le Ψ (119867 (119909

0 1199091))119867 (119909

0 1198791199090) (37)

Now we choose 1199092isin 1198791199091such that

119887 (119867 (1199091 1199092))119867 (119909

1 1199092) le 119867 (119909

1 1198791199091)

119867 (1199092 1198791199092) le Φ (119867 (119909

1 1199092))119867 (119909

1 1199092)

(38)

which lead to

119867(1199092 1198791199092) le Ψ (119867 (119909

1 1199092))119867 (119909

1 1198791199091) (39)

Continuing this process we get an iterative sequence (119909119899)

where 119909119899+1isin 119879119909119899sube 119883 and denoting 119889

119899= 119867(119909

119899 119909119899+1)

and119863119899= 119867(119909

119899 119879119909119899) for all 119899 ge 0 we can write that

119887 (119889119899) 119889119899le 119863119899 119863

119899+1le Φ (119889

119899) 119889119899 (40)

for all 119899 ge 0 Hence

119863119899+1le Ψ (119889

119899)119863119899 (41)

If 119863119896= 0 for some 119896 ge 0 then we trivially have

lim119899rarrinfin

119863119899= 0 and the conclusion is immediate So without

loss of generality we can assume that 119863119899gt 0 for all 119899 ge 0

and from (41) we have119863119899+1lt 119863119899for all 119899 ge 0

Observe also that if for some 119899 119889119899le 119889119899+1

we are led toa contradiction Indeed from (40) and using the fact that thefunction 119887 is nondecreasing we have

119889119899le 119889119899+1leΦ (119889119899)

119887 (119889119899+1)119889119899leΦ (119889119899)

119887 (119889119899)119889119899= Ψ (119889

119899) 119889119899lt 119889119899

(42)

Hence 119889119899+1lt 119889119899for all 119899 ge 0 Thus there exist 120575 ge 0 and

119889 ge 0 such that lim119899rarrinfin

119863119899= 120575 and lim

119899rarrinfin119889119899= 119889 From

(41) we get

120575 le (lim sup119899rarrinfin

Ψ (119889119899)) 120575 = lim sup

119889119899

rarr119889+

Ψ (119889119899) 120575 (43)

and hence 120575 = 0 (since lim sup119903rarr 119905+Φ(119903) lt 1 for all 119905 isin

[0infin)) Moreover since

119886119889119899le 119887 (119889

119899) 119889119899le 119863119899le Ψ (119889

119899minus1)119863119899minus1 (44)

this forces 119889 to be 0 that is lim119899rarrinfin

119889119899= 0

Furthermore setting 120572 = lim sup119889119899

rarr0+Ψ(119889119899) and letting

119902 gt 0 be a positive number such that 120572 lt 119902 lt 1 there is 1198990

such that Ψ(119889119899) lt 119902 for all 119899 ge 119899

0 Hence from (41) and (44)

we get

119886119889119899le Ψ (119889

119899minus1)119863119899minus1le 119902119863119899minus1le sdot sdot sdot le 119902

119899minus11989901198631198990

(45)

for all 119899 ge 1198990 So 119889

119899le (1119886)119902

119899minus11989901198631198990

for all 119899 ge 1198990 Using a

similar argument as the one used in the proof of Theorem 9we conclude that (119909

119899) is a left119870-Cauchy sequence and that its

limit point is a starpoint of 119879

Journal of Operators 5

Example 15 Let 119883 = [0 1) and 119889 119883 times 119883 rarr R be themapping defined by 119889(119886 119887) = max119886 minus 119887 0 Then 119889 is a1198790-quasipseudometric on 119883 We know that (119883 119889) is left 119870-

complete Let 119879 119883 rarr 119862119861(119883) be such that

119879 (119909) = 1

21199092 (46)

Let Φ [0infin) rarr [0 34) sub [0 1) be defined by

Φ (119903) =

3

2119903 for 119903 isin [0 1

2)

0 for 119903 isin [12infin)

(47)

Let 119887 [0infin) rarr [23 1) be defined by

119887 (119905) =

5

6119905 for 119905 isin [0 1

2)

1

2 for 119905 isin [1

2infin)

(48)

An explicit computation of 119891(119909) = 119867(119909 119879119909) gives

119891 (119909) = 119909 minus1

21199092 for 119909 isin [0 1) (49)

Moreover for each 119909 isin [0 1) there exists 119910 = (12)1199092and we have

119867(119910 119879119910) le3

2119867 (119909 119910)119867 (119909 119910)

= Φ (119867 (119909 119910))119867 (119909 119910)

(50)

Therefore all assumptions ofTheorem 14 are satisfied andthe endpoint of 119879 is 119909 = 0

Corollary 16 Let (119883 119889) be a right 119870-complete quasipseudo-metric space Let 119879 119883 rarr 119862119861(119883) be a set-valued mapand define 119891 119883 rarr R as 119891(119909) = 119867

minus1(119909 119879119909) Let

119887 [0infin) rarr [119886 1) 119886 gt 0 be a nondecreasing function LetΦ [0infin) rarr [0 1) be a function such that Φ(119905) lt 119887(119905) foreach 119905 isin [0infin) and lim sup

119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for

each 119905 isin [0infin) Moreover assume that for any 119909 isin 119883 thereexists 119910 isin 119879119909 satisfying

119867minus1(119910 119879119910) le Φ (119867

minus1(119909 119910))119867

minus1(119909 119910) (51)

Φ(119905) lt 119887(119905) and lim sup119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for each

119905 isin [0infin) Moreover assume that for any 119909 isin 119883 there exists119910 isin 119879119909 satisfying and then 119879 has an endpoint

Corollary 17 Let (119883 119889) be a bicomplete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and define119891 119883 rarr R as 119891(119909) = 119867119904(119909 119879119909) Let 119887 [0infin) rarr [119886 1)119886 gt 0 be a nondecreasing function Let Φ [0infin) rarr [0 1)

be a function such that Φ(119905) lt 119887(119905) and lim sup119903rarr 119905+Φ(119903) lt

lim sup119903rarr 119905+119887(119903) for each 119905 isin [0infin) Moreover assume that

for any 119909 isin 119883 there exists 119910 isin 119879119909 satisfying

119867119904(119910 119879119910) le min Φ (119886) 119886 Φ (119887) 119887 (52)

where 119886 = 119867(119909 119910) and 119887 = 119867minus1(119909 119910) Then 119879 has afixed point

Remark 18 All the results given remain true when we replaceaccordingly the bicomplete quasipseudometric space (119883 119889)by a left Smyth sequentially completeleft 119870-complete or aright Smyth sequentially completeright 119870-complete space

Conflict of Interests

The author declares that there is no conflict of interestsregarding the publication of this paper

References

[1] Y U Gaba ldquoStartpoints and (120572 120574)-contractions in quasi-pseudometric spacesrdquo Journal of Mathematics vol 2014 ArticleID 709253 8 pages 2014

[2] H-P Kunzi and C Ryser ldquoThe Bourbaki quasi-uniformityrdquoTopology Proceedings vol 20 pp 161ndash183 1995

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Page 2: Research Article New Results in the Startpoint Theory for ...downloads.hindawi.com/archive/2014/741818.pdf · Research Article New Results in the Startpoint Theory for Quasipseudometric

2 Journal of Operators

convergent to119909 andwedenote it as119909119899

119889119904

997888rarr 119909 or119909119899rarr 119909 when

there is no confusion Hence

119909119899

119889119904

997888997888rarr 119909 lArrrArr 119909119899

119889

997888997888rarr 119909 119909119899

119889minus1

997888997888rarr 119909 (3)

