Open access peer-reviewed chapter - ONLINE FIRST

# Asymptotic Behavior by Krasnoselskii Fixed Point Theorem for Nonlinear Neutral Differential Equations with Variable Delays

Submitted: September 9th 2020Reviewed: January 15th 2021Published: March 15th 2021

DOI: 10.5772/intechopen.96040

## Abstract

In this paper, we consider a neutral differential equation with two variable delays. We construct new conditions guaranteeing the trivial solution of this neutral differential equation is asymptotic stable. The technique of the proof based on the use of Krasnoselskii’s fixed point Theorem. An asymptotic stability theorem with a necessary and sufficient condition is proved. In particular, this paper improves important and interesting works by Jin and Luo. Moreover, as an application, we also exhibit some special cases of the equation, which have been studied extensively in the literature.

### Keywords

• fixed points theory
• stability
• neutral differential equations
• integral equation
• variable delaysAMS Subject Classifications: 34K20
• 34K30
• 34B40

## 1. Introduction

For more than one hundred years, Liapunov’s direct method has been very effectively used to investigate the stability problems of a wide variety of ordinary, functional, and partial differential, integro-differential equations. The success of Liapunov’s direct method depends on finding a suitable Liapunov function or Liapunov functional. Nevertheless, the applications of this method to problems of stability in differential and integro-differential equations with delays have encountered serious difficulties if the delays are unbounded or if the equation has unbounded terms (see [1, 2, 3]). Therefore, new methods and techniques are needed to address those difficulties. Recently, Burton and his co-authors have applied fixed point theory to investigate the stability, which shows that some of these difficulties vanish when applying fixed point theory [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22]. It turns out that the fixed point method is becoming a powerful technique in dealing with stability problems for indeterministic scenes (see for instance [16, 17, 21, 23]).

For example, Luo [16] studied the mean-square asymptotic stability for a class of linear scalar neutral stochastic differential equations by means of Banach’s fixed point theory. The author did not use Lyapunov’s method; he got interesting results for the stability even when the delay is unbounded. The author also obtained necessary and sufficient conditions for the asymptotic stability. Moreover, it possesses the advantage that it can yield the existence, uniqueness, and stability criteria of the considered system in one step.

Neutral delay differential equations are often used to describe the dynamical systems which not only depend on present and past states but also involve derivatives with delays, (see [24, 25, 26, 27, 28]). It has been applied to describe numerous intricate dynamical systems, such as population dynamics [18], mathematical biology [27], heat conduction, and engineering [28], etc.

In particular, qualitative analysis for neutral type equations such as stability and periodicity, oscillation theory, has been an active field of research, both in the deterministic and stochastic cases. We can refer to [6, 7, 13, 15, 16, 17, 19, 20, 21, 23, 29, 30, 31], and the references cited therein.

With this motivation, in this paper, we aim to discuss the boundedness and stability for neutral differential equations with two delays (1). It is worth noting that our research technique is based on Krasnoselskii’s fixed point theory. We will give some new conditions to ensure that the zero solution is asymptotically stable. Namely, a necessary and sufficient condition ensuring the asymptotic stability is proved. Our findings generalize and improve some results that can be found in the literature. In our result, the delays can be unbounded and the coefficients in the equations can change their sign. This paper is organized as follows. In Section 1 we present some basic preliminaries and the form of the neutral functional differential equations which will be studied. In Section 2, we present the inversion of the equation and we state Krasnoselskii’s fixed point theorem. The boundedness and stability of the neutral differential Eq. (1) are discussed in Section 3 via Krasnoselskii’s fixed point theory. Finally, in Section 4 an example is given to illustrate our theory and our method, also to compare our result by using the fixed point theory with the known results by Ardjouni and Djoudi [6].

