We study the multiplicity of periodic solutions of nonautonomous delay differential equations which are asymptotically linear both at zero and at
infinity. By making use of a theorem of Benci, some sufficient conditions are obtained to guarantee the existence of multiple periodic solutions.
1. Introduction
The existence and multiplicity of periodic solutions of delay differential equations have received a great deal of attention. In 1962, Jones [1] firstly investigated the existence of periodic solutions to the following scalar equation:u′(t)=-au(t-1)[1+u(t)].
By making use of Browder fixed point theorem, the author showed that there exist periodic solutions of (1.1) for each a>π/2. Since then, various fixed point theorems have been used to study the existence of periodic solutions of delay differential equations(cf. [2]). As pointed out in [3], by making change of variable 1+u=ex, (1.1) turns intox′(t)=-f(x(t-1)).
In 1974, Kaplan and Yorke [4] studied the following more general form of (1.2)x′(t)=-f(x(t-1))-f(x(t-2))-⋯-f(x(t-n)).
They introduced a technique which translates the problem of the existence of periodic solutions of a scalar delay differential equation to that of the existence of critical points of an associated ordinary differential system. Using this method, they proved that (1.3) has a periodic solution with minimal period 4 (resp., 6) when (1.3) has one delay (resp., two delays). In this direction, Fei, Li and He did some excellent work and got some signification results (cf. [5–8]).
Many other approaches, such as coincidence degree theory, the Hopf bifurcation theorem, and the Poincaré-Bendixson theorem, have also been used to study the existence of periodic solutions of delay differential equations (cf. [9, 10]). However, most of those results are concerned with scalar delay equations. In 2005, Guo and Yu [3] studied vector delay differential system (1.2). They built a variational structure for (1.2) on certain suitable spaces. Then they reduced the existence of periodic solutions of (1.2) to that of critical points of an associated variational functional. By making use of pseudoindex theory, they obtained some sufficient conditions to guarantee the existence of multiple periodic solutions.
In spite of so many papers on periodic solutions of delay differential equations, there are a quite few researches on nonautonomous case (see for example [11]). The main goal of this paper is to investigate the following nonautonomous system:x′(t)=-f(t,x(t-π2)).
We assume that
(f1) there exists F∈C1([0,π/2]×ℝn,ℝ) such that f is the gradient of F with respect to x, and
F(t,x)=F(t,-x),F(t+π2,x)=F(t,x),∀(t,x)∈R×Rn,
(f2)f(t,x)=B0(t)x+o(|x|) as |x|→0 uniformly for t∈[0,π/2],
(f3)f(t,x)=B∞(t)x+o(|x|) as |x|→∞ uniformly for t∈[0,π/2],
where B0,B∞ are n×n symmetric continuous π/2-periodic matrix functions.
Hypothesis (f3) is known as asymptotically linear condition at infinity. Hypothesis (f2) is an asymptotically linear condition at zero, which implies that 0 is a trivial solution of (1.4). We are interested in nontrivial periodic solutions of (1.4). Similar to [3], we build a variational structure for (1.4) and convert the existence of periodic solutions to that of critical points of variational functional. Since the asymptotically linear hypothesis at infinity is given by a periodic loop of symmetric matrix, it will be more difficult to deal with more than a constant matrix. However, we can prove the existence of multiple periodic solutions by making use of a multiple critical points theorem of Benci (cf. [12]).
The rest of this paper is organized as follows: in Section 2, we build the variational functional and state some useful lemmas; in Section 3, the main results will be proved.
2. Variational Tools
Denote S1=ℝ/(2πℤ). The space H=H1/2(S1,ℝn) has been introduced in [3]. The space H can be equipped with inner product as follows: 〈x,y〉=(a0,c0)+∑j=1∞(1+j)[(aj,cj)+(bj,dj)],
where x=a0/2π+1/π∑j=1∞(ajcosjt+bjsinjt), y=c0/2π+1/π∑j=1∞(cjcosjt+djsinjt), a0,c0∈ℝn, aj,cj,bj,dj∈ℝn, j∈ℕ.
Set E={x∈H∣x(t+π)=-x(t),∀t∈R}.
Then E is a closed subspace of H. If x∈E, it has Fourier expansion x(t)=1π∑j=1∞[ajcos(2j-1)t+bjsin(2j-1)t].