Definition 4 A sequence (119909119899) in quasipseudometric (119883 119889) is

called

(a) left 119889-Cauchy if for every 120598 gt 0 there exist 119909 isin 119883 and1198990isin N such that

forall119899 ge 1198990119889 (119909 119909

119899) lt 120598 (4)

(b) left 119870-Cauchy if for every 120598 gt 0 there exists 1198990isin N

such that

forall119899 119896 1198990le 119896 le 119899 119889 (119909

119896 119909119899) lt 120598 (5)

(c) 119889119904-Cauchy if for every 120598 gt 0 there exists 1198990isin N such

that

forall119899 119896 ge 1198990119889 (119909119899 119909119896) lt 120598 (6)

Dually we define right 119889-Cauchy and right 119870-Cauchysequences

Definition 5 A quasipseudometric space (119883 119889) is called

(i) left 119870-complete provided that any left 119870-Cauchysequence is 119889-convergent

(ii) left Smyth sequentially complete if any left 119870-Cauchysequence is 119889119904-convergent

Definition 6 A 1198790-quasipseudometric space (119883 119889) is called

bicomplete provided that the metric 119889119904 on119883 is complete

As usual a subset 119860 of a quasipseudometric space (119883 119889)will be called bounded provided that there exists a positivereal constant119872 such that 119889(119909 119910) lt 119872 whenever 119909 119910 isin 119860

Let (119883 119889) be a quasipseudometric spaceWe setP0(119883) =

2119883 0 where 2119883 denotes the power set of119883 For 119909 isin 119883 and

119860 119861 isin P0(119883) we define

119889 (119909 119860) = inf 119889 (119909 119886) 119886 isin 119860

119889 (119860 119909) = inf 119889 (119886 119909) 119886 isin 119860 (7)

and119867(119860 119861) by

119867(119860 119861) = maxsup119886isin119860

119889 (119886 119861) sup119887isin119861

119889 (119860 119887) (8)

Then 119867 is an extended quasipseudometric on P0(119883)

Moreover we know from [2] that the restriction of 119867 to119878119888119897(119883) = 119860 sube 119883 119860 = (119888119897

120591(119889)119860) cap (119888119897

120591(119889minus1

)119860) is an

extended 1198790-quasipseudometric We will denote by 119862119861(119883)

the collection of all nonempty bounded and 119889119904-closed subsetsof119883

Definition 7 (compare [1]) Let 119865 119883 rarr 2119883 be a set-valued

map An element 119909 isin 119883 is said to be

(i) a fixed point of 119865 if 119909 isin 119865119909(ii) a startpoint of 119865 if119867(119909 119865119909) = 0(iii) an endpoint of 119865 if119867(119865119909 119909) = 0

We complete this section by recalling the followinglemma

Lemma 8 (compare [1]) Let (119883 119889) be a quasipseudometricspace For every fixed 119909 isin 119883 the mapping 119910 997891rarr 119889(119909 119910) is 120591(119889)upper semicontinuous (120591(119889)-119906119904119888 in short) and 120591(119889minus1) lowersemicontinuous (120591(119889minus1)-lsc in short) For every fixed 119910 isin 119883the mapping 119909 997891rarr 119889(119909 119910) is 120591(119889)-lsc and 120591(119889minus1)-usc

2 Main Results

We commence this section with the main result of this paper

Theorem9 Let (119883 119889) be a left119870-complete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and define119891 119883 rarr R as119891(119909) = 119867(119909 119879119909) LetΦ [0infin) rarr [0 1) bea function such that lim sup

119903rarr 119905+Φ(119903) lt 1 for each 119905 isin [0infin)

Moreover assume that for any 119909 isin 119883 there exists 119910 isin 119879119909satisfying

119867(119910 119879119910) le Φ (119867 (119909 119910))119867 (119909 119910) (9)

and then 119879 has a startpoint

Proof First observe that since Φ(119867(119909 119910)) lt 1 for any119909 119910 isin 119883 it follows that 2 minus Φ(119867(119909 119910)) gt 1 for any 119909 119910 isin119883 and hence

119867(119909 119910) le [2 minus Φ (119867 (119909 119910))]119867 (119909 119879119909) (10)

for any 119909 isin 119883 and 119910 isin 119879119909For any initial 119909

0isin 119883 there exists (for all actually) 119909

1isin

1198791199090sube 119883 such that

119867(1199090 1199091) le [2 minus Φ (119867 (119909

0 1199091))]119867 (119909

0 1198791199090)

(11)

From (9) we get

119867(1199091 1198791199091) le Φ (119867 (119909

0 1199091))119867 (119909

0 1199091)

le Φ (119867 (1199090 1199091))

times [2 minus Φ (119867 (1199090 1199091))]119867 (119909

0 1198791199090)

= Ψ (119867 (1199090 1199091))119867 (119909

0 1198791199090)

(12)

where Ψ [0infin) rarr [0 1) is defined by

Ψ (119905) = Φ (119905) (2 minus Φ (119905)) (13)

Now choosing 1199092isin 1198651199091sube 119883 we have that

119867(1199091 1199092) le [2 minus Φ (119867 (119909

1 1199092))]119867 (119909

1 1198791199091)

(14)

Journal of Operators 3

and from (9) we get

119867(1199092 1198791199092) le Ψ (119867 (119909

1 1199092))119867 (119909

1 1198791199091) (15)

Continuing this process we obtain a sequence (119909119899)where

119909119899+1isin 119879119909119899sube 119883 with

119867(119909119899 119909119899+1)

le [2 minus Φ (119867 (119909119899 119909119899+1))]119867 (119909

119899 119879119909119899)

(16)

119867(119909119899+1 119879119909119899+1) le Ψ (119867 (119909

119899 119909119899+1))119867 (119909

119899 119879119909119899)

119899 = 1 2

(17)

For simplicity denote 119889119899= 119867(119909

119899 119909119899+1) and 119863

119899=

119867(119909119899 119879119909119899) for all 119899 ge 0 So from (17) we can write

119863119899+1le Ψ (119889

119899)119863119899le 119863119899 (18)

for all 119899 ge 0 Hence (119863119899) is a strictly decreasing sequence and

hence there exists 120575 ge 0 such that

lim119899rarrinfin

119863119899= 120575 (19)

From (16) it is easy to see that

119889119899lt 2119863119899 (20)

Thus the sequence (119889119899) is bounded and so there is 119889 ge 0

such that lim sup119899rarrinfin

119889119899= 119889 and hence a subsequence (119889

119899119896

)

of (119889119899) such that lim

119896rarrinfin119889119899119896

= 119889+ From (17)we have119863

119899119896

+1le

Ψ(119889119899119896

)119863119899119896

and thus

120575 = lim sup119896rarrinfin

119863119899119896

+1

le (lim sup119896rarrinfin

Ψ(119889119899119896

))(lim sup119896rarrinfin

119863119899119896

)

le lim sup119889119899

119896

rarr119889+

Ψ(119889119899119896

) 120575

(21)

This together with the fact lim sup119903rarr 119905+Φ(119903) lt 1 for each

119905 isin [0infin) implies that 120575 = 0 Then from (19) and (20) wederive that lim