In this work, we consider the following class of neutral differential equations with variable delays,

xt=atxtτ1t+ctxtτ1t+btxσtτ2t,tt0,E1

denote xtRthe solution to (1) with the initial condition

xt=ψtfortmt0t0,E2

where ψCmt0t0R,σ01is a quotient with odd positive integer denominator. We assume that a,bCR+R,cC1R+Rand τiCR+R+satisfy tτitas t,i=1,2and for each t00,

mit0=inftτittt0,mt0=minmit0i=12.E3

Special cases of Eq. (1) have been recently considered and studied under various conditions and with several methods. Particularly, in the case σ=1/3, and ct=0,in [14] Jin and Luo using the fixed point theorem of Krasnoselskii obtained boundedness and asymptotic stability for the following equation:

x't=atxtτ1t+btx13tτ2t,t0.E4

More precisely, the following result was established.

Theorem A(Jin and Luo [14]).Letτ1be differentiable and suppose that there existsα0.1,k1,k2>0,and a functionhCm0R+such that fort1t21,

t1t2buduk1t1t2,E5

and

t1t2huduk2t1t2,E6

while fort0,

tτ1tthudu+0testhuduhssτ1sshududs+0testhuduhsτ1s1τ1'sas+bsdsα.E7

Then there is a solutionxt0ψof(4) onR+withxt0ψ1.

Notice that when ct=0in the second term on the right-hand side of (1), then (1) reduces to (4). On the other hand, in the case, τ1t=τ1, a constant, Eq. (4) reduces to the one in [9]. Therefore, we considered the more general system than in [9, 14].

Very recently, by the same method of Jin and Luo [14], Ardjouni and Djoudi [6] improved the results of Jin and Luo [14] to the generalized nonlinear neutral differential equation with variable delays of the form

xt=atxtτ1t+ctxtτ1t+btGxσtτ2t,t0,E8

where G:RRis locally Lipschitz continuous in x. That is, there is an L>0so that if x,y1then

GxGy<xyandG0=0.

We note that due to the presence of the term ctxtτ1t, once the equation is inverted then once will face with the term ct1τ1txtτ1t, (where, τ1t1for t0) which produces a restrictive condition for the stability of (8) (as described in more detail below).

Theorem B(Ardjouni and Djoudi [6]).Letτ1be twice differentiable and suppose thatτ1t1for all tm0and suppose that there are constantsα0.1,k1,k2>0,and a functionhCm0R+such that fort1t21,

t1t2buduk1t1t2,E9

and

t1t2huduk2t1t2,E10

while fort0,

ct1τ1t+tτ1tthudu+0testhuduhssτ1sshududs+0testhuduhsτ1s1τsasμs+Lbsdsα,E11

where

μt=ctht+ct1τ1t+ctτ1t1τ1t2.

Then there is a solutionxt0ψof(8) onR+withxt0ψ1.

By letting ct=0and Gxσtτ2t=xσtτ2tin (8), the present authors [14] have studied, the asymptotic stability and the stability by using Krasnoselskii’s fixed point theorem, under appropriate conditions, of the Eq. (4) and improved the results claimed in [9]. Consequently, Theorem B improves and generalizes Theorem A. Following the technique of Jin and Luo [14], Ardjouni and Djoudi [6] studied the stability properties of (8). However, the condition (11) in Ardjouni and Djoudi [6] is restrictive. By employing two auxiliary functions pand gfor constructing a fixed point mapping argument, the alternative condition (21) in Theorem 3.1 is obtained. Note that the condition

ct1τ1't<α,

for some constant α01, is not needed in Theorem 3.1. In the present paper, we also adopt Krasnoselskii’s fixed point theory to study the boundedness and stability of (1). A new criteria for asymptotic stability with a necessary and sufficient condition is given. The considered neutral differential equations, the results and assumptions to be given here are different from those that can be found in the literature and complete that one. These are the contributions of this paper to the literature and its novelty and originality. In addition, an example is provided to illustrate the effectiveness and the merits of the results introduced.

## 2. Inversion of equation

In this section, we use the variation of parameter formula to rewrite the equation as an integral equation suitable for the Krasnoselskii theorem. The technique for constructing a mapping for a fixed point argument comes from an idea in [21]. In our consideration we suppose that:

A1)Let τ1be twice differentiable and suppose that τ1t1for all tmt0.

