Research Article Homoclinic Orbits for a Class of Nonperiodic Hamiltonian Qi Wang

advertisement
Hindawi Publishing Corporation
Abstract and Applied Analysis
Volume 2013, Article ID 610906, 11 pages
http://dx.doi.org/10.1155/2013/610906
Research Article
Homoclinic Orbits for a Class of Nonperiodic Hamiltonian
Systems with Some Twisted Conditions
Qi Wang1 and Qingye Zhang2
1
2
Institute of Contemporary Mathematics, School of Mathematics and Information Science, Henan University, Kaifeng 475000, China
Department of Mathematics, Jiangxi Normal University, Nanchang 330022, China
Correspondence should be addressed to Qi Wang; mathwq@henu.edu.cn
Received 4 January 2013; Accepted 1 April 2013
Academic Editor: Changbum Chun
Copyright © 2013 Q. Wang and Q. Zhang. 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.
By the Maslov index theory, we will study the existence and multiplicity of homoclinic orbits for a class of asymptotically linear
nonperiodic Hamiltonian systems with some twisted conditions on the Hamiltonian functions.
1. Introduction and Main Results
Consider the following first-order nonautonomous Hamiltonian systems
𝑧̇ = 𝐽𝐻𝑧 (𝑑, 𝑧) ,
(HS)
where 𝑧 : R → R2𝑁, 𝐽 = ( 𝐼0𝑁 −𝐼0𝑁 ), 𝐻 ∈ 𝐢1 (R × R2𝑁, R) and
∇𝑧 𝐻(𝑑, 𝑧) denotes the gradient of 𝐻(𝑑, 𝑧) with respect to 𝑧. As
usual we say that a nonzero solution 𝑧(𝑑) of (HS) is homoclinic (to 0) if 𝑧(𝑑) → 0 as |𝑑| → ∞.
As a special case of dynamical systems, Hamiltonian systems are very important in the study of gas dynamics, fluid
mechanics, relativistic mechanics and nuclear physics. However it is well known that homoclinic solutions play an important role in analyzing the chaos of Hamiltonian systems. If a
system has the transversely intersected homoclinic solutions,
then it must be chaotic. If it has the smoothly connected
homoclinic solutions, then it cannot stand the perturbation,
and its perturbed system probably produces chaotic phenomena. Therefore, it is of practical importance and mathematical
significance to consider the existence of homoclinic solutions
of Hamiltonian systems emanating from 0.
In the last years, the existence and multiplicity of homoclinic orbits for the first-order system (HS) were studied
extensively by means of critical point theory, and many results
were obtained under the assumption that 𝐻(𝑑, 𝑧) depends
periodically on 𝑑 (see, e.g., [1–12]). Without assumptions of
periodicity, the problem is quite different in nature and there
is not much work done so far. To the best of our knowledge,
the authors in [13] firstly obtained the existence of homoclinic
orbits for a class of first-order systems without any periodicity
on the Hamiltonian function. After this, there were a few
papers dealing with the existence and multiplicity of homoclinic orbits for the first-order system (HS) in this situation
(see, e.g., [14–17]).
In the present paper, with the Maslov index theory of
homoclinic orbits introduced by Chen and Hu in [18], we will
study the existence and multiplicity of homoclinic orbits for
(HS) without any periodicity on the Hamiltonian function.
To the best of the author’s knowledge, the Maslov index
theory of homoclinic orbits is the first time to be used to
study the existence of homoclinic solutions. We are mainly
interested in the Hamiltonian functions of the form
𝐻 (𝑑, 𝑧) = −𝐿 (𝑑) 𝑧 ⋅ 𝑧 + 𝑅 (𝑑, 𝑧) ,
(1)
where 𝐿 is a 2𝑁 × 2𝑁 symmetric matrix valued function. We
2
assume that (𝐿 1 )𝐿 ∈ 𝐢(R, R𝑁 ), and there are 𝛼, 𝑐 > 0, 𝑑0 ≥ 0
and a constant matrix 𝑃, satisfying
𝑃𝐿 (𝑑) − 𝑐|𝑑|𝛼 𝐼2𝑁 ≥ 0,
∀ |𝑑| ≥ 𝑑0 ,
(2)
where 𝐼2𝑁 is the identity map on R2𝑁, and for a 2𝑁 × 2𝑁
matrix 𝑀, we say that 𝑀 ≥ 0 if and only if
inf
πœ‰∈R2𝑁 ,|πœ‰|=1
π‘€πœ‰ ⋅ πœ‰ ≥ 0.
(3)
2
Abstract and Applied Analysis
In (𝐿 1 ), if 𝑃 = ( 𝐼0𝑁 𝐼0𝑁 ), then (𝐿 1 ) is similar to the condition
(𝑅0 ) in [14]. But the restrictions on 𝑅(𝑑, 𝑧) will be different
from [14], and we will give some examples in Remark 5. If 𝑃 =
𝐼
0
±πΌ2𝑁 or 𝑃 = ( 𝑁+π‘š
0 −𝐼𝑁−π‘š ) in condition (𝐿 1 ), for example, it is
quite different from the existing results as authors known. In
short, condition (𝐿 1 ) means that the eigenvalues of 𝐿(𝑑) will
tend to ±∞ with the speed no less than |𝑑|𝛼 . But (𝐿 1 ) does not
contain all of these cases. For examples, let 𝑁 = 1 and 𝐿(𝑑) =
𝛼
2𝑑 sin 2𝑑
|𝑑|𝛼 ( cos
sin 2𝑑 − cos 2𝑑 ), we have the eigenvalues of 𝐿(𝑑) are ±|𝑑| ,
𝛼
but there is no constant matrix 𝑃 satisfying 𝑃𝐿(𝑑) − 𝑐|𝑑| 𝐼2𝑁 ≥
0, for all |𝑑| ≥ 𝑑0 .
Denote by 𝐹̃ the self-adjoint operator −J(𝑑/𝑑𝑑) + 𝐿(𝑑) on
Μƒ = 𝐻1 (R, R2𝑁) if 𝐿(𝑑) is
𝐿2 ≡ 𝐿2 (R, R2𝑁), with domain 𝐷(𝐹)
1
2𝑁
Μƒ ⊂ 𝐻 (R, R ) if 𝐿(𝑑) is unbounded. Let
bounded and 𝐷(𝐹)
Μƒ be the absolute value of 𝐹,
Μƒ and let |𝐹|
Μƒ 1/2 be the square root
|𝐹|
Μƒ 𝐷(𝐹)
Μƒ is a Hilbert space equipped with the norm
of |𝐹|.