119899rarrinfin119889119899= 0

Claim 1 (119909119899) is a left 119870-Cauchy sequence

Now let 120572 = lim sup119889119899

rarr0+Ψ(119889119899) and 119902 such that 120572 lt 119902 lt

1 This choice of 119902 is always possible since 120572 lt 1 Then thereis 1198990such that Ψ(119889

119899) lt 119902 for all 119899 ge 119899

0 So from (18) we have

119863119899+1le 119902119863

119899for all 119899 ge 119899

0 Then by induction we get 119863

119899le

119902119899minus11989901198631198990

for all 119899 ge 1198990+ 1 Combining this and inequality

(20) we get

119898

sum

119896=1198990

119889 (119909119896 119909119896+1) le 2

119898

sum

119896=1198990

119902119896minus11989901198631198990

le 21

1 minus 1199021198631198990

(22)

for all119898 gt 119899 ge 1198990+1 Hence (119909

119899) is a left119870-Cauchy sequence

According to the left 119870-completeness of (119883 119889) thereexists 119909lowast isin 119883 such that 119909

119899

119889

997888rarr 119909lowast

Claim 2 119909lowast is a startpoint of 119879Observe that the sequence 119863

119899= (119891119909

119899) = (119867(119909

119899 119879119909119899))

converges to 0 Since 119891 is 120591(119889)-lower semicontinuous (assupremumof 120591(119889)-lower semicontinuous functions) we have

0 le 119891 (119909lowast) le lim inf119899rarrinfin

119891 (119909119899) = 0 (23)

Hence 119891(119909lowast) = 0 that is119867(119909lowast 119879119909lowast) = 0This completes the proof

We give below an example to illustrate the theorem

Example 10 Let 119883 = [0 1] and 119889 119883 times 119883 rarr R be themapping defined by 119889(119886 119887) = max119886 minus 119887 0 Then 119889 is a1198790-quasipseudometric on119883 Observe that any left119870-Cauchy

sequence in (119883 119889) is 119889-convergent to 0 Indeed if (119909119899) is a left

119870-Cauchy sequence for every 120598 gt 0 there exists 1198990isin N such

that

forall119899 119896 1198990le 119896 le 119899 119889 (119909

119896 119909119899) lt 120598 (24)

This entails that forall119899 1198990lt 119899

119889 (0 119909119899) le 119889 (0 119909

119899minus1) + 119889 (119909

119899minus1 119909119899) = 0 + 119889 (119909

119899minus1 119909119899) lt 120598

(25)

Hence 119889(0 119909119899) rarr 0 that is 119909

119899

119889

997888rarr 0 Therefore (119883 119889) isleft 119870-complete Let 119879 119883 rarr 119862119861(119883) be such that

119879 (119909) =

1

21199092 for 119909 isin [0 15

32) cup (

15

32 1]

17

961

4 for 119909 = 15

32

(26)

Let Φ [0infin) rarr [0 34) sub [0 1) be defined by

Φ (119903) =

3

2119903 for 119903 isin [0 7

24) cup (

7

241

2)

425

768 for 119903 = 7

24

1

2 for 119903 isin [1

2infin)

(27)

An explicit computation of 119891(119909) = 119867(119909 119879119909) gives

119891 (119909) =

119909 minus1

21199092 for 119909 isin [0 15

32) cup (

15

32 1]

28

96 for 119909 = 15

32

(28)

Moreover for each 119909 isin [0 1532) cup (1532 1] 119910 =

(12)1199092 and we have

119867(119910 119879119910) le Φ (119867 (119909 119910))119867 (119909 119910) (29)

Of course inequality (29) also holds in the case of119909 = 1532 and 119910 = 1796 Therefore all assumptions ofTheorem 9 are satisfied and the endpoint of 119879 is 119909 = 0

4 Journal of Operators

Remark 11 In fact every sequence in (119883 119889) 119889-converges to0

Corollary 12 Let (119883 119889) be a right 119870-complete quasipseudo-metric space Let 119879 119883 rarr 119862119861(119883) be a set-valued map anddefine 119891 119883 rarr R as 119891(119909) = 119867(119879119909 119909) Let Φ [0infin) rarr[0 1) be a function such that lim sup

119903rarr 119905+Φ(119903) lt 1 for each

119905 isin [0infin) Moreover assume that for any 119909 isin 119883 there exists119910 isin 119879119909 satisfying

119867(119879119910 119910) le Φ (119867 (119910 119909))119867 (119910 119909) (30)

and then 119879 has an endpoint

Corollary 13 Let (119883 119889) be a bicomplete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and119891 119883 rarr R defined by 119891(119909) = 119867

119904(119879119909 119909) =

max119867(119879119909 119909)119867(119909 119879119909) If there exists 119888 isin (0 1) suchthat for all 119909 isin 119883 there exists 119910 isin 119865119909 satisfying

119867119904(119910 119865119910) le min Φ (119886) 119886 Φ (119887) 119887 (31)

where 119886 = 119867(119910 119909) and 119887 = 119867(119909 119910) then 119879 has a fixedpoint

Proof We give here the main idea of the proof Observe thatinequality (31) guarantees that the sequence (119909

119899) constructed

in the proof ofTheorem 9 is a 119889119904-Cauchy sequence and hence119889119904-converges to some 119909lowast Using the fact that 119891 is 120591(119889119904)-

lower semicontinuous (as supremum of 120591(119889119904)-continuousfunctions) we have

0 le 119891 (119909lowast) le lim inf119899rarrinfin

119891 (119909119899) = 0 (32)

Hence 119891(119909lowast) = 0 that is 119867(119909lowast 119879119909lowast) = 0 =

119867(119879119909lowast 119909lowast) and we are done

The following theorem is the second generalization thatwe propose

Theorem 14 Let (119883 119889) be a left119870-complete quasipseudomet-ric space Let 119879 119883 rarr 119862119861(119883) be a set-valued map anddefine 119891 119883 rarr R as 119891(119909) = 119867(119909 119879119909) Let 119887 [0infin) rarr[119886 1) 119886 gt 0 be a nondecreasing function Let Φ [0infin) rarr[0 1) be a function such that Φ(119905) lt 119887(119905) for each 119905 isin [0infin)and lim sup

119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for each 119905 isin [0infin)

Moreover assume that for any 119909 isin 119883 there exists 119910 isin 119879119909satisfying

119867(119910 119879119910) le Φ (119867 (119909 119910))119867 (119909 119910) (33)

and then 119879 has a startpoint

Proof Observe that because 119887(119867(119909 119910)) lt 1 for all 119909 119910 isin119883

119887 (119867 (119909 119910))119867 (119909 119910) le 119867 (119909 119879119909) (34)

for any 119910 isin 119879119909 Let 1199090isin 119883 be arbitrary Then we can choose

1199091isin 1198791199090such that

119887 (119867 (1199090 1199091))119867 (119909

0 1199091) le 119867 (119909

0 1198791199090) (35)

119867(1199091 1198791199091) le Φ (119867 (119909

0 1199091))119867 (119909

0 1199091) (36)

Define the function Ψ [0infin) rarr [0 1) by Ψ(119905) =Φ(119905)119887(119905) Hence (35) and (36) together sum to

119867(1199091 1198791199091) le Ψ (119867 (119909

0 1199091))119867 (119909

0 1198791199090) (37)

Now we choose 1199092isin 1198791199091such that

119887 (119867 (1199091 1199092))119867 (119909

1 1199092) le 119867 (119909

1 1198791199091)

119867 (1199092 1198791199092) le Φ (119867 (119909

1 1199092))119867 (119909

1 1199092)

(38)

which lead to

119867(1199092 1198791199092) le Ψ (119867 (119909

1 1199092))119867 (119909

1 1198791199091) (39)

Continuing this process we get an iterative sequence (119909119899)

where 119909119899+1isin 119879119909119899sube 119883 and denoting 119889

119899= 119867(119909

119899 119909119899+1)

and119863119899= 119867(119909

119899 119879119909119899) for all 119899 ge 0 we can write that

119887 (119889119899) 119889119899le 119863119899 119863

119899+1le Φ (119889

119899) 119889119899 (40)

for all 119899 ge 0 Hence

119863119899+1le Ψ (119889

119899)119863119899 (41)