Let x∈L2(S1,ℝn). If for every z∈C∞(S1,ℝn)∫02π(x(t),z′(t))dt=-∫02π(y(t),z(t))dt,
then y is called a weak derivative of x, denoted by ẋ.
The variational functional defined on H, corresponding to (1.4), isJ(x)=∫02π[12(x(t+π2),ẋ(t))-F(t,x(t))]dt.
Define a linear bounded operator A:H→H by setting〈Ax,y〉=∫02π(x(t+π2),ẏ(t))dt.
It is easy to prove that E is an invariant subspace of H with respect to A and A is self-adjoint if it is restricted to E.
Lemma 2.1 (see [3]).
The essential spectrum of the operator A restricted to E is just {2,-2}.
Define φ(x)=-∫02πF(t,x(t))dt,∀x∈H.
Then J can be rewritten asJ(x)=12〈Ax,x〉+φ(x),∀x∈H.
Similar to the argument as in [3], we can prove the following two basic lemmas.
Lemma 2.2.
Assume that f satisfies (f1)–(f3). Then J is continuous differentiable on H and
〈J′(x),h〉=∫02π[12(ẋ(t-π2)-ẋ(t+π2),h(t))-(f(t,x(t)),h(t))]dt,∀h∈H.
Moreover, φ′:H→H* is a compact mapping defined as follows:
〈φ′(x),h〉=-∫02π(f(t,x(t)),h(t))dt,∀x,h∈H.
Lemma 2.3.
The existence of 2π-periodic solutions of (1.4) belonging to E is equivalent to the existence of critical points of functional J restricted to E.
Lemma 2.3 implies that we can restrict our discussion on space E. At the end of this section, we recall a useful embedding theorem.
Lemma 2.4 (see [13]).
For every p∈[1,+∞), H is compactly embedded into the Banach space Lp(S1,ℝn). In particular, there is an αp such that
‖x‖Lp≤αp‖x‖,∀x∈H.
Remark 2.5.
Here and hereafter, αp(p∈[1,∞)) denotes the real number satisfying (2.11).
3. Main Results
Let B(t) be an n×n symmetric continuous π/2-periodic matrix function. We define a bounded self-adjoint linear operator B∈L(E) by extending the bilinear forms〈Bx,y〉=∫02π(B(t)x(t),y(t))dt,∀x,y∈E.
It is well known that B is compact (cf. [14]).
Denote by B0,B∞ the operators defined by (3.1), corresponding to B0(t),B∞(t), respectively. Set n0=dim Ker (A-B0),n∞=dim Ker(A-B∞),Gi(t,x)=F(t,x)-12(Bi(t)x,x),ψi(x)=∫02πGi(t,x)dt,i=0,∞.
Then the functional J defined by (2.5) can be rewritten asJ(x)=12〈(A-Bi)x,x〉-ψi(x),∀x∈E,i=0,∞.
Lemma 3.1.
Suppose that f satisfies (f1)–(f3). Then
lim‖x‖→0‖ψ0′(x)‖‖x‖=0,lim‖x‖→+∞‖ψ∞′(x)‖‖x‖=0.
The proof uses the same arguments of [5].
In order to prove our results, we need an abstract theorem by Benci [12].
Proposition 3.2.
Let χ∈C1(E,ℝ) satisfy the following:
χ(x)=1/2〈Lx,x〉+ω(x), where L is a bounded linear self-adjoint operator and ω′ is compact, where ω′ denotes the Frechét derivative of ω;
every sequence {xj} such that χ(xj)→c<φ(0) and ∥χ′(xj)∥→0 as j→+∞ has a convergent subsequence;
ω(x)=ω(-x), x∈E;
there are two closed subspaces of E, E+, and E-, and some constant c0,c∞,ρ with c0<c∞<ω(0) and ρ>0 such that
χ(x)>c0 for x∈E+,
χ(x)<c∞<ω(0) for u∈E-∩Sρ(Sρ={u∈E∥x∥=ρ}).
Then the number of pairs of nontrivial critical points of χ is greater than or equal to dim(E+∩E-)-codim(E-+E+). Moreover, the corresponding critical values belong to [c0,c∞].
Definition 3.3.