󡄨 󡄨 σ΅„©
σ΅„©
‖𝑧‖𝐹̃ = σ΅„©σ΅„©σ΅„©σ΅„©(𝐼 + 󡄨󡄨󡄨󡄨𝐹̃󡄨󡄨󡄨󡄨) 𝑧󡄩󡄩󡄩󡄩𝐿2 ,
Μƒ .
∀𝑧 ∈ 𝐷 (𝐹)
(4)
Μƒ 1/2 ), and define on 𝐸 the inner product and
Let 𝐸 = 𝐷(|𝐹|
norm by
Μƒ 1/2 𝑒, |𝐹|
Μƒ 1/2 V) + (𝑒, V)2 ,
(𝑒, V)𝐸 = (|𝐹|
2
‖𝑒‖𝐸 =
(𝑒, 𝑒)1/2
𝐸 ,
(5)
where (⋅, ⋅)𝐿2 denotes the usual inner product on 𝐿2 (R, R2𝑁).
Then, 𝐸 is a Hilbert space. It is easy to see that 𝐸 is continuously embedded in 𝐻1/2 (R, R2𝑁), and we further have the
following lemma.
𝑧(𝑑) 󳨃→ 𝐡(𝑑)𝑧(𝑑), for any 𝑧 ∈ 𝐿2 , we still denote this operator
by 𝐡 and then 𝐾𝐡 : 𝐸 ⊂ 𝐿2 → 𝐸 is a self-adjoint compact
operator on 𝐸 and satisfies
(𝐾𝐡𝑒, V)𝐸 = (𝐡𝑒, V)𝐿2 ,
∀𝑒, V ∈ 𝐸.
(9)
Before presenting the conditions on 𝑅(𝑑, 𝑧), we need the
concept of Maslov index for homoclinic orbits introduced by
Chen and Hu in [18] which is equivalent to the relative Morse
index. We will give a brief introduction of it by Definition 7,
where for any 𝐡 ∈ B, we denote the associated index pair by
(πœ‡πΉ (𝐾𝐡), 𝜐𝐹 (𝐾𝐡)).
Now we can present the conditions on 𝑅(𝑑, 𝑧) as follows.
For notational simplicity, we set 𝐡0 (𝑑) = ∇𝑧2 𝑅(𝑑, 0), and in
what follows, the letter 𝑐 will be repeatedly used to denote
various positive constants whose exact value is irrelevant.
Besides, for two 2𝑁 × 2𝑁 symmetric matrices 𝑀1 and 𝑀2 ,
𝑀1 ≤ 𝑀2 means that 𝑀2 − 𝑀1 is semipositive definite.
(𝑅1 ) 𝑅 ∈ 𝐢2 (R × R2𝑁, R), and there exists a constant 𝑐 > 0
such that
󡄨󡄨 2
󡄨
󡄨󡄨∇𝑧 𝑅 (𝑑, 𝑧)󡄨󡄨󡄨 ≤ 𝑐, ∀ (𝑑, 𝑧) ∈ R × R2𝑁.
(10)
󡄨
󡄨
(𝑅0 ) ∇𝑧 𝑅(𝑑, 0) ≡ 0 and 𝐡0 ∈ B.
(𝑅∞ ) There exists some 𝑅0 > 0 and continuous symmetric
matrix functions 𝐡1 , 𝐡2 ∈ B with πœ‡πΉ (𝐾𝐡1 ) =
πœ‡πΉ (𝐾𝐡2 ) and 𝜐𝐹 (𝐾𝐡2 ) = 0 such that
𝐡1 (𝑑) ≤ ∇𝑧2 𝑅 (𝑑, 𝑧) ≤ 𝐡2 (𝑑) ,
∀𝑑 ∈ R, |𝑧| > 𝑅0 .
(11)
Then, we have our first result.
Lemma 1. Suppose that 𝐿 satisfies (𝐿 1 ). Then 𝐸 is compactly
embedded in 𝐿𝑝 (R, R2𝑁) with the usual norm β€– ⋅ ‖𝐿𝑝 for any
1 ≤ 𝑝 ∈ (2/(1 + 𝛼), ∞).
Theorem 2. Assume that (𝐿 1 ), (𝑅1 ), (𝑅0 ), and (𝑅∞ ) hold. If
This lemma is similar to Lemmas 2.1–2.3 in [13], and we
will prove it in Section 3. Define the quadratic form Q on 𝐸
by
then (HS) has at least one nontrivial homoclinic orbit. Moreover, if 𝜐𝐹 (𝐾𝐡0 ) = 0 and |πœ‡πΉ (𝐾𝐡1 ) − πœ‡πΉ (𝐾𝐡0 )| ≥ 𝑁, the problem possesses at least two nontrivial homoclinic orbits.
Q (𝑒, V) = ∫ ((−𝐽𝑒,Μ‡ V) + (𝐿 (𝑑) 𝑒, V)) 𝑑𝑑,
R
∀𝑒, V ∈ 𝐸.
(6)
It is easy to check that Q(𝑒, V) is a bounded quadratic form
on 𝐸, and, hence, there exists a unique bounded self-adjoint
operator 𝐹 : 𝐸 → 𝐸 such that
(𝐹𝑒, V)𝐸 = Q (𝑒, V) ,
∀𝑒, V ∈ 𝐸.
(7)
(𝐾𝑒, V)𝐸 = (𝑒, V)𝐿2 ,
∀𝑒 ∈ 𝐿 , V ∈ 𝐸.
(12)
Condition (𝑅∞ ) is a two-side pinching condition near the
±
) as follows.
infinity, and we can relax (𝑅∞ ) to condition (𝑅∞
±
) There exist some 𝑅0 > 0 and a continuous symmetric
(𝑅∞
matrix function 𝐡∞ ∈ B with 𝜐𝐹 (𝐾𝐡∞ ) = 0 such that
±∇𝑧2 𝑅 (𝑑, 𝑧) ≥ ±π΅∞ (𝑑) ,
∀𝑑 ∈ R, |𝑧| > 𝑅0 .
(13)
Then, we have the following results.
Besides, define a linear operator 𝐾 : 𝐿2 → 𝐸 by
2
πœ‡πΉ (𝐾𝐡1 ) ∉ [πœ‡πΉ (𝐾𝐡0 ) , πœ‡πΉ (𝐾𝐡0 ) + 𝜐𝐹 (𝐾𝐡0 )] ,
(8)
In view of Lemma 1, we know that 𝐹 is a Fredholm operator
and 𝐾 is a compact operator.