If 119863119896= 0 for some 119896 ge 0 then we trivially have

lim119899rarrinfin

119863119899= 0 and the conclusion is immediate So without

loss of generality we can assume that 119863119899gt 0 for all 119899 ge 0

and from (41) we have119863119899+1lt 119863119899for all 119899 ge 0

Observe also that if for some 119899 119889119899le 119889119899+1

we are led toa contradiction Indeed from (40) and using the fact that thefunction 119887 is nondecreasing we have

119889119899le 119889119899+1leΦ (119889119899)

119887 (119889119899+1)119889119899leΦ (119889119899)

119887 (119889119899)119889119899= Ψ (119889

119899) 119889119899lt 119889119899

(42)

Hence 119889119899+1lt 119889119899for all 119899 ge 0 Thus there exist 120575 ge 0 and

119889 ge 0 such that lim119899rarrinfin

119863119899= 120575 and lim

119899rarrinfin119889119899= 119889 From

(41) we get

120575 le (lim sup119899rarrinfin

Ψ (119889119899)) 120575 = lim sup

119889119899

rarr119889+

Ψ (119889119899) 120575 (43)

and hence 120575 = 0 (since lim sup119903rarr 119905+Φ(119903) lt 1 for all 119905 isin

[0infin)) Moreover since

119886119889119899le 119887 (119889

119899) 119889119899le 119863119899le Ψ (119889

119899minus1)119863119899minus1 (44)

this forces 119889 to be 0 that is lim119899rarrinfin

119889119899= 0

Furthermore setting 120572 = lim sup119889119899

rarr0+Ψ(119889119899) and letting

119902 gt 0 be a positive number such that 120572 lt 119902 lt 1 there is 1198990

such that Ψ(119889119899) lt 119902 for all 119899 ge 119899

0 Hence from (41) and (44)

we get

119886119889119899le Ψ (119889

119899minus1)119863119899minus1le 119902119863119899minus1le sdot sdot sdot le 119902

119899minus11989901198631198990

(45)

for all 119899 ge 1198990 So 119889

119899le (1119886)119902

119899minus11989901198631198990

for all 119899 ge 1198990 Using a

similar argument as the one used in the proof of Theorem 9we conclude that (119909

119899) is a left119870-Cauchy sequence and that its

limit point is a starpoint of 119879

Journal of Operators 5

Example 15 Let 119883 = [0 1) and 119889 119883 times 119883 rarr R be themapping defined by 119889(119886 119887) = max119886 minus 119887 0 Then 119889 is a1198790-quasipseudometric on 119883 We know that (119883 119889) is left 119870-

complete Let 119879 119883 rarr 119862119861(119883) be such that

119879 (119909) = 1

21199092 (46)

Let Φ [0infin) rarr [0 34) sub [0 1) be defined by

Φ (119903) =

3

2119903 for 119903 isin [0 1

2)

0 for 119903 isin [12infin)

(47)

Let 119887 [0infin) rarr [23 1) be defined by

119887 (119905) =

5

6119905 for 119905 isin [0 1

2)

1

2 for 119905 isin [1

2infin)

(48)

An explicit computation of 119891(119909) = 119867(119909 119879119909) gives

119891 (119909) = 119909 minus1

21199092 for 119909 isin [0 1) (49)

Moreover for each 119909 isin [0 1) there exists 119910 = (12)1199092and we have

119867(119910 119879119910) le3

2119867 (119909 119910)119867 (119909 119910)

= Φ (119867 (119909 119910))119867 (119909 119910)

(50)

Therefore all assumptions ofTheorem 14 are satisfied andthe endpoint of 119879 is 119909 = 0

Corollary 16 Let (119883 119889) be a right 119870-complete quasipseudo-metric space Let 119879 119883 rarr 119862119861(119883) be a set-valued mapand define 119891 119883 rarr R as 119891(119909) = 119867

minus1(119909 119879119909) Let

119887 [0infin) rarr [119886 1) 119886 gt 0 be a nondecreasing function LetΦ [0infin) rarr [0 1) be a function such that Φ(119905) lt 119887(119905) foreach 119905 isin [0infin) and lim sup

119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for

each 119905 isin [0infin) Moreover assume that for any 119909 isin 119883 thereexists 119910 isin 119879119909 satisfying

119867minus1(119910 119879119910) le Φ (119867

minus1(119909 119910))119867

minus1(119909 119910) (51)

Φ(119905) lt 119887(119905) and lim sup119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for each

119905 isin [0infin) Moreover assume that for any 119909 isin 119883 there exists119910 isin 119879119909 satisfying and then 119879 has an endpoint

Corollary 17 Let (119883 119889) be a bicomplete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and define119891 119883 rarr R as 119891(119909) = 119867119904(119909 119879119909) Let 119887 [0infin) rarr [119886 1)119886 gt 0 be a nondecreasing function Let Φ [0infin) rarr [0 1)

be a function such that Φ(119905) lt 119887(119905) and lim sup119903rarr 119905+Φ(119903) lt

lim sup119903rarr 119905+119887(119903) for each 119905 isin [0infin) Moreover assume that

for any 119909 isin 119883 there exists 119910 isin 119879119909 satisfying

119867119904(119910 119879119910) le min Φ (119886) 119886 Φ (119887) 119887 (52)

where 119886 = 119867(119909 119910) and 119887 = 119867minus1(119909 119910) Then 119879 has afixed point

Remark 18 All the results given remain true when we replaceaccordingly the bicomplete quasipseudometric space (119883 119889)by a left Smyth sequentially completeleft 119870-complete or aright Smyth sequentially completeright 119870-complete space

Conflict of Interests

The author declares that there is no conflict of interestsregarding the publication of this paper

References

[1] Y U Gaba ldquoStartpoints and (120572 120574)-contractions in quasi-pseudometric spacesrdquo Journal of Mathematics vol 2014 ArticleID 709253 8 pages 2014

[2] H-P Kunzi and C Ryser ldquoThe Bourbaki quasi-uniformityrdquoTopology Proceedings vol 20 pp 161ndash183 1995

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Page 3: Research Article New Results in the Startpoint Theory for ...downloads.hindawi.com/archive/2014/741818.pdf · Research Article New Results in the Startpoint Theory for Quasipseudometric

Journal of Operators 3

and from (9) we get

119867(1199092 1198791199092) le Ψ (119867 (119909

1 1199092))119867 (119909

1 1198791199091) (15)

Continuing this process we obtain a sequence (119909119899)where

119909119899+1isin 119879119909119899sube 119883 with

119867(119909119899 119909119899+1)

le [2 minus Φ (119867 (119909119899 119909119899+1))]119867 (119909

119899 119879119909119899)

(16)

119867(119909119899+1 119879119909119899+1) le Ψ (119867 (119909

119899 119909119899+1))119867 (119909

119899 119879119909119899)

119899 = 1 2

(17)

For simplicity denote 119889119899= 119867(119909

119899 119909119899+1) and 119863

119899=

119867(119909119899 119879119909119899) for all 119899 ge 0 So from (17) we can write

119863119899+1le Ψ (119889

119899)119863119899le 119863119899 (18)

for all 119899 ge 0 Hence (119863119899) is a strictly decreasing sequence and

hence there exists 120575 ge 0 such that

lim119899rarrinfin

119863119899= 120575 (19)

From (16) it is easy to see that

119889119899lt 2119863119899 (20)