A2)There exists a bounded function p:mt00with pt=1for tmt0t0such that ptexists for all tmt0.

Let yt=ψton tmt0t0,and let

xt=ptytfortt0.E12

Make substitution of (12) into (1) to show

yt=ptptytatptτ1tctptτ1tptytτ1t+ctptτ1tptytτ1t+btpσtτ2tptyσtτ2t,tt0,E13

then it can be verified that xsatisfies (1).

We now re-write Eq. (13) in an equivalent form. To this end, we use the variation of parameter formula and rewrite the equation in an integral from which we derive a Krasnoselskii fixed point theorem. Besides, the integration by parts will be applied.

We need the following lemma in our proof of the main theorem.

Lemma 2.1.Leth:mt0R+be an arbitrary continuous function and suppose that (A1) and (A2) hold. Thenyis a solution of(13) if and only if

yt=ψt0pt0τ1t0pt0ct01τ1t0ψt0τ1t0t0τ1t0t0hupupuyuduet0thsds+ptτ1tptct1τ1tytτ1t+tτ1tthupupuyudu+t0testhuduμ¯s+hsτ1sp'sτ1spsτ1s1τ1sβ¯s×ysτ1sdst0testhuduhssτ1sshupupuyududs+t0testhudubspσsτ2spsyσsτ2sds,E14

where

μ¯t=atptτ1tctptτ1tpt,Ct=ctptτ1tpt.E15

and

β¯t=Ctht+Ct1τ1t+Ctτ1t1τ1t2.E16

Proof.Let ytbe a solution of Eq. (13). Rewrite (13) as

yt+htyt=htptptytatptτ1tctptτ1tptytτ1t+ctptτ1tptytτ1t+btpσtτ2tptyσtτ2t,tt0.E17

Multiply both sides of (17) the previous equality by et0thsdsand then integrate from t0to t,we have

yt=ψt0et0thsds+t0thspspsesthuduysdst0testhuduaspsτ1scspsτ1spsysτ1sds+t0testhuducspsτ1spsysτ1sds+t0testhudubspσsτ2spsyσsτ2sds.E18

Performing an integration by parts, we can conclude, for tt0,

yt=ψt0et0thsds+t0testhududsτ1sshupupuyudu+t0testhuduhsτ1spsτ1spsτ1s×1τ1sysτ1sdst0testhuduaspsτ1scspsτ1spsysτ1sds+t0testhuducspsτ1spsysτ1sds+t0testhudubspσsτ2spsyσsτ2sds.

Thus,

yt=ψt0pt0τ1t0pt0ct01τ1t0ψt0τ1t0t0τ1t0t0hupupuyuduet0thsds+ptτ1tptct1τ1tytτ1t+tτ1tthupupuyudu+t0testhuduμ¯s+hsτ1spsτ1spsτ1s1τ1sβ¯s×ysτ1sdst0testhuduhssτ1sshupupuyududs+t0testhudubspσsτ2spsyσsτ2sds,

where μ¯sand β¯sare defined in (15) and (16), respectively. The proof is complete.

Below we state Krasnoselskii’s fixed point theorem which will enable us to establish a stability result of the trivial solution of (1) For more details on Krasnoselskii’s captivating theorem, we refer to smart [20] or [3].

Theorem 2.1.(see, [Kranoselskii’s fixed point theorem, [20]]). Suppose thatX.is a Banach space andMis a bounded, convex, and closed subset ofX. Suppose further that there exist, two operators, A,BMintoXsuch that:

1. Ax+ByMfor allx,yM;

2. Ais completely continuous;

3. Bis a contraction mapping.

ThenA+Bhas a fixed point inM.