Denote by B the set of all uniformly bounded symmetric
2𝑁 × 2𝑁 matric functions. That is to say, 𝐡 ∈ B if and only
if 𝐡𝑇 (𝑑) = 𝐡(𝑑) for all 𝑑 ∈ R, and 𝐡(𝑑) is uniformly bounded
in 𝑑 as the operator on R2𝑁. For any 𝐡 ∈ B, it is easy to see
that 𝐡 determines a bounded self-adjoint operator on 𝐿2 , by
+
−
Theorem 3. Assumes (𝐿 1 ), (𝑅1 ), (𝑅0 ), (𝑅∞
) (or (𝑅∞
)), and
𝜐𝐹 (𝐾𝐡0 ) = 0 hold. If πœ‡πΉ (𝐾𝐡∞ ) ≥ πœ‡πΉ (𝐾𝐡0 ) + 2 (or πœ‡πΉ (𝐾𝐡∞ ) ≤
πœ‡πΉ (𝐾𝐡0 ) − 2), then (HS) has at least one nontrivial homoclinic
orbit.
+
−
) (or (𝑅∞
)),
Theorem 4. Suppose that (𝐿 1 ), (𝑅1 ), (𝑅0 ), (𝑅∞
and 𝜐𝐹 (𝐾𝐡0 ) = 0 are satisfied. If in addition 𝑅 is even in 𝑧 and
πœ‡πΉ (𝐾𝐡∞ ) ≥ πœ‡πΉ (𝐾𝐡0 ) + 2 (or πœ‡πΉ (𝐾𝐡∞ ) ≤ πœ‡πΉ (𝐾𝐡0 ) − 2), then
(HS) has at least |πœ‡πΉ (𝐾𝐡∞ ) − πœ‡πΉ (𝐾𝐡0 )| − 1 pairs of nontrivial
homoclinic orbits.
Abstract and Applied Analysis
3
Remark 5. Lemma 1 shows that 𝜎(𝐴), the spectrum of 𝐴,
consists of eigenvalues numbered by (counted in their multiplicities):
⋅ ⋅ ⋅ ≤ πœ† −2 ≤ πœ† −1 ≤ 0 < πœ† 1 ≤ πœ† 2 ≤ ⋅ ⋅ ⋅
(14)
with πœ† ±π‘˜ → ±∞ as π‘˜ → ∞. Let 𝐡0 (𝑑) ≡ 𝐡0 and 𝐡∞ (𝑑) ≡
𝐡∞ , with the constants 𝐡0 , 𝐡∞ satisfying πœ† 𝑙 < 𝐡0 < πœ† 𝑙+1 , and
πœ† 𝑙+𝑖 < 𝐡∞ < πœ† 𝑙+𝑖+1 for some 𝑙 ∈ Z and 𝑖 ≥ 1 (or 𝑖 ≤ −1).
Define
1
1
𝑅 (𝑑, 𝑧) = 𝛿 (|𝑧|) 𝐡0 |𝑧|2 + (1 − 𝛿 (|𝑧|)) 𝐡∞ |𝑧|2 ,
2
2
Definition 7. For any bounded self-adjoint Fredholm operator F and a compact self-adjoint operator T on H, the
relative Morse index pair (πœ‡F (T), 𝜐F (T)) is defined by
πœ‡F (T) = ind (𝑃F |H− (F−T) ) ,
(15)
where 𝛿 is a smooth cutoff function satisfying 𝛿(|𝑧|) =
|𝑧|<1,
{ 1,
0, |𝑧|>2. By Proposition 12 that, it is easy to verify 𝑅 satisfies
all the conditions in Theorem 2. Furthermore, let the constant
𝐡∞ satisfy πœ† 𝑙+𝑖 < 𝐡∞ < πœ† 𝑙+𝑖+1 for some 𝑙 ∈ Z and 𝑖 ≥ 2 (or
𝑖 ≤ −2). Define
1
1
𝑅 (𝑑, 𝑧) = 𝛿 (|𝑧|) 𝐡0 |𝑧|2 + (1 − 𝛿 (|𝑧|)) 𝐡∞ |𝑧|2 .
2
2
(see [19, Lemma 2.7]), where 𝑃F : H → H− (F) and 𝑃F−T :
H → H− (F − T) are the respective projections. Then, by
Fredholm operator theory, 𝑃F |H− (F−T) : H− (F − T) →
H− (F) is a Fredholm operator. Here and in the sequel, we
denote by ind(⋅) the Fredholm index of a Fredholm operator.
𝜐F (T) = dim H0 (F − T) .
(18)
2.2. Saddle Point Reduction. In this subsection, we describe
the saddle point reduction in [20–22]. Recall that H is a real
Hilbert space, and 𝐴 is a self-adjoint operator with domain
𝐷(𝐴) ⊂ H. Let Φ ∈ 𝐢1 (H, R), with ΦσΈ€  (πœƒ) = 0. Assume that
(1) there exist real numbers 𝛼 < 𝛽 such that 𝛼, 𝛽 ∉ 𝜎(𝐴),
and that 𝜎(𝐴)∩[𝛼, 𝛽] consists of at most finitely many
eigenvalues of finite multiplicities;
(16)
(2) ΦσΈ€  is Gateaux differentiable in H, which satisfies
Then 𝑅 satisfies all the conditions in Theorems 3 and 4. However, it is easy to see that some conditions of the main results
in [13–15, 17] do not hold for these examples.
σ΅„©σ΅„©
𝛼 + 𝛽 σ΅„©σ΅„©σ΅„©σ΅„© 𝛽 − 𝛼
σ΅„©σ΅„© σΈ€ 
𝐼󡄩 ≤
,
σ΅„©σ΅„©π‘‘Φ (𝑒) −
σ΅„©σ΅„©
2 σ΅„©σ΅„©σ΅„©
2
±
) is
Remark 6. Note that the assumption 𝜐𝐹 (𝐾𝐡∞ ) = 0 in (𝑅∞
+
) with
not essential for our main results. For the case of (𝑅∞
𝜐𝐹 (𝐾𝐡∞ ) =ΜΈ 0, let 𝐡̃∞ = 𝐡∞ − πœ€πΌ2𝑁 with πœ€ > 0 small enough,
where 𝐼2𝑁 is the identity map on R2𝑁; then, πœ‡πΉ (𝐾𝐡̃∞ ) =
+
πœ‡πΉ (𝐾𝐡∞ ) and 𝜐𝐹 (𝐾𝐡̃∞ ) = 0, and, hence, (𝑅∞
) holds for 𝐡̃∞ .
Therefore, Theorems 3 and 4 still hold in this case. While for
−
) with 𝜐𝐹 (𝐾𝐡∞ ) =ΜΈ 0, if we replace πœ‡πΉ (𝐾𝐡∞ ) by
the case of (𝑅∞
πœ‡πΉ (𝐾𝐡∞ )+𝜐𝐹 (𝐾𝐡∞ ) in Theorems 3 and 4, then similar results
hold. Indeed, let 𝐡̃∞ = 𝐡∞ + πœ€πΌ2𝑁 with πœ€ > 0 small enough
such that πœ‡πΉ (𝐾𝐡̃∞ ) = πœ‡πΉ (𝐾𝐡∞ ) + 𝜐𝐹 (𝐾𝐡∞ ) and 𝜐𝐹 (𝐾𝐡̃∞ ) =
−
0, then this case is also reduced to the case of (π‘Š∞
) for 𝐡̃∞
with 𝜐𝐹 (𝐾𝐡̃∞ ) = 0.
∀𝑒 ∈ H,
(19)
and without loss of generality, we may assume that 𝛼 =
−𝛽, 𝛽 > 0;
(3) Φ ∈ 𝐢2 (𝑉, R), 𝑉 = 𝐷(|𝐴|1/2 ), with the norm
1/2
σ΅„©
σ΅„©2
‖𝑧‖𝑉 = (σ΅„©σ΅„©σ΅„©σ΅„©|𝐴|1/2 𝑧󡄩󡄩󡄩󡄩H + πœ€2 ‖𝑧‖2H ) ,
(20)
where πœ€ > 0 small and −πœ€ ∉ 𝜎(𝐴).
Consider the solutions of the following equation:
𝐴𝑧 = ΦσΈ€  (𝑧) ,
𝑧 ∈ 𝐷 (𝐴) .
(21)
Let
2. Preliminaries
𝛽
In this section, we recall the definition of relative Morse index
and saddle point reduction and give the relationship between
them. For this propose, the notion of spectral flow will be
used.
2.1. Relative Morse Index. Let H be a separable Hilbert space;
for any self-adjoint operator 𝐴 on H, there is a unique 𝐴invariant orthogonal splitting
H = H+ (𝐴) ⊕ H− (𝐴) ⊕ H0 (𝐴) ,
𝑃0 = ∫ π‘‘πΈπœ† ,
𝑃+ = ∫
−𝛽
+∞
𝛽
𝑃− = ∫
−𝛽
−∞
π‘‘πΈπœ† ,
(22)
π‘‘πΈπœ† ,
where {πΈπœ† } is the spectral resolution of 𝐴, and let
H∗ = 𝑃∗ H,
∗ = 0, ±.
(23)
Decompose the space 𝑉 as follows:
(17)
where H0 (𝐴) is the null space of 𝐴, 𝐴 is positive definite
on H+ (𝐴) and negative definite on H− (𝐴), and 𝑃𝐴 denotes
the orthogonal projection from H to H− (𝐴). For any
bounded self-adjoint Fredholm operator F and a compact
self-adjoint operator T on H, 𝑃F − 𝑃F−T is compact
𝑉 = 𝑉0 ⊕ 𝑉− ⊕ 𝑉+ ,
(24)
where 𝑉∗ = |𝐴 πœ€ |−1/2 H∗ , ∗ = 0, ±, and 𝐴 πœ€ = 𝐴 + πœ€πΌ.
For each 𝑒 ∈ H, we have the decomposition
𝑒 = 𝑒+ + 𝑒0 + 𝑒− ,
(25)
4
Abstract and Applied Analysis
where 𝑒∗ ∈ H∗ and ∗ = 0, ±; let 𝑧 = 𝑧+ + 𝑧0 + 𝑧− , with
󡄨 󡄨−1/2
𝑧∗ = 󡄨󡄨󡄨𝐴 πœ€ 󡄨󡄨󡄨 𝑒∗ ,
∗ = 0, ±.
(26)
Define a functional 𝑓 on H as follows:
𝑓 (𝑒) =
1 σ΅„©σ΅„© σ΅„©σ΅„©2
σ΅„© σ΅„©2
σ΅„© σ΅„©2
(󡄩𝑒 σ΅„© + 𝑄+ 󡄩󡄩󡄩𝑒0 σ΅„©σ΅„©σ΅„© − 𝑄− 󡄩󡄩󡄩𝑒0 σ΅„©σ΅„©σ΅„© − ‖𝑒‖2 )
2 σ΅„© +σ΅„©
(27)
− Φπœ€ (𝑧) ,
∞
0
where 𝑄+ = ∫0 π‘‘πΈπœ† , 𝑄− = ∫−∞ π‘‘πΈπœ† , and Φπœ€ (𝑧) = (πœ€/2)
‖𝑧‖H + Φ(𝑧).
The Euler equation of this functional is the system
󡄨 󡄨−1/2
𝑒± = ±σ΅„¨σ΅„¨σ΅„¨π΄ πœ€ 󡄨󡄨󡄨 𝑃± ΦσΈ€ πœ€ (𝑧) ,
󡄨 󡄨−1/2
𝑄± 𝑒0 = ±σ΅„¨σ΅„¨σ΅„¨π΄ πœ€ 󡄨󡄨󡄨 𝑄± 𝑃0 ΦσΈ€ πœ€ (𝑧) .
(28)
(29)
Thus, 𝑧 = 𝑧+ + 𝑧0 + 𝑧− is a solution of (21) if and only if
𝑒 = 𝑒+ + 𝑒0 + 𝑒− is a critical point of 𝑓. The implicit function
can be applied, yielding a solution 𝑧± (𝑧0 ) for fixed 𝑧0 ∈ 𝑉0 ,
such that 𝑧± ∈ 𝐢1 (𝑉0 , 𝑉± ). Since dim 𝑉0 is finite, all topologies
on 𝑉0 are equivalent, and we choose β€– ⋅ β€–H as it norm. We have
󡄨 󡄨1/2
𝑒± (𝑧0 ) = 󡄨󡄨󡄨𝐴 πœ€ 󡄨󡄨󡄨 𝑧± (𝑧0 ) ∈ 𝐢1 (H0 , H) ,
1
(𝐴 (𝑧 (π‘₯) , 𝑧 (π‘₯))) − Φ (𝑧 (π‘₯)) ,
2
(32)
(33)
between the critical points of the 𝐢2 -function π‘Ž ∈ 𝐢2 (H0 , R)
with the solutions of the operator equation
𝑧 ∈ 𝐷 (𝐴) .
(34)
Moreover, the functional π‘Ž satisfies
π‘ŽσΈ€ σΈ€  (π‘₯) = [𝐴 − ΦσΈ€ σΈ€  (𝑧 (π‘₯))] 𝑧󸀠 (π‘₯)
= 𝐴𝑃0 − 𝑃0 ΦσΈ€ σΈ€  (𝑧 (π‘₯)) 𝑧󸀠 (π‘₯) .