Thus the sequence (119889119899) is bounded and so there is 119889 ge 0

such that lim sup119899rarrinfin

119889119899= 119889 and hence a subsequence (119889

119899119896

)

of (119889119899) such that lim

119896rarrinfin119889119899119896

= 119889+ From (17)we have119863

119899119896

+1le

Ψ(119889119899119896

)119863119899119896

and thus

120575 = lim sup119896rarrinfin

119863119899119896

+1

le (lim sup119896rarrinfin

Ψ(119889119899119896

))(lim sup119896rarrinfin

119863119899119896

)

le lim sup119889119899

119896

rarr119889+

Ψ(119889119899119896

) 120575

(21)

This together with the fact lim sup119903rarr 119905+Φ(119903) lt 1 for each

119905 isin [0infin) implies that 120575 = 0 Then from (19) and (20) wederive that lim

119899rarrinfin119889119899= 0

Claim 1 (119909119899) is a left 119870-Cauchy sequence

Now let 120572 = lim sup119889119899

rarr0+Ψ(119889119899) and 119902 such that 120572 lt 119902 lt

1 This choice of 119902 is always possible since 120572 lt 1 Then thereis 1198990such that Ψ(119889

119899) lt 119902 for all 119899 ge 119899

0 So from (18) we have

119863119899+1le 119902119863

119899for all 119899 ge 119899

0 Then by induction we get 119863

119899le

119902119899minus11989901198631198990

for all 119899 ge 1198990+ 1 Combining this and inequality

(20) we get

119898

sum

119896=1198990

119889 (119909119896 119909119896+1) le 2

119898

sum

119896=1198990

119902119896minus11989901198631198990

le 21

1 minus 1199021198631198990

(22)

for all119898 gt 119899 ge 1198990+1 Hence (119909

119899) is a left119870-Cauchy sequence

According to the left 119870-completeness of (119883 119889) thereexists 119909lowast isin 119883 such that 119909

119899

119889

997888rarr 119909lowast

Claim 2 119909lowast is a startpoint of 119879Observe that the sequence 119863

119899= (119891119909

119899) = (119867(119909

119899 119879119909119899))

converges to 0 Since 119891 is 120591(119889)-lower semicontinuous (assupremumof 120591(119889)-lower semicontinuous functions) we have

0 le 119891 (119909lowast) le lim inf119899rarrinfin

119891 (119909119899) = 0 (23)

Hence 119891(119909lowast) = 0 that is119867(119909lowast 119879119909lowast) = 0This completes the proof

We give below an example to illustrate the theorem

Example 10 Let 119883 = [0 1] and 119889 119883 times 119883 rarr R be themapping defined by 119889(119886 119887) = max119886 minus 119887 0 Then 119889 is a1198790-quasipseudometric on119883 Observe that any left119870-Cauchy

sequence in (119883 119889) is 119889-convergent to 0 Indeed if (119909119899) is a left

119870-Cauchy sequence for every 120598 gt 0 there exists 1198990isin N such

that

forall119899 119896 1198990le 119896 le 119899 119889 (119909

119896 119909119899) lt 120598 (24)

This entails that forall119899 1198990lt 119899

119889 (0 119909119899) le 119889 (0 119909

119899minus1) + 119889 (119909

119899minus1 119909119899) = 0 + 119889 (119909

119899minus1 119909119899) lt 120598

(25)

Hence 119889(0 119909119899) rarr 0 that is 119909

119899

119889

997888rarr 0 Therefore (119883 119889) isleft 119870-complete Let 119879 119883 rarr 119862119861(119883) be such that

119879 (119909) =

1

21199092 for 119909 isin [0 15

32) cup (

15

32 1]

17

961

4 for 119909 = 15

32

(26)

Let Φ [0infin) rarr [0 34) sub [0 1) be defined by

Φ (119903) =

3

2119903 for 119903 isin [0 7

24) cup (

7

241

2)

425

768 for 119903 = 7

24

1

2 for 119903 isin [1

2infin)

(27)

An explicit computation of 119891(119909) = 119867(119909 119879119909) gives

119891 (119909) =

119909 minus1

21199092 for 119909 isin [0 15

32) cup (

15

32 1]

28

96 for 119909 = 15

32

(28)

Moreover for each 119909 isin [0 1532) cup (1532 1] 119910 =

(12)1199092 and we have

119867(119910 119879119910) le Φ (119867 (119909 119910))119867 (119909 119910) (29)

Of course inequality (29) also holds in the case of119909 = 1532 and 119910 = 1796 Therefore all assumptions ofTheorem 9 are satisfied and the endpoint of 119879 is 119909 = 0

4 Journal of Operators

Remark 11 In fact every sequence in (119883 119889) 119889-converges to0

Corollary 12 Let (119883 119889) be a right 119870-complete quasipseudo-metric space Let 119879 119883 rarr 119862119861(119883) be a set-valued map anddefine 119891 119883 rarr R as 119891(119909) = 119867(119879119909 119909) Let Φ [0infin) rarr[0 1) be a function such that lim sup

119903rarr 119905+Φ(119903) lt 1 for each

119905 isin [0infin) Moreover assume that for any 119909 isin 119883 there exists119910 isin 119879119909 satisfying

119867(119879119910 119910) le Φ (119867 (119910 119909))119867 (119910 119909) (30)

and then 119879 has an endpoint

Corollary 13 Let (119883 119889) be a bicomplete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and119891 119883 rarr R defined by 119891(119909) = 119867

119904(119879119909 119909) =

max119867(119879119909 119909)119867(119909 119879119909) If there exists 119888 isin (0 1) suchthat for all 119909 isin 119883 there exists 119910 isin 119865119909 satisfying

119867119904(119910 119865119910) le min Φ (119886) 119886 Φ (119887) 119887 (31)

where 119886 = 119867(119910 119909) and 119887 = 119867(119909 119910) then 119879 has a fixedpoint

Proof We give here the main idea of the proof Observe thatinequality (31) guarantees that the sequence (119909

119899) constructed

in the proof ofTheorem 9 is a 119889119904-Cauchy sequence and hence119889119904-converges to some 119909lowast Using the fact that 119891 is 120591(119889119904)-

lower semicontinuous (as supremum of 120591(119889119904)-continuousfunctions) we have

0 le 119891 (119909lowast) le lim inf119899rarrinfin

119891 (119909119899) = 0 (32)

Hence 119891(119909lowast) = 0 that is 119867(119909lowast 119879119909lowast) = 0 =

119867(119879119909lowast 119909lowast) and we are done

The following theorem is the second generalization thatwe propose

Theorem 14 Let (119883 119889) be a left119870-complete quasipseudomet-ric space Let 119879 119883 rarr 119862119861(119883) be a set-valued map anddefine 119891 119883 rarr R as 119891(119909) = 119867(119909 119879119909) Let 119887 [0infin) rarr[119886 1) 119886 gt 0 be a nondecreasing function Let Φ [0infin) rarr[0 1) be a function such that Φ(119905) lt 119887(119905) for each 119905 isin [0infin)and lim sup

119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for each 119905 isin [0infin)

Moreover assume that for any 119909 isin 119883 there exists 119910 isin 119879119909satisfying

119867(119910 119879119910) le Φ (119867 (119909 119910))119867 (119909 119910) (33)

and then 119879 has a startpoint

Proof Observe that because 119887(119867(119909 119910)) lt 1 for all 119909 119910 isin119883

119887 (119867 (119909 119910))119867 (119909 119910) le 119867 (119909 119879119909) (34)

for any 119910 isin 119879119909 Let 1199090isin 119883 be arbitrary Then we can choose