## 3. Stability by Krasnoselskii fixed point theorem

From the existence theory, which can be found in Hale [26] or Burton [3], we conclude that for each t0ψR+×Cmt0t0R, a solution of (1) through t0ψis a continuous function x:mt0t0+ρRfor some positive constant ρ>0such that xsatisfies (1) on t0t0+ρand x=ψon mt0t0. We denote such a solution by xt=xtt0ψ. We define ψ=maxψt:mt0tt0.

As we mentioned previously, our results in this section extend and improve the work in [14] by considering more general classes of neutral differential equations presented by (1). Our main results in this part can be applied to the case when

ct1τ1t1,

which improve [14]. In other words, we will establish and prove a necessary and sufficient condition ensuring the boundedness of solutions and the asymptotic stability of the zero solution to Eq. (1). However, the mathematical analysis used in this research to construct the mapping to employ Krasnoselskii’s fixed point theorem is different from that of [14].

The results of this work are news and they extend and improve previously known results. To the best of our knowledge from the literature, there are few authors who have used the fixed point theorem to prove the existence of a solution and the stability of trivial equilibrium of several special cases of (1) all at once [9, 14].

Let us know to recall the definitions of stability that will be used in the next section. For stability definitions, we refer to [3].

Definition 3.1.The zero solution of (1) is said to be:

1. stable, if for any ε>0and t00,there exists a δ=δεt0>0such that ψCmt0t0Rand ψ<δimply xtt0ψ<εfor tt0.

2. asymptotically stable, if the zero solution is stable and for any ε>0and t00,there exists a δ=δεt0>0such that ψCmt0t0Rand ψ<δimply xtt0ψ0as t..

Now, we can state our main result.

Theorem 3.1. Suppose that assumptions (A1) and (A2) hold, and that there are constantsα01,k1,k2>0,and an arbitrary continuous functionhCmt0R+such that fort1t21,

t1t2bupσuτ2upuduk1t1t2,E19

and

t1t2huduk2t1t2,E20

while fortt0

ptτ1tptct1τ1t+tτ1tthupupudu+t0testhuduμ¯s+hsτ1spsτ1spsτ1s1τ1sβ¯sds+t0testhuduhssτ1sshupupududs+t0testhudubspσsτ2spsds<α,E21

whereμ¯sandβ¯sare defined in(15) and(16), respectively. Ifψis a given continuous initial function which is sufficiently small, then there is a solutionxtt0ψof(1) onR+withxtt0ψ1.

We are now ready to prove Theorem 3.1.

Let X.gbe the Banach space of continuous φ:mt0Rwith

φgsuptmt0φt/gt<.

For each t00and ψCmt0t0Rfixed, we define Xψas the following space

Xψ=φX:φt1fortmt0andφt=ψtiftmt0t0.

It is easy to check that Xψis a complete metric space with metric induced by the norm .g.

We note that to apply Krasnoselskii’s fixed point theorem we need to construct two mappings; one is contraction and the other is compact. Therefore, we use (14) to define the operator H:XψXψby

tAφt+Bφt,

where A,B:XψXψare given by

Aφtt0testhudubspσsτ2spsφσsτ2sds,E22

and

Bφt:=ψt0pt0τ1t0pt0ct01τ1t0ψt0τ1t0t0τ1t0t0hupupuφuduet0thsds+ptτ1tptct1τ1tφtτ1t+tτ1tthupupuφudu+t0testhuduμ¯s+hsτ1spsτ1spsτ1s1τ1sβ¯s×φsτ1sdst0testhuduhssτ1sshupupuφududs.E23

If we are able to prove that Hpossesses a fixed point φon the set Xψ,then ytt0ψ=φtfor tt0,ytt0ψ=ψton mt0t0,ytt0ψsatisfies (13) when its derivative exists and ytt0ψ<1for tt0.That Amaps Xψinto itself can be deduced from condition (21).