Let B1(t) and B2(t) be symmetric matrices function in ℝn, continuous and π/2-periodic in t. A index I of B1(t) and B2(t) is defined as follows:
I(B1(t),B2(t))=dim(M+(A-B1)⋂M-(A-B2))-dim[(M-(A-B1)⊕M0(A-B1))⋂(M+(A-B2)⊕M0(A-B2))],
where Bi(i=1,2) are the operators, defined by (3.1), corresponding to Bi(t)(i=1,2) and M+(A-Bi) (resp., M-(A-Bi(t)), M0(A-Bi)) denotes the subspace of E on which A-Bi is positive definite (resp., negative definite, null).
Lemma 3.4.
If f satisfies (f1)–(f3), then J, defined by (2.8), satisfies (J1), (J3) and (J4).
Proof.
Hypothesis (f1), (2.8), and Lemma 2.2 imply both (J1) and (J3). By definition of ψ0 and Lemma 3.1, we have
ψ0(x)=-φ(0)+o(‖x‖2),for‖x‖⟶0.
Since B0 and B∞ are compact operators from E to E, it follows from Lemma 2.1 and a well-known theorem (cf. [15]) that the essential spectrum of A-B0 and A-B∞ is {2,-2}. Thus 0 is either an isolated eigenvalue of finite multiplicity or it belongs to the resolvent. Hence, we decompose E as follows:
E=M+(A-B0)⊕M0(A-B0)⊕M-(A-B0)=M+(A-B∞)⊕M0(A-B∞)⊕M-(A-B∞).
Setting E+=M+(A-B∞), E-=M-(A-B0), there exists positive constant α,β such that
〈(A-B0)x,x〉≤-α‖x‖2,∀x∈E-,〈(A-B∞)x,x〉≥β‖x‖2,∀x∈E+.
It follows that, for any x∈E-, it is
J(x)≤-α2‖x‖2+φ(0)-o(‖x‖2),as‖x‖⟶0.
Then there exist constants ρ>0 and γ>0 such that
J(x)<-γ+φ(0),∀x∈E-∩Sρ.
Setting c∞=-γ/2+φ(0), (J4)(b) is satisfied.
By (f3), there exists R0>0 such that
G∞(t,x)≤β4α22|x|2,∀|x|>R0.
Since G∞(t,x) is continuous with respect to (t,x), denote by M=max0≤t≤π/2,|x|=R0{G∞(t,x)}. Then M is finite. Thus|ψ∞(x)|≤∫02π|G∞(t,x)|dt≤∫02π[β4α22|x|2+M]dt≤β4‖x‖2+2πM.
Then, for every x∈E+,
J(x)=12〈(A-B∞)x,x〉-ψ∞(x)≥β2‖x‖2-|ψ∞(x)|≥β4‖x‖2-2πM.
Thus J is bounded from below on E+. Setting
c0=infx∈E+J(x)-ω
with ω>0 such that c0<c∞, then (J4)(a) is satisfied.
Remark 3.5.
Supposing that 0∉σe(A-B∞), any bounded sequence has a convergent subsequence (cf. [12]).
Theorem 3.6.
Suppose that f satisfies (f1)–(f3), and n0=n∞=0, then (1.4) has at least |I(B∞,B0)| pairs of nonconstant 2π-periodic solutions if |I(B∞,B0)|>0.
Proof.
Since n∞=0, dimM0(A-B∞)=0. By Proposition 3.2 and Lemma 3.4, we only need to check (J2). Let {xj} be a sequence such that
J′(xj)→0,J(xj)⟶c,
where c∈ℝ, c<φ(0). Suppose to the contrary that we can choose ∥xj∥→+∞ as j→+∞. Clearly, xj can be written as xj=xj++xj-∈M+(A-B∞)⊕M-(A-B∞). On one hand,
|〈J′(xj),xj+-xj-〉||〈xj,xj〉|≤‖J′(xj)‖‖xj‖‖xj‖2,
then we have
0≤limsupj→+∞|〈J′(xj),xj+-xj-〉||〈xj,xj〉|≤limsupj→+∞‖J′(xj)‖‖xj‖‖xj‖2=0.
Thus
limsupj→+∞|〈J′(xj),xj+-xj-〉||〈xj,xj〉|=0.
On the other hand,
〈J′(xj),xj+-xj-〉=〈(A-B∞)xj,xj+-xj-〉-〈ψ∞′(t,xj),xj+-xj-〉.
Since
|〈ψ∞′(xj),xj+-xj-〉||〈xj,xj〉|≤‖ψ∞′(xj)‖‖xj‖‖xj‖2=‖ψ∞′(xj)‖‖xj‖,
it follows by Lemma 3.1 that
limj→+∞|〈ψ∞′(xj),xj+-xj-〉||〈xj,xj〉|=0.