Proof. Note that
≤
1 σ΅„©σ΅„© σΈ€  +
σ΅„©
σ΅„©σ΅„©Φ (𝑧 + 𝑧− + π‘₯) + πœ€(𝑧+ + 𝑧− + π‘₯)σ΅„©σ΅„©σ΅„© 2
󡄩𝐿
σ΅„©
√𝛽 − πœ€
≤
𝐢𝑅 + πœ€ σ΅„©σ΅„© +
σ΅„©
−
󡄩𝑧 + 𝑧 + π‘₯󡄩󡄩󡄩𝐿2
√𝛽 − πœ€ σ΅„©
≤
σ΅„© + σ΅„©
σ΅„© − σ΅„©
𝐢𝑅 + πœ€ 󡄩󡄩󡄩𝑒 (π‘₯)󡄩󡄩󡄩𝐿2 󡄩󡄩󡄩𝑒 (π‘₯)󡄩󡄩󡄩𝐿2
(
+
+ β€–π‘₯‖𝐿2 ) .
√𝛽 − πœ€
√𝛽
√𝛽
(38)
(39)
Therefore,
2√𝛽 (𝐢𝑅 + πœ€)
σ΅„© − σ΅„©
σ΅„©σ΅„© + σ΅„©σ΅„©
β€–π‘₯‖𝐿2 .
󡄩󡄩𝑒 (π‘₯)󡄩󡄩𝐿2 + 󡄩󡄩󡄩𝑒 (π‘₯)󡄩󡄩󡄩𝐿2 ≤
𝛽 − 2𝐢𝑅 − 3πœ€
(40)
Next, since
σΈ€ 
󡄨 󡄨−1/2
(𝑒± ) (π‘₯) = ±σ΅„¨σ΅„¨σ΅„¨π΄ πœ€ 󡄨󡄨󡄨 𝑃± ΦσΈ€ πœ€
σΈ€ 
π‘ŽσΈ€  (π‘₯) = 𝐴 (𝑧 (π‘₯)) − ΦσΈ€  (𝑧 (π‘₯))
= 𝐴π‘₯ − 𝑃0 Φ (𝑧 (π‘₯)) ,
(37)
σΈ€ 
× (𝑧+ + 𝑧− + π‘₯) ((𝑧+ ) (π‘₯) + (𝑧− ) (π‘₯) + 𝐼) ,
(41)
1
π‘Ž (π‘₯) = (𝐴 (𝑧 (π‘₯) , 𝑧 (π‘₯))) − Φ (𝑧 (π‘₯)) ,
2
σΈ€ 
𝛽 󳨀→ ∞.
(36)
(31)
Theorem 8. Under assumptions (1), (2), and (3), there is a oneone correspondence
𝐴𝑧 = ΦσΈ€  (𝑧) ,
Moreover, one has
σ΅„©σ΅„© ± σΈ€  σ΅„©σ΅„©
σ΅„©σ΅„©(𝑒 ) (π‘₯)σ΅„©σ΅„© 󳨀→ 0,
σ΅„©σ΅„©
󡄩󡄩𝐿2
∀π‘₯ ∈ H0 .
From ∇𝑧 𝑅(𝑑, 0) = 0, |∇𝑧2 𝑅| ≤ 𝐢𝑅 , we have ΦσΈ€  (0) = 0 and
β€–ΦσΈ€  (𝑧)‖𝐿2 ≤ 𝐢𝑅 ‖𝑧‖𝐿2 . Since β€–|𝐴 πœ€ |−1/2 𝑃± β€– ≤ 1/√(𝛽 − πœ€), we
have
σ΅„©σ΅„© ± σ΅„©σ΅„©
󡄩󡄩𝑒 (π‘₯)󡄩󡄩𝐿2
where 𝑧(π‘₯) = πœ‰(π‘₯)+π‘₯, πœ‰(π‘₯) = 𝑧+ (π‘₯)+𝑧− (π‘₯) ∈ 𝐷(𝐴). Then, we
have the following theorem due to Amann [20], Chang [21],
and Long [22].
π‘₯ 󳨃󳨀→ 𝑧 = 𝑧 (π‘₯) = 𝑧+ (π‘₯) + 𝑧− (π‘₯) + π‘₯
2√𝛽 (𝐢𝑅 + 1)
σ΅„©σ΅„© ± σ΅„©σ΅„©
β€–π‘₯‖𝐿2 ,
󡄩󡄩𝑒 (π‘₯)󡄩󡄩𝐿2 ≤
𝛽 − 2𝐢𝑅 − 3πœ€
󡄨 󡄨−1/2
𝑒± (π‘₯) = ±σ΅„¨σ΅„¨σ΅„¨π΄ πœ€ 󡄨󡄨󡄨 𝑃± ΦσΈ€ πœ€ (𝑧+ + 𝑧− + π‘₯) .
where 𝑒0 (𝑧0 ) = |𝐴 πœ€ |1/2 𝑧0 , and let 𝑧0 = π‘₯ we have
π‘Ž (π‘₯) =
Lemma 9. Assume that 𝑅 ∈ 𝐢2 (R × R2𝑁, R), |∇𝑧2 𝑅| ≤ 𝐢𝑅 , for
all (𝑑, 𝑧) ∈ R × R2𝑁 and ∇𝑧 𝑅(𝑑, 0) ≡ 0; then one has
(30)
which solves the system (28).
Let
π‘Ž (𝑧0 ) = 𝑓 (𝑒+ (𝑧0 ) + 𝑒− (𝑧0 ) + 𝑒0 (𝑧0 )) ,
Since H0 is a finite dimensional space, for every critical
point π‘₯ of π‘Ž in H0 , the Morse index and nullity are finite, and
we denote them by (π‘šπ‘Ž− (π‘₯), π‘šπ‘Ž0 (π‘₯)).
Now, let the Hilbert space H be 𝐿2 (R, R2𝑁), and the oper∞
ator 𝐴 be 𝐹̃ = −𝐽(𝑑/𝑑𝑑)+𝐿, Φ(𝑒) = ∫−∞ 𝑅(𝑑, 𝑒). Then we have
𝑉 = 𝐸. For 𝑅 ∈ 𝐢2 (R×R2𝑁, R) and |∇𝑧2 𝑅| ≤ 𝐢𝑅 , for all (𝑑, 𝑧) ∈
Μƒ we have 𝐴 and
R × R2𝑁, let −𝛼 = 𝛽 ≥ 2(𝐢𝑅 + 1) and 𝛽 ∉ 𝜎(𝐹);
Φ satisfying the previous conditions. Thus, from Theorem 8,
we can solve our problems on the finite dimensional space.
Similar to Lemma 2.2 and Remark 2.3 in [23], we have the
following estimates.
(35)
where 𝐼 is the identity map on H0 , we have
σ΅„©
σ΅„©σ΅„© + σΈ€  σ΅„©σ΅„©
σ΅„©
σ΅„©σ΅„©(𝑒 ) (π‘₯)σ΅„©σ΅„© + σ΅„©σ΅„©σ΅„©(𝑒− )σΈ€  (π‘₯)σ΅„©σ΅„©σ΅„©
󡄩󡄩𝐿2 σ΅„©σ΅„©
󡄩󡄩𝐿2
σ΅„©σ΅„©
≤
2√𝛽 (𝐢𝑅 + πœ€)
,
𝛽 − 2𝐢𝑅 − 3πœ€
(42)
∀π‘₯ ∈ H0 .
Abstract and Applied Analysis
5
Remark 10. For 𝑧(π‘₯), we also have that there is a constant 𝐢 >
0 dependent of 𝐢𝑅 , but independent of 𝛽, such that
𝐢
σ΅„©σ΅„© ± σ΅„©σ΅„©
β€–π‘₯‖𝐿2 ,
󡄩󡄩𝑧 (π‘₯)󡄩󡄩𝑉 ≤
√𝛽
𝐢
σ΅„©σ΅„© σΈ€ ± σ΅„©σ΅„©
󡄩󡄩𝑧 (π‘₯)σ΅„©σ΅„© ≤
󡄩𝑉 √𝛽 ,
σ΅„©
∀π‘₯ ∈ H0 .
(43)
If 𝑅 satisfies the condition (𝑅1 ), then for any homoclinic
orbit 𝑧 of (HS), ∇𝑧2 𝑅(⋅, 𝑧) ∈ B, and, hence, we have the associated index pair (πœ‡πΉ (𝐾𝐡), 𝜐𝐹 (𝐾𝐡)). For notation simplicity,
in what follows, we set
πœ‡πΉ (𝑧) = πœ‡πΉ (𝐾∇𝑧2 (𝑅 (𝑑, 𝑧))) ,
(44)
𝜐𝐹 (𝑧) = 𝜐𝐹 (𝐾∇𝑧2 (𝑅 (𝑑, 𝑧))) .
Theorem 11. Let 𝑅 ∈ 𝐢2 (R × R2𝑁, R) satisfying |∇𝑧2 𝑅| ≤ 𝐢𝑅 ,
for all (𝑑, 𝑧) ∈ R × R2𝑁 and ∇𝑧 𝑅(𝑑, 0) ≡ 0. For each critical
point π‘₯ of π‘Ž in H0 , 𝑧(π‘₯) is a homoclinic orbit of (HS) and one
has
π‘šπ‘Ž− (π‘₯) = dim (𝐸− (H0 )) + πœ‡πΉ (𝐾∇𝑧2 𝑅 (𝑑, 𝑧 (π‘₯)))
= dim (𝐸 (H0 )) + πœ‡πΉ (𝑧 (π‘₯)) ,
(π‘₯) =
𝜐𝐹 (𝐾∇𝑧2 𝑅 (𝑑, 𝑧 (π‘₯))
(45)
= 𝜐𝐹 (𝑧 (π‘₯)) ,
0
where dim(𝐸− (H0 )) is the dimension of the space ∫−𝛽 π‘‘πΈπœ†
(H0 ).
This theorem shows the relations between the relative
Morse index and the Morse index of the saddle point reduction, and it will play an important role in the proof of our
main results. The proof of this theorem will be postponed to
the next subsection where the notion of spectral flow will be
used.
2.3. The Relationship between πœ‡πΉ (𝑇), Spectral Flow, and the
Morse Index of Saddle Point Reduction. It is well known that
the concept of spectral flow was first introduced by Atiyah
et al. in [24] and then extensively studied in [19, 25–28].
Here, we give a brief introduction of the spectral flow as
introduced in [18]. Let H be a separable Hilbert space as
defined before, and {Fπœƒ | πœƒ ∈ [0, 1]} be a continuous path
of self-adjoint Fredholm operators on the Hilbert space H.
The spectral flow of Fπœƒ represents the net change in the
number of negative eigenvalues of Fπœƒ as πœƒ runs from 0
to 1, where the counting follows from the rule that each
negative eigenvalue crossing to the positive axis contributes
+1 and each positive eigenvalue crossing to the negative axis
contributes −1, and for each crossing, the multiplicity of
eigenvalue is taken into account. In the calculation of spectral
flow, a crossing operator introduced in [28] will be used. Take
a 𝐢1 path {Fπœƒ πœƒ |∈ [0, 1]} and let Pπœƒ be the projection from
H to H0 (Fπœƒ ). When eigenvalue crossing occurs at Fπœƒ , the
operator
πœ•
F P : H0 (Fπœƒ ) 󳨀→ H0 (Fπœƒ )
πœ•πœƒ πœƒ πœƒ
A crossing occurring at Fπœƒ is called simple crossing if
dim H0 (Fπœƒ ) = 1.
As indicated in [19], the spectral flow Sf(Fπœƒ ) will remain
the same after a small disturbance of Fπœƒ , that is, Sf(Fπœƒ ) =
Sf(Fπœƒ + πœ€id) for πœ€ > 0 and small enough, where id is the
identity map on H. Furthermore, we can choose suitable πœ€
such that all the eigenvalue crossings occurring in Fπœƒ , 0 ≤
πœƒ ≤ 1 are regular [28]. Thus, without loss of generality, we
may assume that all the crossings are regular. Let D be the set
containing all the points in [0, 1] at which the crossing occurs.
The set D contains only finitely many points. The spectral
flow of Fπœƒ is
= ∑ sign πΆπ‘Ÿ [Fπœƒ ] − dim H− (πΆπ‘Ÿ [F0 ])
πœƒ∈D∗
−
Pπœƒ
sign πΆπ‘Ÿ [Fπœƒ ] = dim H+ (πΆπ‘Ÿ [Fπœƒ ]) − dim H− (πΆπ‘Ÿ [Fπœƒ ]) .
(47)
Sf (Fπœƒ , 0 ≤ πœƒ ≤ 1)
−
π‘šπ‘Ž0
is called a crossing operator, denoted by πΆπ‘Ÿ [Fπœƒ ]. As mentioned in [28], an eigenvalue crossing at Fπœƒ is said to be
regular if the null space of πΆπ‘Ÿ [Fπœƒ ] is trivial. In this case, we
define
(46)
(48)
+ dim H+ (πΆπ‘Ÿ [F1 ]) ,
where D∗ = D ∩ (0, 1). In what follows, the spectral flow of
Fπœƒ will be simply denoted by Sf(Fπœƒ ) when the starting and
end points of the flow are clear from the contents. And 𝑃Fπœƒ
will be simply denoted by π‘ƒπœƒ .
