AAA Abstract and Applied Analysis 1687-0409 1085-3375 Hindawi Publishing Corporation 175479 10.1155/2013/175479 175479 Research Article On the Existence and Stability of Periodic Solutions for a Nonlinear Neutral Functional Differential Equation Yuan Yueding 1, 2, 3 0000-0002-8351-8571 Guo Zhiming 1, 3 Huang Chuangxia 1 School of Mathematics and Information Sciences Guangzhou University Guangzhou 510006 China gzhu.edu.cn 2 School of Mathematics and Computer Sciences Yichun University Yichun 336000 China yichununiversity.com/ 3 Key Laboratory of Mathematics and Interdisciplinary Science of Guangdong Higher Education Institutes Guangzhou University Guangzhou 510006 China gzhu.edu.cn 2013 31 03 2013 2013 31 01 2013 01 03 2013 2013 Copyright © 2013 Yueding Yuan and Zhiming Guo. This 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.

This paper deals with the existence and stability of periodic solutions for the following nonlinear neutral functional differential equation (d/dt)Dut=p(t)-au(t)-aqu(t-r)-h(u(t),u(t-r)). By using Schauder-fixed-point theorem and Krasnoselskii-fixed-point theorem, some sufficient conditions are obtained for the existence of asymptotic periodic solutions. Main results in this paper extend the related results due to Ding (2010) and Lopes (1976).

1. Introduction

In recent years, the existence and stability of periodic solutions for differential equation has been extensively studied. Many researchers used the Lyapunov functional method, Hopf bifurcation techniques, and Mawhin continuation theorems to obtain the existence and stability of periodic solutions for neutral functional differential equation (NFDE); see papers  and their references therein. However, researches on the existence and stability of periodic solutions for NFDE by using fixed-point theorem are relatively rare [15, 16]. The reason lies in the fact that it is difficult to construct an appropriate completely continuous operator and an appropriate bounded closed convex set.

In this paper, we will investigate the existence and stability of periodic solutions for the following nonlinear NFDE (1)ddtDut=p(t)-au(t)-aqu(t-r)-h(u(t),u(t-r)), where Dut=u(t)-qu(t-r),|q|1,a  0,hC(×,), and pC(,). Such a kind of NFDE has been used for the study of distributed networks containing a transmission line [17, 18]. For example, suppose a system consists of a long electrical cable of length l, and one end of which isconnected to a power source E(t) with resistance R0, while the other end is connected to an oscillating circuit formed of a condenser C0 and a nonlinear element, the volt-ampere characteristic of which is i=f(u). Let L,C,R,and  G be the parameters of the transmission line, respectively, Z0 the characteristic impedance of the line, v=1/LC the propagation velocity and assume the losses can not be disregarded. The process of the final end volt u(t) in such a system can be described by the following NFDES: (2)u(t)-qu(t-r)=p(t)-au(t)-aqu(t-r)-bf(u(t))+bqf(u(t-r)) or (3)C0(u(t)-ku(t-r))=p(t)-Zu(t)-Zku(t-r)-f(u(t)-ku(t-r)), where a=1/Z0C0, b=1/C0, q=(Z0-R0)/A2  (Z0+R0), k=(1-ZR0)/(1+ZR0), p(t)=2E(t-(r/2))/A(Z0+R0)C0, r=2l/v, A=eRl/Z0, Z=C/L, and f(u) is a given nonlinear function. If R0>0, then |q|<1, |k|<1. Obviously, we see that (2) (or (3)) is a special case of (1). The aim of this paper is to establish some criteria to guarantee the existence and stability of periodic solution for (1) by using Schauder's fixed-point theorem and Krasnoselskii's fixed-point theorem. The interesting is that main results obtained in this paper extend the related existing results. Furthermore, our results can also be applied to solve the problem of the existence and stability of periodic solutions for (2) and (3).

2. Main Results and Proofs

In this section, let C1(N) (C(N)) denote the set of all continuously differentiable functions (all continuous functions) ϕ:N, where N=1,2.Cω={ϕϕC(),ϕ(t+ω)=ϕ(t)} is a Banach space with the supremum norm ·0, Cω1=C1()Cω with the norm ϕ1=ϕ0+ϕ0 in a period interval, and ω is a positive constant. The next lemma will be used in the sequel.

Lemma 1.