Using a similar discussion as (3.8), there exists β0>0 such that 〈(A-B∞)x,x〉≤-β0∥x∥2 for all x∈M-(A-B∞). Choosing β′=min(β,β0)>0, we have
〈(A-B∞)xj,xj+-xj-〉=〈(A-B∞)xj+,xj+〉-〈(A-B∞)xj-,xj-〉≥β′‖x‖2.
Thus,
liminfj→+∞|〈J′(xj),xj+-xj-〉||〈xj,xj〉|=liminfj→+∞|〈(A-B∞)xj,xj+-xj-〉-〈g∞′(xj),xj+-xj-〉||〈xj,xj〉|=liminfj→+∞|〈(A-B∞)xj,xj+-xj-〉||〈xj,xj〉|≥β′>0
which contradicts (3.18). This proves (J2). By Lemma 3.1, (1.4) has at least dim(E+∩E-)-codim(E-+E+)=I(B∞,B0) pairs of nontrivial solutions if I(B∞,B0)>0. Since the Sobolev space E does not contain ℝn as its subspace, all nontrivial periodic solutions are nonconstant periodic solutions.
If I(B∞,B0)<0, then I(B0,B∞)=-I(B∞,B0)>0. In this case, we replace J by -J and let E+=M-(A-B0) and E-=M+(A-B∞). It is easy to see that (J1)–(J4) are satisfied. Similarly, we can show that (1.4) has at least I(B0,B∞) pairs of nonconstant solutions.
Remark 3.7.
When Theorem 3.6 is applied to autonomous delay differential equations, we obtain the same number of periodic solutions as that in [3].
Theorem 3.8.
Suppose f satisfies (f1)–(f3) and
G∞′(t,x) is bounded, where G∞′ denotes the derivative of G∞ with respect to x,
G∞(t,x)→±∞ as |x|→+∞, uniformly for t∈[0,π/2].
Then (1.4) has at least I(B∞,B0) pairs of nonconstant 2π-periodic solutions provided I(B∞,B0)>0.
Proof.
By Proposition 3.2 and Lemma 3.4, it suffices to check condition (J2). Let {xj} be a sequence satisfying (3.15). Suppose to the contrary that {xj} is unbounded. Clearly, xj can be written as xj=xj++xj0+xj-∈M+(A-B∞)⊕M0(A-B∞)⊕M-(A-B∞). Since J′(xj)→0, for j large enough, we get
|〈(A-B∞)xj,xj+〉-∫02π(G∞′(t,xj),xj+)dt|≤‖xj+‖.
By (f4), there exists c1>0 such that |G∞′(t,x)|≤c1. Then the above inequality and (3.8) imply
β‖xj+‖2≤|〈(A-B∞)xj,xj+〉|≤‖xj+‖+c1α22π‖xj+‖.
This gives a uniform bound for {xj+}. In the same manner, one gets a uniform bound for {xj-}. Since {J(xj)} is convergent, it is bounded and there exist positive constants c2,c3,c4 such that
c2≤J(xj)≤-ψ∞(xj)+12|〈(A-B∞)xj,xj〉|≤-ψ∞(xj0)+(ψ∞(xj0)-ψ∞(xj))+c3≤-ψ∞(xj0)+c1∫02π|xj0-xj|dt+c3≤-ψ∞(xj0)+c4.
Therefore, ψ∞(xj0) is bounded from above. (f5)+ implies that ∥xj0∥ is bounded. Otherwise, since the kernel of A-B∞ is a finite dimensional space, thus ψ∞(xj0)=∫02πG∞(t,xj0)dt→∞ as j→∞, which contradicts to (3.26).
If (f5)- holds, we replace (3.26) by
c2≥J(xj)≥-ψ∞(xj)-|〈(A-B∞)xj,xj〉|.
Arguing as above, we can get a contradiction and complete our proof.
Theorem 3.9.
Suppose that f satisfies (f1)–(f3) and
there exist constants r>0, p∈(1,2), a1>0, and a2>0 such that
pG∞(t,x)≥(x,G∞′(t,x))>0for|x|≥r,t∈[0,π2];G∞(t,x)≥a1|x|p-a2,∀x∈Rn,t∈[0,π2].
Then (1.4) has at least I(B∞,B0) pairs of nonconstant 2π-periodic solutions provided I(B∞,B0)>0.