For α01, we choose δ>0such that

1+pt0τ1t0pt0ct01τ1t0+t0τ1t0t0hupupuduet0thsdsδ+α1.E24

Let ψ:mt0t0Rbe a given continuous initial function with ψ<δ.Let g:mt01be any strictly increasing and continuous function with gmt0=1,gsas s,such that

ptτ1tptct1τ1t+tτ1tthupupugu/gtdu+t0testhuduhssτ1sshupupugu/gtduds+t0testhuduμ¯s+hsτ1spsτ1spsτ1s1τ1sβ¯s×gsτ1s/gtds<α.E25

Now we split the rest of our proof into three steps.

First step:We now show that φ,ϕXψimplies that Aφ+BϕXψ.Now, let .be the supremum norm on mt0of φXψif φis bounded. Note that if φ,ϕXψthen

Aφt+Bϕt
ψ1+pt0τ1t0pt0ct01τ1t0+t0τ1t0t0hupupuduet0thsds+ϕptτ1tptct1τ1t+ϕtτ1tthupupudu+ϕt0testhuduμ¯s+hsτ1spsτ1spsτ1s1τ1sβ¯sds+ϕt0testhuduhssτ1sshupupududs+φσt0testhudubspσsτ2spsds1+pt0τ1t0pt0ct01τ1t0+t0τ1t0t0hupupuduet0thsdsδ+α1.

By applying (24), we see that Aφt+Bϕt1for tmt0.

We see that also Bmaps Xψinto itself by letting φ=0in the preceding sum.

Second step:Next, we will show that AXψis equicontinuous and Ais continuous. We first show that AXψis equicontinuous. If φXψand if 0t1<t2with t2t1<1,then

Aφt2Aφt1=t0t2est2hudubspσsτ2spsdst0t1bspσsτ2spsest1hududst1t2bspσsτ2spsest2hududs+t0t1est2huduest1hudubspσsτ2spsdst1t2est2hududt1sbυpσυτ2υpυ+et1t2hudu1t0t1est1hudubspσsτ2spsdst1t2bupσuτ2upudu1+t1t2huest2hududs+αet1t2hudu1
2t1t2bupσuτ2upudu+αt1t2hudu2k1+αk2t2t1,

by (19)–(21). In the space X.g, the set AXψis uniformly bounded and equicontinuous. Hence by Ascoli-Arzela theorem AXψresides in a compact set.

Next, we need to show that Ais continuous. Let ε>0be given and let φ,ϕXψ. Now yσ,is uniformly continuous on 1+1so for a fixed T>0with 4/gT<εthere is an η>0such that y1y2<ηgTimplies y1σy2σ<ε/2.Thus for φtϕt<ηgtand for t>Twe have

AφtAϕt/gt=1/gtt0testhudubspσsτ2spsφσsτ2sϕσsτ2sds1/gtt0Testhudubspσsτ2spsφσsτ2sϕσsτ2sds+2Ttesthudubspσsτ2spsdsαε/2gt+2α/gTαε.

Third step:Finally, we show that Bis a contraction with respect to the norm .gwith constant α.Let Bbe defined by (23). Then for ϕ1,ϕ2Xψwe have

Bϕ1tBϕ2t/gtptτ1tptct1τ1tϕ1tτ1tϕ2tτ1t/gt+tτ1tthupupuϕ1uϕ2u/gtdu+t0testhuduμ¯s+hsτ1spsτ1spsτ1s1τ1sβ¯s×ϕ1sτ1sϕ2sτ1s/gtds+t0testhuduhssτ1sshupupuϕ1uϕ2u/gtdudsϕ1ϕ2gptτ1tptct1τ1t+tτ1tthupupugu/gtdu+t0testhuduhssτ1sshupupugu/gtduds+t0testhudu×μ¯s+hsτ1spsτ1spsτ1s1τ1sβ¯sgsτ1s/gtds

αϕ1ϕ2g,by (22).

Since α01,we can conclude that Bis a contraction on Xψ.g.

The conditions of Krasnoselskii’s theorem are satisfied with M=Xψ. Hence, we deduce that H:XψXψhas a fixed point yt, which is a solution of (13) with ys=ψson smt0,t0]and ytt0ψ1for tmt0. Since there exists a bounded function p:mt00with pt=1for tmt0t0, by hypotheses (12) and from the above arguments we deduce that there exists a solution xof (1) with x=ψon mt0t0satisfies xtt0ψ1for all tmt0. The proof is complete.