1199091isin 1198791199090such that

119887 (119867 (1199090 1199091))119867 (119909

0 1199091) le 119867 (119909

0 1198791199090) (35)

119867(1199091 1198791199091) le Φ (119867 (119909

0 1199091))119867 (119909

0 1199091) (36)

Define the function Ψ [0infin) rarr [0 1) by Ψ(119905) =Φ(119905)119887(119905) Hence (35) and (36) together sum to

119867(1199091 1198791199091) le Ψ (119867 (119909

0 1199091))119867 (119909

0 1198791199090) (37)

Now we choose 1199092isin 1198791199091such that

119887 (119867 (1199091 1199092))119867 (119909

1 1199092) le 119867 (119909

1 1198791199091)

119867 (1199092 1198791199092) le Φ (119867 (119909

1 1199092))119867 (119909

1 1199092)

(38)

which lead to

119867(1199092 1198791199092) le Ψ (119867 (119909

1 1199092))119867 (119909

1 1198791199091) (39)

Continuing this process we get an iterative sequence (119909119899)

where 119909119899+1isin 119879119909119899sube 119883 and denoting 119889

119899= 119867(119909

119899 119909119899+1)

and119863119899= 119867(119909

119899 119879119909119899) for all 119899 ge 0 we can write that

119887 (119889119899) 119889119899le 119863119899 119863

119899+1le Φ (119889

119899) 119889119899 (40)

for all 119899 ge 0 Hence

119863119899+1le Ψ (119889

119899)119863119899 (41)

If 119863119896= 0 for some 119896 ge 0 then we trivially have

lim119899rarrinfin

119863119899= 0 and the conclusion is immediate So without

loss of generality we can assume that 119863119899gt 0 for all 119899 ge 0

and from (41) we have119863119899+1lt 119863119899for all 119899 ge 0

Observe also that if for some 119899 119889119899le 119889119899+1

we are led toa contradiction Indeed from (40) and using the fact that thefunction 119887 is nondecreasing we have

119889119899le 119889119899+1leΦ (119889119899)

119887 (119889119899+1)119889119899leΦ (119889119899)

119887 (119889119899)119889119899= Ψ (119889

119899) 119889119899lt 119889119899

(42)

Hence 119889119899+1lt 119889119899for all 119899 ge 0 Thus there exist 120575 ge 0 and

119889 ge 0 such that lim119899rarrinfin

119863119899= 120575 and lim

119899rarrinfin119889119899= 119889 From

(41) we get

120575 le (lim sup119899rarrinfin

Ψ (119889119899)) 120575 = lim sup

119889119899

rarr119889+

Ψ (119889119899) 120575 (43)

and hence 120575 = 0 (since lim sup119903rarr 119905+Φ(119903) lt 1 for all 119905 isin

[0infin)) Moreover since

119886119889119899le 119887 (119889

119899) 119889119899le 119863119899le Ψ (119889

119899minus1)119863119899minus1 (44)

this forces 119889 to be 0 that is lim119899rarrinfin

119889119899= 0

Furthermore setting 120572 = lim sup119889119899

rarr0+Ψ(119889119899) and letting

119902 gt 0 be a positive number such that 120572 lt 119902 lt 1 there is 1198990

such that Ψ(119889119899) lt 119902 for all 119899 ge 119899

0 Hence from (41) and (44)

we get

119886119889119899le Ψ (119889

119899minus1)119863119899minus1le 119902119863119899minus1le sdot sdot sdot le 119902

119899minus11989901198631198990

(45)

for all 119899 ge 1198990 So 119889

119899le (1119886)119902

119899minus11989901198631198990

for all 119899 ge 1198990 Using a

similar argument as the one used in the proof of Theorem 9we conclude that (119909

119899) is a left119870-Cauchy sequence and that its

limit point is a starpoint of 119879

Journal of Operators 5

Example 15 Let 119883 = [0 1) and 119889 119883 times 119883 rarr R be themapping defined by 119889(119886 119887) = max119886 minus 119887 0 Then 119889 is a1198790-quasipseudometric on 119883 We know that (119883 119889) is left 119870-

complete Let 119879 119883 rarr 119862119861(119883) be such that

119879 (119909) = 1

21199092 (46)

Let Φ [0infin) rarr [0 34) sub [0 1) be defined by

Φ (119903) =

3

2119903 for 119903 isin [0 1

2)

0 for 119903 isin [12infin)

(47)

Let 119887 [0infin) rarr [23 1) be defined by

119887 (119905) =

5

6119905 for 119905 isin [0 1

2)

1

2 for 119905 isin [1

2infin)

(48)

An explicit computation of 119891(119909) = 119867(119909 119879119909) gives

119891 (119909) = 119909 minus1

21199092 for 119909 isin [0 1) (49)

Moreover for each 119909 isin [0 1) there exists 119910 = (12)1199092and we have

119867(119910 119879119910) le3

2119867 (119909 119910)119867 (119909 119910)

= Φ (119867 (119909 119910))119867 (119909 119910)

(50)

Therefore all assumptions ofTheorem 14 are satisfied andthe endpoint of 119879 is 119909 = 0

Corollary 16 Let (119883 119889) be a right 119870-complete quasipseudo-metric space Let 119879 119883 rarr 119862119861(119883) be a set-valued mapand define 119891 119883 rarr R as 119891(119909) = 119867

minus1(119909 119879119909) Let

119887 [0infin) rarr [119886 1) 119886 gt 0 be a nondecreasing function LetΦ [0infin) rarr [0 1) be a function such that Φ(119905) lt 119887(119905) foreach 119905 isin [0infin) and lim sup

119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for

each 119905 isin [0infin) Moreover assume that for any 119909 isin 119883 thereexists 119910 isin 119879119909 satisfying

119867minus1(119910 119879119910) le Φ (119867

minus1(119909 119910))119867

minus1(119909 119910) (51)

Φ(119905) lt 119887(119905) and lim sup119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for each

119905 isin [0infin) Moreover assume that for any 119909 isin 119883 there exists119910 isin 119879119909 satisfying and then 119879 has an endpoint

Corollary 17 Let (119883 119889) be a bicomplete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and define119891 119883 rarr R as 119891(119909) = 119867119904(119909 119879119909) Let 119887 [0infin) rarr [119886 1)119886 gt 0 be a nondecreasing function Let Φ [0infin) rarr [0 1)

be a function such that Φ(119905) lt 119887(119905) and lim sup119903rarr 119905+Φ(119903) lt

lim sup119903rarr 119905+119887(119903) for each 119905 isin [0infin) Moreover assume that

for any 119909 isin 119883 there exists 119910 isin 119879119909 satisfying

119867119904(119910 119879119910) le min Φ (119886) 119886 Φ (119887) 119887 (52)

where 119886 = 119867(119909 119910) and 119887 = 119867minus1(119909 119910) Then 119879 has afixed point

Remark 18 All the results given remain true when we replaceaccordingly the bicomplete quasipseudometric space (119883 119889)by a left Smyth sequentially completeleft 119870-complete or aright Smyth sequentially completeright 119870-complete space

Conflict of Interests

The author declares that there is no conflict of interestsregarding the publication of this paper

References

[1] Y U Gaba ldquoStartpoints and (120572 120574)-contractions in quasi-pseudometric spacesrdquo Journal of Mathematics vol 2014 ArticleID 709253 8 pages 2014

[2] H-P Kunzi and C Ryser ldquoThe Bourbaki quasi-uniformityrdquoTopology Proceedings vol 20 pp 161ndash183 1995

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Page 4: Research Article New Results in the Startpoint Theory for ...downloads.hindawi.com/archive/2014/741818.pdf · Research Article New Results in the Startpoint Theory for Quasipseudometric