Proposition 12 (see [18, Proposition 3]). Suppose that, for
each πœƒ ∈ [0, 1], Fπœƒ − F0 is a compact operator on H then
ind (𝑃0 |H− (F1 ) ) = −Sf (Fπœƒ ) .
(49)
Thus, from Definition 7,
πœ‡F0 (T) = −Sf (Fπœƒ , 0 ≤ πœƒ ≤ 1) ,
(50)
where Fπœƒ = F − πœƒT, T is a compact operator. Moreover, if
𝜎(T) ⊂ [0, ∞) and 0 ∉ πœŽπ‘ƒ (T), from the definition of spectral
flow, we have
πœ‡F0 (T) = −Sf (Fπœƒ , 0 ≤ πœƒ ≤ 1)
= ∑ 𝜐F (πœƒT)
πœƒ∈[0,1)
(51)
= ∑ dim H0 (F − πœƒT) .
πœƒ∈[0,1)
The proof of Theorem 11 is the direct consequence of the
aformentioned Proposition 12 and Theorem 3.2 in [19], so we
omit it here.
−
can be transformed into the case
Remark 13. The case of 𝑅∞
+
−
+
case follows from the 𝑅∞
of 𝑅∞ . More concretely, the 𝑅∞
Μƒ 𝑧) = −𝑅(−𝑑, 𝑧). If 𝑧(𝑑)
case by applying to the function 𝑅(𝑑,
6
Abstract and Applied Analysis
Μƒ
is a homoclinic solution of 𝐹𝑧(𝑑)
= ∇𝑧 𝑅(𝑑, 𝑧(𝑑)), let 𝑧̃(𝑑) =
𝑧(−𝑑), it is easy to check that 𝑧̃(𝑑) is a homoclinic solution of
Μƒ 𝑧̃(𝑑)), and this is a one-one correspondence
̃𝑧(𝑑) = ∇𝑧 𝑅(𝑑,
𝐹̃
between the two systems. By the definition of spectral flow
and it is catenation property [19], we have πœ‡πΉ (−𝐡∞ (−𝑑)) −
πœ‡πΉ (−𝐡0 (−𝑑)) = πœ‡πΉ (𝐡0 (𝑑)) − πœ‡πΉ (𝐡∞ (𝑑)). Thus, we only consider
+
from now on.
the case of 𝑅∞
From (55) and (57), we have ‖𝑧 − 𝑧𝑖 ‖𝐿2 < πœ€; thus, 𝐾 has a finite
πœ€-net in 𝐿2 , and so the embedding 𝐸 󳨅→ 𝐿2 is compact.
Step 2 (the case of 𝑝 > 2). Since 𝐸 is continuously embedded
in 𝐻1/2 , hence, by the Sobolev embedding theorem, 𝐸 is
continuously embedded in 𝐿𝑝 , for all 𝑝 > 2. For any 𝑝 > 2,
by the Hölder inequality we have
𝑝−1
Proof of Lemma 1. Recall the operator 𝐹̃ = −J(𝑑/𝑑𝑑) + 𝐿(𝑑),
Μƒ = 𝐻1 (R, R2𝑁) if 𝐿(𝑑) is bounded and
with domain 𝐷(𝐹)
1
Μƒ is a Hilbert
Μƒ
𝐷(𝐹) ⊂ 𝐻 (R, R2𝑁) if 𝐿(𝑑) is unbounded. 𝐷(𝐹)
Μƒ
space equipped with the norm ‖𝑧‖𝐹̃ = β€–(𝐼 + |𝐹|)𝑧‖𝐿2 , for all
Μƒ Recall the Hilbert space 𝐸 = 𝐷(|𝐹|
Μƒ 1/2 ), with the
𝑧 ∈ 𝐷(𝐹).
inner product and norm by
󡄨 󡄨1/2 󡄨 󡄨1/2
(𝑒, V)𝐸 = (󡄨󡄨󡄨󡄨𝐹̃󡄨󡄨󡄨󡄨 𝑒, 󡄨󡄨󡄨󡄨𝐹̃󡄨󡄨󡄨󡄨 V) + (𝑒, V)2 ,
2
‖𝑒‖𝐸 =
Step 1 (the case of 𝑝 = 2). For 𝑅 > 0, from (𝐿 1 ) and (53) we
have
|𝑑|>𝑅
|𝑧|2 𝑑𝑑 ≤ 𝑐|𝑅|𝛼−2 ∫
|𝑑|>𝑅
⟨𝐿 (𝑑) 𝑧, 𝑃𝑇 π‘§βŸ©R2𝑁 𝑑𝑑
(54)
≤ 𝑐|𝑅|𝛼−2 ‖𝑧‖2𝐸 .
For any πœ€ > 0, from (54), we can choose 𝑅0 large enough, such
that
∫
|𝑑|>𝑅0
|𝑧|2 𝑑𝑑 <
πœ€2
,
4
∀𝑧 ∈ 𝐾.
(55)
On the other hand, by the definition of β€– ⋅ ‖𝐸 , we have
∫
2
|𝑑|≤𝑅0
Step 3 (the case of 1 ≤ 𝑝 ∈ (2/(1 + 𝛼), 2)). First, we have
(𝛼/(2 − 𝑝)) ⋅ 𝑝 > 1; so we can choose 𝛼𝑝 satisfying 𝛼𝑝 ∈ (0, 𝛼)
and (𝛼𝑝 /(2 − 𝑝) ⋅ 𝑝) > 1. Denote that π‘Ÿ = 𝛼𝑝 /(2 − 𝑝). For 𝑅 > 0
and 𝑧 ∈ 𝐸, denote that 𝐸𝑅1 (𝑧) = {𝑑; |𝑑| ≥ 𝑅 and |𝑑|π‘Ÿ |𝑧(𝑑)| > 1}
and 𝐸𝑅2 (𝑧) = {𝑑; |𝑑| ≥ 𝑅 and |𝑑|π‘Ÿ |𝑧(𝑑)| ≤ 1}. Then, from (53),
∫
𝐸𝑅1 (𝑧)
𝑐‖𝑧‖2𝐸 .
|𝑧| 𝑑𝑑 ≤ ‖𝑧‖𝐸 ≤ 𝐢,
∀𝑧 ∈ 𝐾.
(56)
Thus, by the Sobolev compact embedding theorem, there
exist 𝑧1 , 𝑧2 , . . . , π‘§π‘š ∈ 𝐾, such that for any 𝑧 ∈ 𝐾, there is 𝑧𝑖
satisfying
πœ€2
󡄩𝑝
σ΅„©σ΅„©
󡄩󡄩𝑧 − 𝑧𝑖 󡄩󡄩󡄩𝐿𝑝 ((−𝑅 ,𝑅 ),R2𝑁 ) < .