Letting σ=1/3,and ct=0in Theorem 3.1. Then we have the following corollary.

Corollary 3.1.Let(19) and(20) hold, and(21) be replaced by

tτ1tthupupudu+t0testhuduhsτ1spsτ1spsτ1s1τ1saspsτspsds+t0testhuduhssτ1sshupupududsα.E26

Then there is a solutionxtt0ψof(4) onR+withxtt0ψ1.

Remark 3.2:When pt=1,then Corollary 3.1 reduces to Theorem A, which was recently stated in Jin and Luo [14]. Therefore, the paper (Jin and Luo [14]) is a particular case of ours.

For the next Theorem, we manipulate function spaces defined on infinite t-intervals. So, for compactness, we need an extension of the Arzelà-Ascoli theorem. This extension is taken from ([3], Theorem 1.2.2 p. 20).

Theorem 3.2. Let(19)–(21) hold and assume that

t0testhudubspσsτ2spsds0ast,E27

and

limtinft0thsds>.E28

Ifψis given continuous initial function which is sufficiently small, then(1) has a solutionxtt0ψ0astif and only if

t0thsdsast.E29

Proof.We set

K=suptt0et0thsds,E30

by (28), Kis well defined. Suppose that (29) holds.

Since pis bounded, it remains to prove that the zero solution of (1) is asymptotically stable.

All of the calculations in the proof of Theorem 3.1 hold with gt=1when .gis replaced by the supremum norm ..

For

φXψ,
Aφtt0testhudubspσsτ2spsdsqt,E31

where qt0as tby (27).

Add to Xψthe condition that φXψimplies that φt0as t. We can see that for φXψthen Aφt0as tby (31), and Bφt0as tby (29).

Since AXψhas been shown to be equicontinuous, Amaps Xψinto a compact subset of Xψ. By Krasnoselskii’s theorem, there is yXψwith Ay+By=y. As yXψ,ytt0ψ0as t.By condition (12), it is very easy to show that there exists a solution xXψof (1) with xtt0ψ0as t.

Conversely, we suppose that (29) fails. From (28) there exists a sequence tnwith tnas nsuch that limnt0tnhudu=ξfor some ξR+.We may also choose a positive constant Jsatisfying

Jt0tnhudu+J,

for all n1.To simplify the expression, we define

ωsμ¯s+hsτ1spsτ1spsτ1s1τ1sβ¯s+hssτ1sshupupudu+bspσsτ2sps,

for all s0.By (21), we have

t0tnestnhuduωsdsα.

This yields

t0tne0shuduωsdsαe0tnhudueJ.

The sequence t0tne0shuduωsdsis bounded, hence there exists a convergent subsequence. Without loss of generality, we can assume that

limnt0tne0shuduωsds=θ,

for some θR+.Let mbe an integer such that

tmtne0shuduωsdsδ04K,

for all nm, where δ0>0satisfies 2δ0KeJ+α1.

We now consider the solution yt=yttmψof (1) with ψtm=δ0and ψsδ0for stm.We may choose ψso that yt1for ttmand

ψtmptmτ1tmptmctm1τ1'tmψtmτ1tmtmτ1tmtmhup'upuzudu12δ0.

In follows from (22) and (23) with yt=Ayt+Bytthat for nm

ytnptnτ1tnptnctn1τ1tnytnτ1tntnτ1tntnhspspsysds12δ0etmtnhudutmtnestnhuduωsds=etmtnhudu12δ0e0tmhudutmtne0shuduωsdsetmtnhudu12δ0Ktmtne0shuduωsds14δ0etmtnhudu14δ0e2J>0.E32

On the other hand, if the zero solution of (13) yt=yttmψ0as t,since tnτitnas t,i=1,2,and (21) holds, we have

ytnptnτ1tnptnctn1τ1'tnytnτ1tntnτ1tntnhsp'spsysds0

as t,which contradicts (32). Hence condition (29) is necessary for the asymptotic stability of the zero solution of (13), and hence the zero solution of (1) is asymptotically stable. The proof is complete.