Proof.
Let {xj} be a sequence satisfying (3.15). We want to show that {xj} is a bounded sequence in E. Decompose xj as xj=xj++xj0+xj-∈M+(A-B∞)⊕M0(A-B∞)⊕M-(A-B∞). Then
〈J′(xj),xj+〉=〈(A-B∞)xj+,xj+〉-〈ψ∞′(xj),xj+〉≥β‖xj+‖2-‖ψ∞′(xj)‖⋅‖xj+‖.
Combining the above inequality with (3.15) and Lemma 3.1, we have
‖xj+‖‖xj‖→0,asj→∞.
Similarly, we have
‖xj-‖‖xj‖⟶0,asj⟶∞.
Then by (3.30) and (3.31), there exists a positive integer j0 such that for j≥j0‖xj0‖≥‖xj++xj-‖.
It follows that
‖xj‖=‖xj++xj-+xj0‖≤‖xj++xj-‖+‖xj0‖≤2‖xj0‖.
By (f6), there exist positive constants M1,M2,M3,M4 such that for j large
M1+12‖xj‖≥12〈J′(xj),xj〉-J(xj)=∫02π[G∞(t,xj)-12(G∞′(t,xj),xj)]dt≥(1-p2)∫02πG∞(t,xj)dt-M2≥M3‖xj‖Lpp-M4.
Let q be such that p-1+q-1=1. Since E⊂Lq(S1,ℝn), the embedding being continuous, the dual space E* of E, contains Lp(S1,ℝn) with continuous embedding. Therefore, by (3.34)
M5(1+‖xj‖)≥‖xj‖E*p.
Since ∥xj∥E*=sup∥w∥E≤1(xj,w)L2=sup∥w∥E≤1[(xj0,w0)L2+(xj-,w-)L2+(xj+,w+)L2], taking w=xj0/∥xj0∥E, it follows that
‖xj‖E*≥1‖xj0‖E‖xj0‖L22.
Owing to the fact that M0(A-B∞) is a finite dimensional subspace of E, there exist two positive constants c1 and c2 such that
c1∥xj0∥E≤∥xj0∥L2≤c2∥xj0∥E.
Therefore by (3.35), (3.36), and (3.37),
M6(1+‖xj‖)≥‖xj0‖p.
Both (3.33) and (3.38) imply that there exists M8>0 such that
M7(1+‖xj0‖)≥‖xj0‖p
which yields a bound for ∥xj0∥ and hence xj via (3.33). Thus (J2) holds.
Theorem 3.10.
Suppose that f satisfies (f1)–(f3) and
(f7)± there exist positive constants c1,c2>0 such that
±[2G∞(t,x)-(G∞′(t,x),x)]≥c1|x|-c2∀x∈Rn,t∈[0,π2].
Then (1.4) has at least I(B∞,B0) pairs of nonconstant 2π-periodic solutions provided I(B∞,B0)>0.
Proof.
Let {xj} be a sequence satisfying (3.15). We want to prove that {xj} is bounded in E. Suppose, to the contrary, {xj} is unbounded in E. Decompose xj as xj=xj++xj0+xj-∈M+(A-B∞)⊕M0(A-B∞)⊕M-(A-B∞). Clearly, (3.30)–(3.33) still hold.
Assume that (f7)+ holds. Since M0(A-B∞) is a finite dimensional subspace of E, we have
〈J′(xj),xj〉-2J(xj)=∫02π[2G∞(t,xj)-(G∞′(t,xj),xj)]dt≥c1∫02π|xj|dt-2πc2≥c1∫02π|xj0|dt-c1∫02π(|xj+|+|xj-|)dt-2πc2≥c3‖xj0‖-c4(‖xj+‖+‖xj-‖+1).
Combining the above inequality with (3.30), (3.31), we have
‖xj0‖‖xj‖⟶0asj⟶∞.
But this implies the following contradiction:
1=‖xj‖‖xj‖=‖xj0‖+‖xj-‖+‖xj+‖‖xj‖⟶0asj⟶+∞,
therefore, {xj} must be a bounded sequence.
If (f7)- holds, using a similar argument, we can get a contradiction which completes our proof.
Theorem 3.11.
Suppose that f satisfies (f1)–(f3) and
there exist constants 1≤γ<2, 0<δ<γ/2, and b1,b2,L>0 such that
|G∞′(t,x)|≤b1|x|δ,±G∞(t,x)≥b2|x|γ,∀|x|≥L,t∈[0,π2].