If a0, fCω, then the scalar equation x(t)=ax(t)+f(t) has a unique ω-periodic solution: (4)x(t)=(1-eaω)-1tt+ωea(t+ω-s)f(s)ds.

Proof.

It is easy to prove. We can find it in many ODE textbooks (e.g., see Example  2 on page 35 of ).

Theorem 2.

Suppose that hC(2) and pCT. If there exists a constant H>0 such that (5)p0<(1-3|q|)aH-sup|x|,|y|H|h(x,y)|, then (1) has a T-periodic solution.

Proof.

According to the condition (5), we can find a sufficiently small L>0 such that (6)(2L+1a)[p0+2a|q|H+sup|x|,|y|H|h(x,y)|]+|q|HH. Let v(t)=u(Lt), τ=r/L, p1(t)=p(Lt), and ω=T/L; then (1) can be rewritten as (7)v(t)-qv(t-τ)=Lp1(t)-aLv(t)-aqLv(t-τ)-Lh(v(t),v(t-τ)), where p1(t)Cω with p0=p10. It suffices to prove that (7) has a ω-periodic solution. Let (8)M={ϕϕCω1,ϕ1H}. Then M is a bounded closed convex set of the Banach space C1(). For any given ϕM, consider the nonhomogeneous equation: (9)v(t)=-aLv(t)+Lp1(t)-aqLϕ(t-τ)-Lh(ϕ(t),ϕ(t-τ))+qϕ(t-τ).

According to Lemma 1, (9) has a unique ω-periodic solution: (10)v(t)=(1-e-aLω)-1×tt+ωe-aL(t+ω-s)×[Lh(ϕ(s),ϕ(s-τ))Lp1(s)-aqLϕ(s-τ)-Lh(ϕ(s),ϕ(s-τ))+qϕ(s-τ)]ds.

Define an operator A by (11)(Aϕ)(t)=(1-e-aLω)-1×tt+ωe-aL(t+ω-s)×[-Lh(ϕ(s),ϕ(s-τ))+qϕ(s-τ)Lp1(s)-aqLϕ(s-τ)-Lh(ϕ(s),ϕ(s-τ))+qϕ(s-τ)]ds=(1-e-aLω)-1×tt+ωe-aL(t+ω-s)×[-Lh(ϕ(s),ϕ(s-τ))]Lp1(s)-2aqLϕ(s-τ)-Lh(ϕ(s),ϕ(s-τ))]ds+qϕ(t-τ).

In order to prove that (7) has a periodic solution, we shall make sure that A satisfies the conditions of Schauder's fixed-point theorem (see Lemma  2.2.4 on page 40 of ).

Note that for any xM, we have x(t+ω)=x(t) and x1H(12)(Ax)(t+ω)=(1-e-aLω)-1×t+ωt+2ωe-aL(t+2ω-s)×[Lh(x(s),x(s-τ))Lp1(s)-2aqLx(s-τ)-Lh(x(s),x(s-τ))Lp1(s)-2aqLx(s-τ)]ds+qx(t+ω-τ)=(1-e-aLω)-1×tt+ωe-aL(t+ω-s)×[(x(s),x(s-τ))Lp1(s)-2aqLx(s-τ)-Lh(x(s),x(s-τ))Lp1(s)-2aqLx(s-τ)]ds+qx(t-τ)=(Ax)(t); Therefore, (Ax)(t+ω)=(Ax)(t). Meanwhile, we get (13)(Ax)(t)=(1-e-aLω)-1×{tt+ωe-aL(t+ω-s)(-aL)×[Lp1(s)-2aqLx(s-τ)-Lh(x(s),x(s-τ))Lp1(s)-2aqLx(s-τ)Lp1(s)-2aqLx(s-τ)]ds+(1-e-aLω)[Lp1(t)  -2aqLx(t-τ)-Lh(x(t),x(t-τ))Lp1(t)  -2aqLx(t-τ)]tt+ωe-aL(t+ω-s)}+qx(t-τ)=(1-e-aLω)-1(-aL)×tt+ωe-aL(t+ω-s)×[Lp1(s)-2aqLx(s-τ)-Lh(x(s),x(s-τ))Lp1(s)-2aqLx(s-τ)]ds+[Lp1(t)-2aqLx(t-τ)-Lh(x(t),x(t-τ))]+qx(t-τ).