For the special case ct=0and σ=13, we can get.

Corollary 3.2.Let(19), (20) and(27) hold and(21) be replaced by

tτ1tthupupudu+t0testhuduhsτ1spsτ1spsτ1s1τ1saspsτspsds+t0testhuduhssτ1sshupupududsα.

Then the zero solutionxtt0ψof(4) with a small continuous functionψtis asymptotically stable if only if

t0thsdsast.

Remark 3.3.The method in this paper can be applied to more general nonlinear neutral differential equations than Eq. (1).

Remark 3.4.Theorem 3.1 is still true if condition (21) is satisfied for ttρwith some tρR+.

## 4. Example

In this section, we now give an example to show the applicability of Theorem 3.1.

Example.Let us consider the following neutral differential equation of first order with two variable delays, which is a special case of (1):

xt=atxtτ1t+ln0.95t+14t+1xtτ1t+0.60.95t+113t+12x13tτ2t,E33

for t0where τ2t=0.5t,τ1t=0.05t,and atsatisfies

μ¯t+htτ1tptτ1tptτ1t1τ1tβ¯t0.03t+1,

where μ¯tand β¯tare defined in (15) and (16), respectively. Choosing ht=1.5t+1and pt=1t+1. By straightforward computations, we can check that condition (21) in Theorem 3.1 holds true. As t,we have

ptτ1tptct1τ1t14×0.950.263,tτ1tthupupudu0.026,0testhuduhssτ1sshupupududs0.026,0testhuduμ¯s+hsτ1spsτ1spsτ1s1τ1sβ¯sds0test1.5u+1du0.3s+1ds0.2,

and

0testhudubspσsτ2spsds0.4,and since 0thsdsas t,pt1.Let α=0.263+0.026+0.026+0.2+0.4. It is easy to see that all the conditions of Theorem 3.1 hold for α0.915<1.Thus, Theorem 3.1 implies that the zero solution of (33) is asymptotic stable.

However, for the asymptotic stable of the zero solution of (33), the corresponding conditions used by the fixed point theory in Ardjouni and Djoudi [6] are

limct1τ1't=lim10.95ln0.95t+14t+1=1.513ast.

This implies that condition (11) does not hold. So it is clear that the reduction of the conservatism by our method is quite significant when compared to Ardjouni and Djoudi [6].

Remark 4.1. It is an open problem whether the zero solution of (1) is uniform asymptotically stable, perseverance, and so on.

## 5. Conclusion

This work is a new attempt at applying the fixed point theory to the stability analysis of neutral differential equations with variable delays, several special cases of which have been studied in [9, 14]. Some of the results, like Theorem B, is mainly dependent on the constraint

ct1τ1't<1.

But in many environments, the constraint does not hold. So by employing two auxiliary continuous functions gand pto define an appropriate mapping, and present new criteria for asymptotic stability of Eq. (1) which makes stability conditions more feasible and the results in [14] are improved and generalized. From what has been discussed above, we see that Krasnoselskii’s fixed point theorem is effective for not only the investigation of the existence of solution but also for the boundedness and the stability analysis of trivial equilibrium. We introduce an example to verify the applicability of the results established. In the future, we will continue to explore the application of other kinds of fixed point theorems to the stability research of fractional neutral systems with variable delays.

chapter PDF

## More

© 2021 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution 3.0 License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

## How to cite and reference

### Cite this chapter Copy to clipboard

Benhadri Mimia (March 15th 2021). Asymptotic Behavior by Krasnoselskii Fixed Point Theorem for Nonlinear Neutral Differential Equations with Variable Delays [Online First], IntechOpen, DOI: 10.5772/intechopen.96040. Available from: