- Research
- Open Access
- Published:
Periodic solutions of second-order differential equations with multiple delays
Advances in Difference Equations volume 2012, Article number: 43 (2012)
Abstract
By using the critical point theory and S1 index theory, we obtain a new result for the existence and multiplicity of periodic solutions for a class of second-order delay differential equations x" (t) = f (x(t))-[f (x(t- 1))+f (x(t- 2))+...+f (x(t-(N -1)))].
1 Introduction
Inspired by the excellent study in [1], many authors [2–16] studied the following differential delay equations
where f ∈ C (ℝ, ℝ),and N ≥ 2 is an integer.
Kaplan and Yorke [2] introduced a technique of couple system which allows them to reduce the search for periodic solutions of a differential delay equation to the problem of finding periodic solutions for a related system of ordinary differential equations. They study periodic solutions of (1.1) with N = 2, f ∈ C (ℝ, ℝ) is odd, xf (x) > 0 for x ≠ 0 and f satisfies some suitable conditions near 0 and ∞. More precisely, if the solution x(t) of (1.1) with N = 2 satisfies x(t) = -x(t-2), let
then X(t) = (x1(t), x2(t))T satisfies
where,
i.e., A2 is a skew symmetric matrix, and
∇H(X) is the gradient of H.
In fact, by direct computation, one has the following proposition:
Proposition 1.1. (i) Any solution x(t) of (1.1) with N = 2, and x(t) = -x(t - 2) will give a solution of (1.3) X(t) = (x1(t), x2(t))T by (1.2). Moreover, X(t) has the following symmetric structure
(ii) Any solution X(t) = (x1(t), x2(t))T of (1.3) with the symmetric structure (1.5) will give a solution of (1.1) by letting x(t) = x1 (t). Moreover, x(t + 2) = -x(t).
Kaplan and Yorke proved that (1.3) has periodic solutions with the symmetric structure (1.5), which give the Kaplan-Yorke type periodic solutions of (1.1) with period 4, i.e., x(t) satisfying x(t) = -x(t - 2). They further conjectured that similar result should be true for the general case N ≥ 2, i.e., under similar conditions for f, (1.1) has a 2N-periodic Kaplan-Yorke type periodic solution x(t), i.e., x(t) satisfying x(t) = -x(t - N).
Li and He [6–8], in an attempt to reuse Kaplan and Yorke's original idea, applied Lyapunov Center Theorem and some known results about convex Hamiltonian systems [17, Theorem 7.2] to obtain 4-periodic solutions of (1.4). But those 4-periodic solutions obtained by [17, Theorem 7.2] give no information about the symmetric structure (1.5) or the minimal period. The solutions of (1.3), which will not generate noncontact solutions of (1.1), see [14, Remark 3.3].
Herz [12] study (1.1) with N = 2 by Lyapunov direct method. And, Jekel and Johnston [13], proved the existence of a 2N-periodic Kaplan-Yorke type periodic solution for (1.1) by Kaplan-Yorke original method and homotopic method.
Fei [14, 15] applied the pseudo-index theory [17–20] to obtain periodic solution in a subspace, which surely have the required symmetric structure (1.5) and give solutions to (1.1).
In recent years, Guo and Yu [16] considered (1.1) with N = 2 by variational methods directly, and they obtain the Kaplan-Yorke type periodic solutions. That is to say that they do not necessarily transform the existence problem of (1.1) to existence problems for related systems (1.3). Afterwards, Cheng and Hu [21] studied (1.1) with N = 2 by Guo-Yu's method in [16]. Guo [22] studied the following second-order differential delay equation by Guo-Yu's method in [16]
they obtained the multiplicity results for periodic solutions, but the solutions are not Kaplan- Yorke type.
The authors [23, 24] considered the Kaplan-Yorke type periodic solutions of (1.1) with N = 2 by Maslov-type index [25] and Morse theory [26], respectively.
Recently, some researchers [27–31] have begun to study the existence of solutions for second-order differential delay equation by using a variational method. However, to the best of authors' knowledge, the study of Kaplan-Yorke type periodic solutions of second- order differential delay equation using a variational method has received considerably less attention. We find the method apply in [6–15], such as the structure of variational does not directly apply to second-order differential delay equation.
Motivated by the study in [7–16, 21–24, 27–32], in this article we are concerned with the existence of Kaplan-Yorke type periodic solutions of the following second-order differential delay equations
where f ∈ C (ℝ,ℝ) is odd, and N ≥ 2 is an integer
If x(t) = x(t+N), let
then z(t) = (x1(t), x2(t),..., x N (t))T satisfies
where
i.e., A N is a n × n symmetric matrix, and
In fact, by direct computation, one has the following proposition.
Proposition 1.2. (i) Any solution x(t) of (1.7) with x(t) = x(t - N) will give a solution of (1.9) X(t) = (x1(t), x2 (t), . . . , x N (t))T by (1.8). Moreover, X(t) has the following symmetric structure
(ii) Any solution X(t) = (x1(t), x2(t), . . . , x N (t))T of (1.9) with the symmetric structure (1.11) will give a solution of (1.7) by letting x(t) = x1(t). Moreover, x(t + N) = x(t).
Throughout this article, we always assume that:
(f 1) f ∈ C (ℝ, ℝ) is odd and 0 <α, β < +∞
(f 2±) |f (x) - βx| is bounded and G β (x) → ± ∞, as |x| → ∞;
(f 3±) ±Gα (x) > 0 for |x| > 0 being small,
where , and
Similarly to the argument in [14, 20, 33], for given a number α ∈ ℝ, and N ≥ 2, k ≥ 1 being two integer, we set:
and
For convenience, denote #(1.7) = the number of geometrically different nonconstant periodic solutions of (1.7) which satisfy x(t - N/2) = -x(t), ∀t ∈ R.
Our main result reads as:
Theorem 1.1. Suppose f satisfies (f 1) and N ≥ 2 being an integer in (1.7). We have the following conclusions:
(i) #(1.7) ≥ i(α, N) - i(β, N) provided v(β, N) = 0 or (f 2-) holds.
(ii) #(1.7) ≥ i(α, N) - i(β, N) + v(α, N) provided (f 3+)holds and either v(β, N) = 0 or (f 2-) holds.
(iii) #(1.7) ≥ i(β, N) + v(β, N) - i(α, N) provided (f 3-) and (f 2+) holds.
2 Variational structure
For S1 = ℝ/(N ℤ),let E = H1(S1, ℝN). Then E is a Hilbert space with norm ||·|| and inner product〈,〉, and E consists of those z(t) in L2(S1, ℝN) whose Fourier series
satisfies
where a m , b m ∈ ℝN and .
We can define an operator
on E. By direct computation, L0 is a bounded self-adjoint linear operator on E and
By (f1), one can show that H(z) ∈ C1 (ℝℕ, ℝ) and satisfies
where d1, d2 > 0. By using similar arguments as in [14, 21], we know that
and critical points of φ in E are classic solutions of (1.9).
Let T N be the N × N matrix given by
For z(t) ∈ E, define
Then we have δNz(t) = z(t - N), and G = {δ, δ2, . . . , δN} is a compact group action over E. Moreover, if δz(t) = z(t) holds, z(t) has the symmetric structure (1.11).
Lemma 2.1. Denote Then we have
where and
Proof. For any, , we must have z(t-N/2) = -z(t), which implies that
for even m.
this means that for any z ∈ SE,
Note that δz(t) = z(t) ⇔ T N z(t - 1) = z(t), i.e.,
This implies that for k ≥ 1
If we introduce complex vector C m = a m + ib m , the above (2.7) becomes
Note that
the eigenvalues of T N are
Equation (2.7) implies that C m must be the eigenvector associated with the eigenvalue
We obtain that
By direct computation, iC m is also the eigenvector associated with the eigenvalue.
Therefore, we have the conclusion. The proof is complete. □
Lemma 2.2. Let φ be given in (2.3) and φ | SE be the restriction of φ on SE. Then critical points of φ | SE over SE are critical points of φ over E.
Proof. By (1.9) and direct computation, we have
Combining these with (2.3) and the fact that any z(t) ∈ E is N-periodic, one can easily verify that
i.e., φ is G-invariant, and φ', is G-equivariant. The conclusion follows directly. □
For any α ∈ ℝ, define an operator L α by extending the bilinear form
on E. Moreover, L α is G-equivariant. By direct computation, L α is a bounded self-adjoint linear operator on E and if z(t) ∈ SE
For m ≥ 1, denote
Then . Since , and , so SE(m) is the eigen-subspace of L0 corresponding to eigenvalue λ m , here λ m = 2 - N, as (2m - 1 mod N) = 0; λ m = 2, as (2m - 1 mod N) ≠ 0. Denote by M -(·), M +(·) and M0(·) the positive definite, negative definite and null subspaces of the self-adjoint linear operator defining it, respectively.
Lemma 2.3. For k ≥ 1, . Then L α , as an operator on SE, has the following properties on SE.
Proof. For m ≥ 1 and , consider the eigenvalue problem
By (2.7) and (2.11), we have
Since a m , b m are the eigenvector of A N corresponding to eigenvalue λ m , then . Therefore L α is positive definite, negative definite, or null on SE(k) if and only if is positive, negative, or zero, respectively.
This implies (2.12) directly. □
For S1 = ℝ/(N ℤ), there is a natural S1-action over SE, defined by
It is easy to see that φ is S1-invariant, φ' is S1-equivariant, and
By directly applying [20, Theorem 2.4] to φ over SE, we have the following lemma.
Lemma 2.4. Assume there exist two closed S1-invariant linear subspaces, SE+ and SE-, of SE and r > 0 such that (a) (SE+ + SE-) is closed and of finite codimension in SE,
(b) L(SE-)∈ SE-with L = L α or L = L β ,
(c) there exist c0 ∈ ℝ, c0 > -∞ such that
(d) there exists c ∞ ∈ℝ such that
(e) φ satisfies (PS)c condition for c0 ≤ c ≤ c ∞ , i.e., every sequence {z m } ⊆ SE with φ (z m ) → c and φ'(z m ) → 0 possesses a convergent subsequence. Then φ possesses at least geometrically different critical orbits in φ-1([c0, c ∞ ]).
3 Proof of main results
Proof of Theorem 1.1 As we already proved in Lemma 2.2, critical points of φ over SE are critical points of φ over E. Hence they are nonconstant classic N -periodic solutions of (1.7) with the symmetric structure (1.11). By Proposition 1.2, they give solutions of (1.7) with the property x(t - N/2) = -x(t). Therefore, we can seek critical points of φ on SE directly.
Set:
Then
Case (i): If i(α, N ) > i(β, N ). We shall carry out the proof in several steps.
Step 1: let
Then Lemma 2.4(a)and (b) hold with L = L α . By (f1), using the same argument as [18, Lemma 5.5], it is easy to show that
We denote
and
Since ψ α (0) = 0, , as ||z|| → 0, for , there exists a constant r > 0 such that for any z ∈ H1 (S1, ℝN)
Furthermore,
Therefore, we have for z ∈ (E- ∩ S r ) = {z ∈ E-| ||z|| = r}
Thus (d) of Lemma 2.4 holds.
If M 0 (Lβ)= {0}, by (f1), using the same argument as [18, Lemma 5.5], it is easy to show that
it follows that given there exists r > 0, such that
Moreover, there is d1 > 0 such that
thus
Using the above formula we get
Since for any z ∈ SE+= M+(Lβ)⊕M0 (Lβ)
Then for any z ∈ SE+ = M+(Lβ) ⊕ M 0(Lβ)= M + (Lβ) ⊕ {0}
Thus there exists c0 > -∞ such that (c) of Lemma 2.4 holds.
If M0(L β ) ≠ {0}, by (f 2-) and (1.12), for any M > 0, there exists a constant d2 > 0 such that
Let
and
Since F (z) ∈ C2 (ℝN, ℝ), by (3.3), there exists a constant d3 > 0 such that
By (3.1) - (3.4), we have
Then we have for z = z+ + z0 ∈ M+(L β ) ⊕ M 0 (L β )
Thus φ is bounded from below on E+. Therefore, condition (c) of Lemma 2.4 is verified.
Step 2: Using the same argument as [22, 25] and [26, Lemma 4.2], one can prove that φ satisfies (PS)c condition for any c ∈ ℝ under the condition either (f 1) with v(β, N) = 0 holds or (f 2±) holds.
Step 3: By Lemma (2.4), φ has at least geometrically different critical orbits in φ-1 ([c0, c∞]).
Now by Lemma (2.3), it is easy to show that
This means #(1.7) ≥ i(α, N)-i(β, N), and this completes the proof of case (i).
For case (ii), (iii), using the same idea and similar arguments, one can show that the conclusions hold. We omit the details. The proof is complete. □
Example. Consider the following equation
and its coupled system
with f ∈ C (ℝ, ℝ) being odd and
Here N = 2, and by Definition
and
Let α = 16, β = 1. Then it is easy to see that
By Theorem 1.1, Equation (3.5) has at least i(16, 2) - i(1, 2) = 2 geometrically different nonconstant 2-periodic solutions which satisfy x(t - 1) = -x(t).
References
Grafton R: A periodicity theorem for autonomous functional differential equations. J Diff Equ 1969, 6: 87–109. 10.1016/0022-0396(69)90119-3
Kaplan JL, Yorke JA: Ordinary differential equations which yield periodic solution of delay equations. J Math Anal Appl 1974, 48: 317–324. 10.1016/0022-247X(74)90162-0
Kaplan JL, Yorke JA: On the stability of a periodic solution of a differential delay equation. SIAM J Math Anal 1975, 6: 268–282. 10.1137/0506028
Kaplan JL, Yorke JA: On the nonlinear differential delay equation x '( t ) = - f ( x ( t ), x ( t - 1)). J Diff Equ 1977, 23: 294–314.
Nussbaum RD: Uniqueness and nonuniqueness for periodic solutions of x '( t ) = - g ( x ( t -1)). J Diff Equ 1979, 34: 25–54. 10.1016/0022-0396(79)90016-0
Li JB, He XZ: Multiple periodic solutions of differential delay equations created by asymptotically linear Hamiltonian systems. Nonlinear Anal TMA 1998, 31: 45–54. 10.1016/S0362-546X(96)00058-2
Li JB, He XZ: Proof and generalization of Kaplan-Yorke's conjecture on periodic solution of differential delay equations. Sci China (Ser A) 1999, 42(9):957–964. 10.1007/BF02880387
Li JB, He XZ, Liu ZR: Hamiltonian symmetric groups and multiple periodic solutions of differential delay equations. Nonlinear Anal 1999, 35: 457–474. 10.1016/S0362-546X(97)00623-8
Ge WG: Number of simple periodic solutions of differential difference equation on x '( t ) = - f ( x ( t - 1)). Chinese Ann Math 1993, 14A: 480–491.
Chen YS: The existence of periodic solutions of the equation x '( t ) = - f ( x ( t ), x ( t - r )). J Math Anal Appl 1992, 163: 227–237. 10.1016/0022-247X(92)90290-T
Chen YS: The existence of periodic solutions for a class of neutral differential difference equations. Bull Austral Math Soc 1992, 33: 508–516. 10.1017/S0334270000007190
Herz AVM: Solutions of x '( t ) = - g ( x ( t - 1)) approach the Kaplan-Yorke orbits for odd sigmoidg. J Diff Equ 1995, 118: 36–53. 10.1006/jdeq.1995.1066
Jekel S, Johnston C: A Hamiltonian with periodic orbits having several delays. J Diff Equ 2006, 222: 425–438. 10.1016/j.jde.2005.08.013
Fei GH: Multiple periodic solutions of differential delay equations via Hamiltonian systems (I). Nonlinear Anal 2006, 65: 25–39. 10.1016/j.na.2005.06.011
Fei GH: Multiple periodic solutions of differential delay equations via Hamiltonian systems (II). Nonlinear Anal 2006, 65: 40–58. 10.1016/j.na.2005.06.012
Guo ZM, Yu JS: Multiplicity results for periodic solutions to delay differential difference equations via critical point theory. J Diff Equ 2005, 218: 15–35. 10.1016/j.jde.2005.08.007
Mawhin J, Willem M: Critical point theory and Hamiltonian systems. Springer, New York; 1989.
Benci V: On critical point theory for indefinite functionals in the presence of symmetries. Trans Am Math Soc 1982, 274: 533–572. 10.1090/S0002-9947-1982-0675067-X
Chang KC: Infinite Dimensional Morse Theory and Multiple Solution Problems. Birkhaăser, Bostoi; 1993.
Fannio LO: Multiple periodic solution of Hamiltonian systems with strong resonance at infinity. Discrete Cont Dyn Syst 1997, 3: 251–264.
Cheng R, Hu JH: Variational Approaches for the Existence of Multiple Periodic Solutions of Differential Delay Equations. J Abstr Appl Anal 2010, 2010: 14. Article ID 978137 doi:10.1155/2010/978137
Guo CJ, Guo ZH: Existence of multiple periodic solutions for a class of second-order delay differential equations. Nonlinear Anal RWA 2009, 10: 3285–3297. 10.1016/j.nonrwa.2008.10.023
Cheng R, Zhang DF: Existence results of periodic solutions for non-autonomous differential delay equations with asymptotically linear properties. Nonlinear Anal RWA 2010, 11: 3406–3412. 10.1016/j.nonrwa.2009.11.031
Zhang XS, Meng Q: Nontrivial periodic solutions for delay differential systems via Morse theory. Nonlinear Anal 2011, 74: 1960–1968. 10.1016/j.na.2010.11.003
Long YM, Zehnder E: Morse theory for forced oscillations of asymptotically linear Hamiltonian systems. In Stochastic processes, Physics and Geometry. Edited by: Albeverio S. Proceedings of the Conference in Asconal/Locarno, Switzerland, World Scientific, Singapore; 1990:528–563.
Li SJ, Liu JQ: Morse theory and asymptotically linear Hamiltonian systems. J Diff Equ 1989, 78: 53–73. 10.1016/0022-0396(89)90075-2
Shu XB, Xu YT: Multiple periodic solutions to a class of second-order functional differential equations of mixed type. Acta Math Appl Sin 2006, 29(5):821–831.
Wang GQ, Yan JR: Existence of periodic solutions for second-order nonlinear neutral delay equations. Acta Math Sin 2004, 47(2):379–384.
Wang GQ, Cheng SS: Even periodic solutions of higher order duffing differential equations. J Czechoslovak Math 2007, 57(132):331–343.
Xu YT, Guo ZM: Applications of a Zp index theory to periodic solutions for a class of functional differential equations. J Math Anal Appl 2001, 257(1):189–205. 10.1006/jmaa.2000.7342
Xu YT, Guo ZM: Applications of a geometrical index theory to functional differential equations. Acta Math Sin 2001, 44(6):1027–1036.
Lu SP, Ge WG: Periodic solutions of the second-order differential equation with deviating Arguments. Acta Math Sin 2002, 45: 811–818.
Schechter M: Spectra of Partial Differential Equation. North-Holland, Amsterdam; 1971.
Acknowledgements
The authors thank the referees for valuable comments and suggestions which improved the presentation of this manuscript. This study was partially supported by NFSC (10871206, 10961011, 60964006, 11161013) Guangxi Natural Science Foundation (2011GXNSFA018134).
Author information
Authors and Affiliations
Corresponding author
Additional information
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors carried out the proof and authors conceived of the study. All authors read and approved the final manuscript.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 2.0 International License (https://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
About this article
Cite this article
Chen, H., Tang, H. & Sun, J. Periodic solutions of second-order differential equations with multiple delays. Adv Differ Equ 2012, 43 (2012). https://doi.org/10.1186/1687-1847-2012-43
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/1687-1847-2012-43
Keywords
- delay differential equations
- multiple periodic so1utiois
- critical point theory
- index theory