By (6), we have (14)Ax1=Ax0+(Ax)0supt|[Lp1(t)-2aqLx(t-τ)-Lh(x(t),x(t-τ))](1-e-aLω)-1×tt+ωe-aL(t+ω-s)×[-Lh(x(s),x(s-τ))Lp1(s)-2aqLx(s-τ)(1-e-aLω)-1tt+ω-Lh(x(s),x(s-τ))[-Lh(x(s),x(s-τ))Lp1(s)-2aqLx(s-τ)]ds|tt+ω+supt|(1-e-aLω)-1(-aL)×tt+ωe-aL(t+ω-s)×[Lh(x(s),x(s-τ))Lp1(s)-2aqLx(s-τ)(1-e-aLω)-1tt+ω-Lh(x(s),x(s-τ))]ds+[Lp1(t)-2aqLx(t-τ)-Lh(x(t),x(t-τ))]×tt+ωe-aL(t+ω-s)|+|q|x1(2+1aL)×[Lp10+2a|q|LH+Lsup|x|,|y|H|h(x,y)|]+|q|H=(2L+1a)×[p10+2a|q|H+sup|x|,|y|H|h(x,y)|]+|q|HH. Thus, AxM.

For any xM, Ax0H, (Ax)0H. According to Arzela-Ascoli Theorem (see Theorem 4.9.6 on page 84 of ), the subset AM of Cω is a precompact set; therefore, A:MC1()Cω is a compact operator.

Suppose that {xn}M, xnx, then xn-x00 and xn-x00 as n. Also, we have (15)Axn-Ax0=supt|(1-e-aLω)-1×tt+ωe-aL(t+ω-s)×[(h(xn(s),xn(s-τ))-h(x(s),x(s-τ)))-2aqL(xn(s-τ)-x(s-τ))-L(h(xn(s),xn(s-τ))-h(x(s),x(s-τ)))]ds+q(xn(t-τ)-x(t-τ))(1-e-aLω)-1tt+ω|1aL[+Lsupt[0,ω]|h(xn(t),xn(t-τ))-h(x(t),x(t-τ)xn(t),xn(t-τ))|2a|q|Lxn-x0+Lsupt[0,ω]|h(xn(t),xn(t-τ))-h(x(t),x(t-τ)xn(t),xn(t-τ))|+Lsupt[0,ω]|h(xn(t),xn(t-τ))]+|q|xn-x0=3|q|xn-x0+1asupt[0,ω]×|h(xn(t),xn(t-τ))-h(x(t),x(t-τ))|,(Axn)-(Ax)0=supt|(xn(t-τ)-x(t-τ))(1-e-aLω)-1(-aL)×tt+ωe-aL(t+ω-s)×[(h(xn(s),xn(s-τ))-h(x(s),x(s-τ)))-2aqL×(xn(s-τ)-x(s-τ))-L(h(xn(s),xn(s-τ))-h(x(s),x(s-τ))ef22xn(s-τ))]ds-2aqL(xn(t-τ)-x(t-τ))-L(h(xn(t),xn(t-τ))-h(x(t),x(t-τ))h(xn(t),xn(t-τ)))+q(xn(t-τ)-x(t-τ))tt+ω|2L[supt[0,ω]|h(xn(t),xn(t-τ))2a|q|xn-x0+supt[0,ω]|h(xn(t),xn(t-τ))-h(x(t),x(t-τ))h(xn(t),xn(t-τ))|2a|q|xn-x0+supt[0,ω]|h(xn(t),xn(t-τ))]+|q|xn-x0.

When xn-x10 as n, |xn(t)-x(t)|0 for t[0,ω] uniformly. And since h is continuous, Axn-Ax00, (Axn)-(Ax)00. Consequently, A is continuous.

Thus, by Schauder-fixed-point theorem (see Lemma  2.2.4 on page 40 of ), there is a ϕM such that ϕ=Aϕ. Therefore, (7) has at least one ω-periodic solution. Since v(t)=u(Lt) and p(Lt)=p1(t), (1) has at least one T-periodic solution. The proof is completed.

Next, we explore the stability of this T-periodic solution u*(t) for (1). We assume that theconditions of Theorem 2 are satisfied. Therefore, (1) has at least one T-periodic solution u*(t). Let v(t)=u(t)-u*(t) then (1) is transformed as (16)ddtDvt=-av(t)-aqv(t-r)-g(v(t),v(t-r)), where Dvt=v(t)-qv(t-r) and g(v(t),v(t-r))=h(u*(t)+v(t),u*(t-r)+v(t-r))-h(u*(t),u*(t-r)). Clearly, (16) has trivial solution v(t)0. Now we use Krasnoselskiis-fixed-point theorem (see  or [15, Lemma  2.2]) to prove that trivial solution v(t)0 to (16) is asymptotically stable.