Then (1.4) has at least I(B∞,B0) pairs of nonconstant 2π-periodic solutions provided I(B∞,B0)>0.
Proof.
Let {xj} be a sequence satisfying (3.15). Suppose, to the contrary, ∥xj∥→+∞ as j→+∞. Decompose xj as xj=xj++xj0+xj-∈M+(A-B∞)⊕M0(A-B∞)⊕M-(A-B∞). First, we show that for j large enough
‖xj++xj-‖≤b3‖xj0‖δ+η,
where b3>0 and η>0 are constants independent of j. Since |xj|≥L for sufficiently large j, therefore, |G∞′(t,xj)|≤b1|xj|δ, and we have
|G∞′(t,xj)|2≤b12|xj|2δ+b4;|〈ψ∞′(xj),y〉|≤∫02π|G∞′(t,xj)||y|dt≤(∫02π|G∞′(t,xj)|2dt)1/2‖y‖L2≤α2[b12(2π)1-δα22δ‖xj‖2δ+2πb4]1/2‖y‖,for anyy∈E.
This implies that for j large enough
‖ψ∞′(xj)‖‖xj‖δ≤b5.
By (3.15), (3.22), (3.33) and (3.47), for j large enough, we have
|〈J′(xj),xj+-xj-〉|=|〈(A-B∞)xj,xj+-xj-〉-〈ψ∞′(xj),xj+-xj-〉|≥β′‖xj++xj-‖2-b5‖xj‖δ‖xj+-xj-‖≥β′‖xj++xj-‖2-b52δ‖xj0‖δ‖xj+-xj-‖.
Therefore, for sufficiently large j,
‖J′(xj)‖≥β′‖xj++xj-‖-b52δ‖xj0‖δ.
This implies that (3.45) holds, where b3=b52δ/β′.
By (3.15) and (3.45), for j large enough, there exist positive constants b6,b7,b7′,b8 such that
ψ∞(xj)=12〈(A-B∞)(xj++xj-),xj++xj-〉-J(xj)≤b6‖xj++xj-‖2+b7′≤b8‖xj0‖2δ+b7.
Now, we claim that there exists b9>0 such that, for j large enough,
∫02π|xj|γdt≥b9‖xj0‖γ,
In fact, for γ>1, by (3.45) and the fact that δ<1, we have
∫02π(xj,xj0)dt≤(∫02π|xj|γdt)1/γ(∫02π|xj0|γ/γ-1dt)γ-1/γ≤b10(∫02π|xj|γdt)1/γ‖xj0‖;∫02π(xj,xj0)dt=∫02π(xj0,xj0)dt+∫02π(xj++xj-,xj0)dt≥∫02π|xj0|2dt-‖xj++xj-‖L2‖xj0‖L2≥b11‖xj0‖2-b3α22‖xj0‖1+δ-α22η‖xj0‖≥b12‖xj0‖2,
for j large enough. This implies (3.51) for γ>1.
For γ=1, since M0(A-B∞) is a finite dimensional subspace of E, we know that for any j,
b13‖xj0‖≤‖xj0‖∞≤b14‖xj0‖.
where b13,b14>0 are constants independent of j. Now we have
∫02π(xj,xj0)dt≤∫02π|xj||xj0|dt≤(∫02π|xj|dt)‖xj0‖∞≤b15‖xj0‖(∫02π|xj|dt).
Combining (3.52) with (3.54), we get (3.51) for γ=1.
On the other hand, by (f8)+ψ∞(xj)=∫02πG∞(t,xj)dt≥∫02πb2|xj|γdt-2πb16≥b17‖xj0‖γ-2πb16.
Since that γ>2δ, we get a contradiction from (3.50) and (3.55). Therefore, {xj} is bounded.
If (f8)- holds, using a similar argument as above, we get a contradiction and completes our proof.
Theorem 3.12.
Suppose f satisfies (f1)–(f3) and
there exist positive constants 1≤γ<2, 0<δ<γ/2, and b1,b2,L such that
|G∞′(t,x)|≤b1|x|δ,±〈G∞′(t,x),x〉≥b2|x|γ,∀|x|≥L,t∈[0,π2].
Then (1.4) has at least I(B∞,B0) pairs of nonconstant 2π-periodic solutions provided I(B∞,B0)>0.
Proof.
If (f9)+ holds, for j large enough, by (3.45) and (3.51), we have
∫02π(G∞′(t,xj),xj)dt≤|-〈J′(xj),xj〉+〈(A-B∞)(xj++xj-),(xj++xj-)〉|≤‖xj‖+M1‖xj++xj-‖2≤‖xj0‖+M2‖xj0‖δ+M3‖xj0‖2δ+M4;∫02π(G∞′(t,xj),xj)dt≥b2∫02π|xj|γdt-M5≥M6‖xj0‖γ-M5.
Since γ>2δ, {xj0} is bounded, so is {xj}. Therefore, J satisfies (J2).
In the case that (f9)- holds, using a similar argument, we can verify (J2). This completes the proof.
Example 3.13.
Consider the following nonautonomous delay differential equation
x′(t)=-Mx(t-π2)a+b(t)|x(t-(π/2))|5/2+c|x(t-(π/2))|41+|x(t-(π/2))|4,
where M is a 4×4 matrix, a,c are constants, b∈C([0,π/2],ℝ+).
Case 1.
Let A=diag(0.3,2.7,7.3,9.3), a=1, c=2, and b arbitrary. Computing directly, we have I(B0(t),B∞(t))=10. Applying Theorem 3.6, equation (3.58) has at least 10 pairs of 2π-periodic solutions.
Case 2.
Let A=diag(0.3,2.7,5,10.5), a=2, c=1, and b arbitrary. Then by Theorem 3.8, (3.58) has at least 8 pairs of nonconstant 2π-periodic solutions.
Acknowledgments
This paper is supported by Xinmiao Program of Guangzhou University and National Natural Science Foundation of China (no. 11026059), SRF of Guangzhou Education Bureau (no. 10A012), and Foundation for Distinguished Young Talents in Higher Education of Guangdong (no. LYM09105).
JonesG. S.The existence of periodic solutions of f′(x)=−af(x(t−1))[1+f(x)]19625435450014183710.1016/0022-247X(62)90017-3ZBL0106.29504GraftonR. B.A periodicity theorem for autonomous functional differential equations1969687109024317610.1016/0022-0396(69)90119-3ZBL0175.38503GuoZ.YuJ.Multiplicity results for periodic solutions to delay differential equations via critical point theory200521811535217496410.1016/j.jde.2005.08.007ZBL1095.34043KaplanJ. L.YorkeJ. A.Ordinary differential equations which yield periodic solutions of differential delay equations197448317324036481510.1016/0022-247X(74)90162-0ZBL0293.34102FeiG.Multiple periodic solutions of differential delay equations via Hamiltonian systems. I2006651253910.1016/j.na.2005.06.0112226257ZBL1136.34329FeiG.Multiple periodic solutions of differential delay equations via Hamiltonian systems. II2006651405810.1016/j.na.2005.06.0122226258ZBL1136.34330LiJ.HeX.-Z.Multiple periodic solutions of differential delay equations created by asymptotically linear Hamiltonian systems1998311-2455410.1016/S0362-546X(96)00058-21487530LiJ.HeX.Proof and generalization of Kaplan-Yorke's conjecture on periodic
solution of differential delay equations,1999429957964173658610.1007/BF02880387Mallet-ParetJ.SellG. R.The Poincaré-Bendixson theorem for monotone cyclic feedback systems with delay1996125244148910.1006/jdeq.1996.00371378763ZBL0849.34056MawhinJ.Equivalence theorems for nonlinear operator equations and coincidence degree theory for some mappings in locally convex topological vector spaces197212610636032870310.1016/0022-0396(72)90028-9ZBL0244.47049O'ReganD.WangH.Positive periodic solutions of systems of first order ordinary differential equations2005483-43103252215583ZBL1117.34067BenciV.On critical point theory for indefinite functionals in the presence of symmetries1982274253357210.2307/1999120675067ZBL0504.58014RabinowitzP. H.198665New York, NY, USACBMSCBMS Regional Conference Series in Mathematics845785SzulkinA.ZouW.Infinite dimensional cohomology groups and periodic solutions of asymptotically linear Hamiltonian systems20011742369391184674010.1006/jdeq.2000.3942ZBL0997.37040SchechterM.1986142ndAmsterdam, The NetherlandsNorth-Hollandxiv+310North-Holland Series in Applied Mathematics and Mechanics869254