4 Journal of Operators

Remark 11 In fact every sequence in (119883 119889) 119889-converges to0

Corollary 12 Let (119883 119889) be a right 119870-complete quasipseudo-metric space Let 119879 119883 rarr 119862119861(119883) be a set-valued map anddefine 119891 119883 rarr R as 119891(119909) = 119867(119879119909 119909) Let Φ [0infin) rarr[0 1) be a function such that lim sup

119903rarr 119905+Φ(119903) lt 1 for each

119905 isin [0infin) Moreover assume that for any 119909 isin 119883 there exists119910 isin 119879119909 satisfying

119867(119879119910 119910) le Φ (119867 (119910 119909))119867 (119910 119909) (30)

and then 119879 has an endpoint

Corollary 13 Let (119883 119889) be a bicomplete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and119891 119883 rarr R defined by 119891(119909) = 119867

119904(119879119909 119909) =

max119867(119879119909 119909)119867(119909 119879119909) If there exists 119888 isin (0 1) suchthat for all 119909 isin 119883 there exists 119910 isin 119865119909 satisfying

119867119904(119910 119865119910) le min Φ (119886) 119886 Φ (119887) 119887 (31)

where 119886 = 119867(119910 119909) and 119887 = 119867(119909 119910) then 119879 has a fixedpoint

Proof We give here the main idea of the proof Observe thatinequality (31) guarantees that the sequence (119909

119899) constructed

in the proof ofTheorem 9 is a 119889119904-Cauchy sequence and hence119889119904-converges to some 119909lowast Using the fact that 119891 is 120591(119889119904)-

lower semicontinuous (as supremum of 120591(119889119904)-continuousfunctions) we have

0 le 119891 (119909lowast) le lim inf119899rarrinfin

119891 (119909119899) = 0 (32)

Hence 119891(119909lowast) = 0 that is 119867(119909lowast 119879119909lowast) = 0 =

119867(119879119909lowast 119909lowast) and we are done

The following theorem is the second generalization thatwe propose

Theorem 14 Let (119883 119889) be a left119870-complete quasipseudomet-ric space Let 119879 119883 rarr 119862119861(119883) be a set-valued map anddefine 119891 119883 rarr R as 119891(119909) = 119867(119909 119879119909) Let 119887 [0infin) rarr[119886 1) 119886 gt 0 be a nondecreasing function Let Φ [0infin) rarr[0 1) be a function such that Φ(119905) lt 119887(119905) for each 119905 isin [0infin)and lim sup

119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for each 119905 isin [0infin)

Moreover assume that for any 119909 isin 119883 there exists 119910 isin 119879119909satisfying

119867(119910 119879119910) le Φ (119867 (119909 119910))119867 (119909 119910) (33)

and then 119879 has a startpoint

Proof Observe that because 119887(119867(119909 119910)) lt 1 for all 119909 119910 isin119883

119887 (119867 (119909 119910))119867 (119909 119910) le 119867 (119909 119879119909) (34)

for any 119910 isin 119879119909 Let 1199090isin 119883 be arbitrary Then we can choose

1199091isin 1198791199090such that

119887 (119867 (1199090 1199091))119867 (119909

0 1199091) le 119867 (119909

0 1198791199090) (35)

119867(1199091 1198791199091) le Φ (119867 (119909

0 1199091))119867 (119909

0 1199091) (36)

Define the function Ψ [0infin) rarr [0 1) by Ψ(119905) =Φ(119905)119887(119905) Hence (35) and (36) together sum to

119867(1199091 1198791199091) le Ψ (119867 (119909

0 1199091))119867 (119909

0 1198791199090) (37)

Now we choose 1199092isin 1198791199091such that

119887 (119867 (1199091 1199092))119867 (119909

1 1199092) le 119867 (119909

1 1198791199091)

119867 (1199092 1198791199092) le Φ (119867 (119909

1 1199092))119867 (119909

1 1199092)

(38)

which lead to

119867(1199092 1198791199092) le Ψ (119867 (119909

1 1199092))119867 (119909

1 1198791199091) (39)

Continuing this process we get an iterative sequence (119909119899)

where 119909119899+1isin 119879119909119899sube 119883 and denoting 119889

119899= 119867(119909

119899 119909119899+1)

and119863119899= 119867(119909

119899 119879119909119899) for all 119899 ge 0 we can write that

119887 (119889119899) 119889119899le 119863119899 119863

119899+1le Φ (119889

119899) 119889119899 (40)

for all 119899 ge 0 Hence

119863119899+1le Ψ (119889

119899)119863119899 (41)

If 119863119896= 0 for some 119896 ge 0 then we trivially have

lim119899rarrinfin

119863119899= 0 and the conclusion is immediate So without

loss of generality we can assume that 119863119899gt 0 for all 119899 ge 0

and from (41) we have119863119899+1lt 119863119899for all 119899 ge 0

Observe also that if for some 119899 119889119899le 119889119899+1

we are led toa contradiction Indeed from (40) and using the fact that thefunction 119887 is nondecreasing we have

119889119899le 119889119899+1leΦ (119889119899)

119887 (119889119899+1)119889119899leΦ (119889119899)

119887 (119889119899)119889119899= Ψ (119889

119899) 119889119899lt 119889119899

(42)

Hence 119889119899+1lt 119889119899for all 119899 ge 0 Thus there exist 120575 ge 0 and

119889 ge 0 such that lim119899rarrinfin

119863119899= 120575 and lim

119899rarrinfin119889119899= 119889 From

(41) we get

120575 le (lim sup119899rarrinfin

Ψ (119889119899)) 120575 = lim sup

119889119899

rarr119889+

Ψ (119889119899) 120575 (43)

and hence 120575 = 0 (since lim sup119903rarr 119905+Φ(119903) lt 1 for all 119905 isin

[0infin)) Moreover since

119886119889119899le 119887 (119889

119899) 119889119899le 119863119899le Ψ (119889

119899minus1)119863119899minus1 (44)

this forces 119889 to be 0 that is lim119899rarrinfin

119889119899= 0

Furthermore setting 120572 = lim sup119889119899

rarr0+Ψ(119889119899) and letting

119902 gt 0 be a positive number such that 120572 lt 119902 lt 1 there is 1198990

such that Ψ(119889119899) lt 119902 for all 119899 ge 119899

0 Hence from (41) and (44)

we get

119886119889119899le Ψ (119889

119899minus1)119863119899minus1le 119902119863119899minus1le sdot sdot sdot le 119902

119899minus11989901198631198990

(45)

for all 119899 ge 1198990 So 119889

119899le (1119886)119902

119899minus11989901198631198990

for all 119899 ge 1198990 Using a

similar argument as the one used in the proof of Theorem 9we conclude that (119909

119899) is a left119870-Cauchy sequence and that its

limit point is a starpoint of 119879

Journal of Operators 5

Example 15 Let 119883 = [0 1) and 119889 119883 times 119883 rarr R be themapping defined by 119889(119886 119887) = max119886 minus 119887 0 Then 119889 is a1198790-quasipseudometric on 119883 We know that (119883 119889) is left 119870-

complete Let 119879 119883 rarr 119862119861(119883) be such that

119879 (119909) = 1

21199092 (46)

Let Φ [0infin) rarr [0 34) sub [0 1) be defined by

Φ (119903) =

3

2119903 for 119903 isin [0 1

2)

0 for 119903 isin [12infin)

(47)

Let 119887 [0infin) rarr [23 1) be defined by

119887 (119905) =

5

6119905 for 119905 isin [0 1

2)

1

2 for 119905 isin [1

2infin)

(48)

An explicit computation of 119891(119909) = 119867(119909 119879119909) gives

119891 (119909) = 119909 minus1

21199092 for 119909 isin [0 1) (49)

Moreover for each 119909 isin [0 1) there exists 119910 = (12)1199092and we have

119867(119910 119879119910) le3

2119867 (119909 119910)119867 (119909 119910)

= Φ (119867 (119909 119910))119867 (119909 119910)

(50)

Therefore all assumptions ofTheorem 14 are satisfied andthe endpoint of 119879 is 119909 = 0

Corollary 16 Let (119883 119889) be a right 119870-complete quasipseudo-metric space Let 119879 119883 rarr 119862119861(119883) be a set-valued mapand define 119891 119883 rarr R as 119891(119909) = 119867

minus1(119909 119879119909) Let

119887 [0infin) rarr [119886 1) 119886 gt 0 be a nondecreasing function LetΦ [0infin) rarr [0 1) be a function such that Φ(119905) lt 119887(119905) foreach 119905 isin [0infin) and lim sup

119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for

each 119905 isin [0infin) Moreover assume that for any 119909 isin 119883 thereexists 119910 isin 119879119909 satisfying

119867minus1(119910 119879119910) le Φ (119867

minus1(119909 119910))119867

minus1(119909 119910) (51)

Φ(119905) lt 119887(119905) and lim sup119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for each

119905 isin [0infin) Moreover assume that for any 119909 isin 119883 there exists119910 isin 119879119909 satisfying and then 119879 has an endpoint

Corollary 17 Let (119883 119889) be a bicomplete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and define119891 119883 rarr R as 119891(119909) = 119867119904(119909 119879119909) Let 119887 [0infin) rarr [119886 1)119886 gt 0 be a nondecreasing function Let Φ [0infin) rarr [0 1)

be a function such that Φ(119905) lt 119887(119905) and lim sup119903rarr 119905+Φ(119903) lt

lim sup119903rarr 119905+119887(119903) for each 119905 isin [0infin) Moreover assume that

for any 119909 isin 119883 there exists 119910 isin 119879119909 satisfying

119867119904(119910 119879119910) le min Φ (119886) 119886 Φ (119887) 119887 (52)

where 119886 = 119867(119909 119910) and 119887 = 119867minus1(119909 119910) Then 119879 has afixed point

Remark 18 All the results given remain true when we replaceaccordingly the bicomplete quasipseudometric space (119883 119889)by a left Smyth sequentially completeleft 119870-complete or aright Smyth sequentially completeright 119870-complete space

Conflict of Interests

The author declares that there is no conflict of interestsregarding the publication of this paper

References

[1] Y U Gaba ldquoStartpoints and (120572 120574)-contractions in quasi-pseudometric spacesrdquo Journal of Mathematics vol 2014 ArticleID 709253 8 pages 2014

[2] H-P Kunzi and C Ryser ldquoThe Bourbaki quasi-uniformityrdquoTopology Proceedings vol 20 pp 161ndash183 1995

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Page 5: Research Article New Results in the Startpoint Theory for ...downloads.hindawi.com/archive/2014/741818.pdf · Research Article New Results in the Startpoint Theory for Quasipseudometric

Journal of Operators 5

Example 15 Let 119883 = [0 1) and 119889 119883 times 119883 rarr R be themapping defined by 119889(119886 119887) = max119886 minus 119887 0 Then 119889 is a1198790-quasipseudometric on 119883 We know that (119883 119889) is left 119870-

complete Let 119879 119883 rarr 119862119861(119883) be such that

119879 (119909) = 1

21199092 (46)

Let Φ [0infin) rarr [0 34) sub [0 1) be defined by

Φ (119903) =

3

2119903 for 119903 isin [0 1

2)

0 for 119903 isin [12infin)

(47)

Let 119887 [0infin) rarr [23 1) be defined by

119887 (119905) =

5

6119905 for 119905 isin [0 1

2)

1

2 for 119905 isin [1

2infin)

(48)

An explicit computation of 119891(119909) = 119867(119909 119879119909) gives

119891 (119909) = 119909 minus1

21199092 for 119909 isin [0 1) (49)

Moreover for each 119909 isin [0 1) there exists 119910 = (12)1199092and we have

119867(119910 119879119910) le3

2119867 (119909 119910)119867 (119909 119910)

= Φ (119867 (119909 119910))119867 (119909 119910)

(50)

Therefore all assumptions ofTheorem 14 are satisfied andthe endpoint of 119879 is 119909 = 0

Corollary 16 Let (119883 119889) be a right 119870-complete quasipseudo-metric space Let 119879 119883 rarr 119862119861(119883) be a set-valued mapand define 119891 119883 rarr R as 119891(119909) = 119867

minus1(119909 119879119909) Let

119887 [0infin) rarr [119886 1) 119886 gt 0 be a nondecreasing function LetΦ [0infin) rarr [0 1) be a function such that Φ(119905) lt 119887(119905) foreach 119905 isin [0infin) and lim sup

119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for

each 119905 isin [0infin) Moreover assume that for any 119909 isin 119883 thereexists 119910 isin 119879119909 satisfying

119867minus1(119910 119879119910) le Φ (119867

minus1(119909 119910))119867

minus1(119909 119910) (51)

Φ(119905) lt 119887(119905) and lim sup119903rarr 119905+Φ(119903) lt lim sup

119903rarr 119905+119887(119903) for each

119905 isin [0infin) Moreover assume that for any 119909 isin 119883 there exists119910 isin 119879119909 satisfying and then 119879 has an endpoint

Corollary 17 Let (119883 119889) be a bicomplete quasipseudometricspace Let 119879 119883 rarr 119862119861(119883) be a set-valued map and define119891 119883 rarr R as 119891(119909) = 119867119904(119909 119879119909) Let 119887 [0infin) rarr [119886 1)119886 gt 0 be a nondecreasing function Let Φ [0infin) rarr [0 1)

be a function such that Φ(119905) lt 119887(119905) and lim sup119903rarr 119905+Φ(119903) lt

lim sup119903rarr 119905+119887(119903) for each 119905 isin [0infin) Moreover assume that

for any 119909 isin 119883 there exists 119910 isin 119879119909 satisfying

119867119904(119910 119879119910) le min Φ (119886) 119886 Φ (119887) 119887 (52)

where 119886 = 119867(119909 119910) and 119887 = 119867minus1(119909 119910) Then 119879 has afixed point

Remark 18 All the results given remain true when we replaceaccordingly the bicomplete quasipseudometric space (119883 119889)by a left Smyth sequentially completeleft 119870-complete or aright Smyth sequentially completeright 119870-complete space

Conflict of Interests

The author declares that there is no conflict of interestsregarding the publication of this paper

References

[1] Y U Gaba ldquoStartpoints and (120572 120574)-contractions in quasi-pseudometric spacesrdquo Journal of Mathematics vol 2014 ArticleID 709253 8 pages 2014

[2] H-P Kunzi and C Ryser ldquoThe Bourbaki quasi-uniformityrdquoTopology Proceedings vol 20 pp 161ndash183 1995

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of

Page 6: Research Article New Results in the Startpoint Theory for ...downloads.hindawi.com/archive/2014/741818.pdf · Research Article New Results in the Startpoint Theory for Quasipseudometric

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Mathematical PhysicsAdvances in

Complex AnalysisJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OptimizationJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Operations ResearchAdvances in

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Algebra

Discrete Dynamics in Nature and Society

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Decision SciencesAdvances in

Discrete MathematicsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of