Set S as the Banach space of bounded continuous function ϕ:[-r,)R with the supremum norm ·. Also, Given the initial function ψ, denote the norm of ψ by ψ=supt[-r,0]|ψ(t)|, which should not cause confusion with the same symbol for the norm in S.

Theorem 3.

Let H be as in Theorem 2. Assume that all conditions of Theorem 2 are satisfied. Suppose that h satisfies the Lipschitz condition and (17)sup|x|,|y|H|h(x,y)|aH(1+|q|). If there exists Q>0 such that (18)ψQ-3|q|Q1+|q|-H-1a(1+|q|)sup|x|,|y|H+Q|h(x,y)|, then the solution v(ψ)(t) to (16) through ψ satisfies limtv(ψ)(t)=0.

Proof.

By (18), we have (19)(1+|q|)ψ+3|q|Q+1asup|x|,|y|H+Q|h(x,y)|+H(1+|q|)Q.

Given the initial function ψ, by [20, Theorem  12.2.3], there exists a unique solution v(ψ)(t) for (16). Let (20)Mψ={ϕϕS,ϕQ,ϕ0=ψ,ϕ(t)0ast}; then Mψ is a bounded convex closed set of S. We write (16) as (21)[v(t)-qv(t-r)]=-a[v(t)-qv(t-r)]-2aqv(t-r)-g(v(t),v(t-r)); then we have (22)v(t)=[ψ(0)-qψ(-r)]e-at+qv(t-r)+0t[-2aqv(s-r)-g(v(s),v(s-r))]e-a(t-s)ds.

For all ϕMψ, define the operators A and B by (23)(Aϕ)(t)={0,t[-r,0],0t[-2aqϕ(s-r)-g(ϕ(s),ϕ(s-r))]e-a(t-s)ds,0t[-2aqϕ(s-r)-g(ϕ(s),ϕ(s-r))]t0,(Bϕ)(t)={ψ(t),t[-r,0],(ψ(0)-qψ(-r))e-at+qϕ(t-r),t0.

For any x,yMψ, x(t)0, y(t)0 as t, and xQ, yQ. Therefore, we have (24)limt(Ax)(t)=limt0t[-2aqx(s-r)-g(x(s),x(s-r))]easdseat=limt1a[-2aqx(t-r)-g(x(t),x(t-r))]=0,limt(By)(t)=limt[(ψ(0)-qψ(-r))e-at+qy(t-r)]=0. Thus, limt(Ax+By)(t)=0. Again by (17) and (19), we have (25)Ax+ByAx+By=supt0|0t[-2aqx(s-r)-g(x(s),x(s-r))]e-a(t-s)ds|+supt-r|(By)(t)|[2a|q|Q+sup|x|,|y|H+Q|h(x,y)|+sup|x|,|y|H|h(x,y)|]×supt0|0te-a(t-s)ds|+max{ψ,supt0|(ψ(0)-qψ(-r))e-at+qy(t-r)|}(1+|q|)ψ+3|q|Q+1asup|x|,|y|H+Q|h(x,y)|+1asup|x|,|y|H|h(x,y)|(1+|q|)ψ+3|q|Q+1asup|x|,|y|H+Q|h(x,y)|+H(1+|q|)Q. Thus, Ax+ByMψ.

Since |(Ax)(t)|=0,t[-r,0], and (26)|(Ax)(t)|=|-a0t[-2aqx(s-r)-g(x(s),x(s-r))]e-a(t-s)ds-2aqx(t-r)-g(x(t),x(t-r))0t|a[2a|q|Q+sup|x|,|y|H+Q|h(x,y)|+sup|x|,|y|H|h(x,y)|]×supt0|0te-a(t-s)ds|+2a|q|Q+sup|x|,|y|H+Q  |h(x,y)|+sup|x|,|y|H|h(x,y)|=2[2a|q|Q+sup|x|,|y|H+Q|h(x,y)|+sup|x|,|y|H|h(x,y)|],t0; here, the derivative of (Ax)(t) at zero means the left-hand side derivative when t0 and the right-hand side derivative when t0, one can see (Ax)(t) is bounded for all xMψ. Therefore, AMψ is a precompact set of S. Thus, A is compact.

Suppose that {xn}Mψ, xS, xnx as n; then |xn(t)-x(t)|0 uniformly for t-r as n. Since (27)Axn-Ax=supt0|0t{-2aq[xn(s-r)-x(s-r)]-g(xn(s),xn(s-r))+g(x(s),x(s-r))[xn(s-r)-x(s-r)]}e-a(t-s)ds0t|[|h(u*(t)+xn(t),u*(t-r)+xn(t-r))2a|q|xn-x+supt0|h(u*(t)+xn(t),u*(t-r)+xn(t-r))-h(u*(t)+x(t),u*(t-r)+x(t-r))|2a|q|xn-x|h(u*(t)+xn(t),u*(t-r)+xn(t-r))]×supt0|0te-a(t-s)ds|=2|q|xn-x+1asupt0|h(u*(t)+xn(t),u*(t-r)+xn(t-r))-h(u*(t)+x(t),u*(t-r)+x(t-r))|, and h is continuous, we have Axn-Ax0 as n. Thus, A is continuous. Due to the fact that (28)Bx-By=supt0|qx(t-r)-qy(t-r)||q|x-y,x,yMψ, and |q|<1, we know that B is a contractive operator.

According to Krasnoselskii's fixed-point theorem (see  or [15, Lemma  2.2]), there is a ϕMψ such that (A+B)ϕ=ϕ. Therefore, ϕ(t) is a solution for (16). Because the solution through ψ for the equation is unique, the solution v(ψ)(t)=ϕ(t)0 as t.

When h satisfies the Lipschitz condition, then there is a constant L>0 such that (29)|h(u(t)+u*(t),u(t-r)+u*(t-r))-h(u*(t),u*(t-r))|L|u(t)|2+|u(t-r)|2,uS. Since ϕ satisfies (30)ϕ(t)=[ψ(0)-qψ(-r)]e-at+qϕ(t-r)+0t[-2aqϕ(s-r)-g(ϕ(s),ϕ(s-r))]e-a(t-s)ds, then (31)ϕ(1+|q|)ψ+3|q|ϕ+2Laϕ, that is (32)(1-3|q|-2La)ϕ(1+|q|)ψ. Therefore, if 1-3|q|-(2L/a)>0, then there clearly exists a δ>0 for any ε>0 such that |ϕ(t)|<ε for all t-r if ψ<δ. Thus, we have the following.

Theorem 4.

If the Lipschitz constant L for h corresponding to 2 satisfies (33)1-3|q|-2La>0, then the zero solution for (16) is stable.

When p is constant and the equation p-a(1+q)u=h(u,u) has only one solution u*, then u* is an equilibrium of (1) and (1) can be transformed to the following equation: (34)ddtDut=-au(t)-aqu(t-r)-g(u(t),u(t-r)), where Dut=u(t)-qu(t-r) and g(u(t),u(t-r))=h(u*+u(t),u*+u(t-r))-h(u*,u*). Now, we consider the stability of the zero solution for (34).

Theorem 5.

Suppose that is hC1(2) and (hx(u*,u*),  hy(u*,u*))=(0,0); then the zero solution of (34) is exponentially asymptotically stable.

Proof.

For all ϕ in C=C([-r,0],), let (35)Dϕ=ϕ(0)-qϕ(-r),Lϕ=-aϕ(0)-aqϕ(-r),Fϕ=-g(ϕ(0),ϕ(-r)). Then D is stable, and D and L are linear and continuous. Consider the equation (d/dt)Dut=Lut. Let (36)V(ϕ)=(Dϕ)2+2aq2-r0ϕ2(θ)dθ. Then (37)V˙(ϕ)=2(Dϕ)(-aϕ(0)-aqϕ(-r))+2aq2(ϕ2(0)-ϕ2(-r))=-2a(1-q2)ϕ2(0). Thus, according to the last conclusion of Theorem  12.7.1 in [20, Page 297], the zero solution of u(t)-qu(t-r)=-au(t)-aqu(t-r) is uniformly asymptotically stable. On the other hand, one can see that (38)Fϕu=(-gx(u(0),u(-r)),-gy(u(0),u(-r))). Thus, F(0)=Fϕ(0)=0. According to [20, Theorem  12.9.1], the zero solution of (34) is exponentially asymptotically stable.

3. Examples

In this section, we present two examples to illustrate the applicability of our main results.

Example 6 (Lopes et al. [<xref ref-type="bibr" rid="B8">8</xref>, <xref ref-type="bibr" rid="B9">9</xref>, <xref ref-type="bibr" rid="B13">13</xref>, <xref ref-type="bibr" rid="B15">15</xref>, <xref ref-type="bibr" rid="B23">23</xref>]).

Consider the NFDE (2) which arises from a transmission line model, where a>0,b>0,  r>0,  |q|<1,  pC(), and f is a given nonlinear function. Now, let h(u(t),u(t-r))=bf(u(t))-bqf(u(t-r)). It is not difficult to see that (2) is a special case of (1). Therefore, by Theorems 25, we have the following.

Theorem 7.

Suppose that fC() and pCT. If there exists a constant H>0 such that (39)p0<(1-3|q|)aH-b(1+|q|)sup|x|H|f(x)|, then (2) has a T-periodic solution.

Remark 8.

Theorem 7 implies that the conditions in  (40)l<1,|q|<1-l3+l, where l=(b/aH)sup|x|H|f(x)|, are unnecessary for the existence of periodic solutions for (2).

Theorem 9.

Let H be as in Theorem 7. Assume that all conditions of Theorem 7 are satisfied. If f satisfies the Lipschitz condition, (b/aH)sup|x|H|f(x)|1 and there exists Q>0 such that (41)ψ-u*Q-3|q|Q1+|q|-H-basup|x|H+Q|f(x)|, then the solution u(ψ)(t) through ψ to (2) satisfying u(ψ)(t)u*(t) as t, where u*(t) is a T-periodic solution of (2).

Remark 10.

The sufficient conditions for the existence of periodic solutions in  are very complicated. For example, they need extra condition Q>H, m<Q-H and (42)|q|<Q-H-m3Q+H+m,ψ-u*Q-3|q|Q1+|q|-H-m1+|q|, where m=(b/a)sup|x|H+Q    |f(x)|.

Theorem 11.

If all conditions of Theorem 7 are satisfied, and the Lipschitz constant L for f corresponding to (-,+) satisfies 1-3|q|-(b/a)(1+|q|)L>0, then the T-periodic solution u*(t) of (2) is stable.

Theorem 12.

Suppose that p is constant, the equation p-a(1+q)u=b(1-q)f(u) has only one solution u*, fC1(), and f(u*)=0; then the equilibrium u* of (2) is exponentially asymptotically stable.

Example 13 (Lopes [<xref ref-type="bibr" rid="B9">9</xref>]).

Consider the NFDE (3) which arises from a transmission line model, where C0>0,  Z>0,  r>0,|k|<1,  pC() and f is a given nonlinear function. Let p~(t)=(1/C0)p(t), a~=Z/C0, and f~(u)=(1/C0)f(u); then (3) can be rewritten as (43)u(t)-ku(t-r)=p~(t)-a~u(t)-a~ku(t-r)-f~(u(t)-ku(t-r)). Now, let h(u(t),u(t-r))=f~(u(t)-ku(t-r)). It is not difficult to see that (43) is a special case of (1). Therefore, by Theorems 25, we have the following.

Theorem 14.

Suppose that fC() and pCT. If there exists a constant H>0 such that (44)p0<(1-3|k|)ZH-sup|x|(1+|k|)H|f(x)|, then (3) has a T-periodic solution.

Theorem 15.

Let H be as in Theorem 14. Assume that all conditions of Theorem 14 are satisfied. If f satisfies the Lipschitz condition, sup|x|(1+|k|)H|f(x)|ZH(1+|k|), and there exists Q>0 such that (45)ψ-u*Q-3|k|Q1+|k|-H-1Z(1+|k|)sup|x|(1+|k|)(H+Q)|f(x)|, then the solution through ψ of (3) u(ψ)(t)u*(t) as t, where u*(t) is a T-periodic solution of (3).

Theorem 16.

If all conditions of Theorem 14 are satisfied, and the Lipschitz constant L for f corresponding to (-,+) satisfies 1-3|k|-(L/Z)(1+|k|)>0, then T-periodic solution u*(t) of (3) is stable.

Theorem 17.

Suppose that p is constant, the equation p-Z(1+k)u=f(u-ku) has only one solution u*, fC1(), and f(u*-ku*)=0, then the equilibrium u* of (3) is exponentially asymptotically stable.

Acknowledgment

This work was partially supported by the National Natural Science Foundation of China nos. 11201411 and 11031002.

Brayton R. K. Bifurcation of periodic solutions in a nonlinear difference-differential equations of neutral type Quarterly of Applied Mathematics 1966 24 215 224 MR0204800 Brayton R. K. Nonlinear oscillations in a distributed network Quarterly of Applied Mathematics 1967 24 289 301 Brayton R. K. Miranker W. L. A stability theory for nonlinear mixed initial boundary value problems Archive for Rational Mechanics and Analysis 1964 17 358 376 MR0168864 Brayton R. K. Moser J. K. A theory of nonlinear networks. I Quarterly of Applied Mathematics 1964 22 1 33 MR0169746 Cooke K. L. Krumme D. W. Differential-difference equations and nonlinear initial-boundary value problems for linear hyperbolic partial differential equations Journal of Mathematical Analysis and Applications 1968 24 372 387 MR0232089 Ferreira J. M. On the stability of a distributed network SIAM Journal on Mathematical Analysis 1986 17 1 38 45 10.1137/0517005 MR819210 Krawcewicz W. Wu J. Xia H. Global Hopf bifurcation theory for condensing fields and neutral equations with applications to lossless transmission problems The Canadian Applied Mathematics Quarterly 1993 1 2 167 220 MR1226773 Lopes O. Forced oscillations in nonlinear neutral differential equations SIAM Journal on Applied Mathematics 1975 29 196 207 MR0382811 Lopes O. Stability and forced oscillations Journal of Mathematical Analysis and Applications 1976 55 3 686 698 MR0446657 Siqueira Ceron S. Lopes O. α -contractions and attractors for dissipative semilinear hyperbolic equations and systems Annali di Matematica Pura ed Applicata. Serie Quarta 1991 160 193 206 10.1007/BF01764127 MR1163208 Lu S. Ge W. On the existence of periodic solutions for neutral functional differential equation Nonlinear Analysis: Theory, Methods & Applications 2003 54 7 1285 1306 10.1016/S0362-546X(03)00187-1 MR1995931 Miranker W. L. Periodic solutions of the wave equation with a nonlinear interface condition International Business Machines Corporation 1961 5 2 24 MR0118307 Slemrod M. Nonexistence of oscillations in a nonlinear distributed network Journal of Mathematical Analysis and Applications 1971 36 22 40 MR0282028 Wei J. J. Ruan S. G. Stability and global Hopf bifurcation for neutral differential equations Acta Mathematica Sinica. Chinese Series 2002 45 1 93 104 MR1923668 Ding L. Li Z. Periodicity and stability in neutral equations by Krasnoselskii's fixed point theorem Nonlinear Analysis: Real World Applications 2010 11 3 1220 1228 10.1016/j.nonrwa.2009.02.013 MR2646539 Burton T. A. Perron-type stability theorems for neutral equations Nonlinear Analysis: Theory, Methods & Applications 2003 55 3 285 297 10.1016/S0362-546X(03)00240-2 MR2007475 Kolmanovskiĭ V. B. Nosov V. R. Stability of functional-differential equations 1986 180 London, UK Academic Press xiv+217 Mathematics in Science and Engineering MR860947 Kuang Y. Delay Differential Equations with Applications in Population Dynamics 1993 191 Boston, Mass, USA Academic Press xii+398 Mathematics in Science and Engineering MR1218880 Ding T.-R. Li C. Ordinary Differential Equations Tutorial 2004 Bejing, China Higher Education Press Hale J. K. Functional differential equations Analytic Theory of Differential Equations 1971 183 Berlin, Germany Springer 2 Lecture Notes in Mathematics MR0390425 Xia D. X. Wu Z. R. Yan S. Z. Shu W. C. Real Variable Function and Functional Analysis 2010 2nd Beijing, China Higher Education Press Smart D. R. Fixed Point Theorems 1974 London, UK Cambridge University Press viii+93 MR0467717 Angelov V. G. Lossy transmission lines terminated by R-loads with exponential V-I characteristics Nonlinear Analysis: Real World Applications 2007 8 2 579 589 10.1016/j.nonrwa.2006.01.002 MR2289570