0 0
2
(57)
(58)
thus, the embedding 𝐸 󳨅→ 𝐿𝑝 is compact, for all 𝑝 > 2.
(52)
Let 𝐾 ⊂ 𝐸 be a bounded set. We will show that 𝐾 is precompact in 𝐿𝑝 for 1 ≤ 𝑝 ∈ (2/(1 + 𝛼), ∞). We divide the
proof into three steps.
∫
R
(𝑒, 𝑒)1/2
𝐸 ,
where (⋅, ⋅)𝐿2 denotes the usual inner product on 𝐿2 (R, R2𝑁).
From (𝐿 1 ), there is a matrix 𝐿 0 such that 𝑃(𝐿(𝑑) − 𝐿 0 ) ≥ 0 for
Μƒ 𝐹̃0 +𝐿 0 , with 𝐹̃0 = −J(𝑑/𝑑𝑑)+𝐿 0
all 𝑑 ∈ R. We have 𝐿(𝑑) = 𝐹−
1
and 𝐷(𝐹̃0 ) = 𝐻 . Thus, for any 𝑧 ∈ 𝐸,
󡄨󡄨󡄨(𝐿 (𝑑) 𝑧, 𝑃𝑇 𝑧) 󡄨󡄨󡄨
󡄨󡄨
𝐿2 󡄨󡄨
󡄨󡄨
󡄨 󡄨
󡄨 󡄨
󡄨
Μƒ 𝑃𝑇 𝑧) 󡄨󡄨󡄨 + 󡄨󡄨󡄨(𝐹̃0 𝑧, 𝑃𝑇 𝑧) 󡄨󡄨󡄨 + 󡄨󡄨󡄨(𝐿 0 𝑧, 𝑃𝑇 𝑧) 󡄨󡄨󡄨 (53)
≤ 󡄨󡄨󡄨(𝐹𝑧,
󡄨
󡄨
󡄨
󡄨
𝐿2 󡄨 󡄨
𝐿2 󡄨 󡄨
𝐿 2 󡄨󡄨
󡄨
≤
𝑝−1
∫ |𝑧|𝑝 𝑑𝑑 ≤ ‖𝑧‖𝐿2 ‖𝑧‖𝐿2(𝑝−1) ≤ 𝐢‖𝑧‖𝐿2 ‖𝑧‖𝐸 ,
3. Proof of Our Main Results
𝑝
|𝑧|𝑝 𝑑𝑑 = ∫
𝐸𝑅1 (𝑧)
≤∫
𝐸𝑅1 (𝑧)
≤
≤
𝑐
(|𝑑|π‘Ÿ |𝑧|) |𝑑|−π‘Ÿπ‘ 𝑑𝑑
|𝑧|2 |𝑑|𝛼𝑝 𝑑𝑑
|𝑅|𝛼−𝛼𝑝
󡄨󡄨
󡄨
󡄨󡄨(𝐿 (𝑑) 𝑧, 𝑃𝑇 𝑧) 2 󡄨󡄨󡄨
󡄨
𝐿 󡄨
(59)
𝑐
‖𝑧‖2𝐸 ,
|𝑅|𝛼−𝛼𝑝
and so
∫
|𝑑|≥𝑅
|𝑧|𝑝 𝑑𝑑
=∫
𝐸𝑅1 (𝑧)
≤
|𝑧|𝑝 𝑑𝑑 + ∫
𝐸𝑅2 (𝑧)
𝑐
𝛼−𝛼𝑝
|𝑅|
‖𝑧‖2𝐸 +
|𝑧|𝑝 𝑑𝑑
(60)
2
,
(π‘Ÿπ‘ − 1) π‘…π‘Ÿπ‘−1
∀𝑧 ∈ 𝐸.
Let 𝐾 ⊂ 𝐸 be a bounded set. For any πœ€ > 0, from (60), choose
𝑅0 > 0 large enough, such that
∫
|𝑑|≥𝑅0
|𝑧|𝑝 𝑑𝑑 <
πœ€π‘
,
4
∀𝑧 ∈ 𝐾.
(61)
On the other hand, by the Sobolev compact embedding theorem, there are 𝑧1 , 𝑧2 , . . . , π‘§π‘š ∈ 𝐾, such that for any 𝑧 ∈ 𝐾,
there exists 𝑧𝑖 satisfying
πœ€π‘
σ΅„©σ΅„©
󡄩𝑝
󡄩󡄩𝑧 − 𝑧𝑖 󡄩󡄩󡄩𝐿𝑝 ((−𝑅 ,𝑅 ),R2𝑁 ) < .
0 0
2
(62)
From (61) and (62), we have
σ΅„©
σ΅„©σ΅„©
󡄩󡄩𝑧 − 𝑧𝑖 󡄩󡄩󡄩𝐿𝑝 < πœ€;
(63)
that is to say, 𝐾 has a finite πœ€-net in 𝐿𝑝 , and the embedding
𝐸 󳨅→ 𝐿𝑝 is compact. The proof of the lemma is complete.
Abstract and Applied Analysis
7
Consider the homoclinic orbits of the linear Hamiltonian
systems
𝑧̇ (𝑑) = 𝐽𝐡 (𝑑) 𝑧 (𝑑) ,
𝑧 (𝑑) 󳨀→ 0,
∀𝑑 ∈ R,
|𝑑| 󳨀→ ∞,
the saddle point reduction, choose a suitable number 𝛽,
which is used in the projection for the saddle point reduction
in Section 2.2. Let
𝛽
(64)
𝑃 = ∫ π‘‘πΈπœ† ,
−𝛽
2
( 𝐼0𝑁 −𝐼0𝑁 ),
and 𝐡(𝑑) is a
where 𝑧(𝑑) : R → R2𝑁, 𝐽 =
continuous symmetric matrix function. Denote by 𝑆 the set
of homoclinic orbits of linear systems (64); then, 𝑆 is a linear
subspace of 𝐿2 (R, R2𝑁) and we have the following lemma.
Lemma 14. The dimension of the solution space 𝑆 will be less
than or equal to 𝑁. Thus for any homoclinic orbit 𝑧(𝑑) of (HS),
if 𝑅 satisfies (𝑅1 ), one has
0 ≤ 𝜐𝐹 (𝑧) ≤ 𝑁.
(65)
Proof. As usual, we define the symplectic groups on R2𝑁 by
𝑆𝑝 (2𝑁) = {𝑀 ∈ L (R2𝑁) , | 𝑀𝑇 𝐽𝑀 = 𝐽} ,
(66)
where L(R2𝑁) is the set of all 2𝑁 × 2𝑁 real matrices, and
𝑀𝑇 denotes the transpose of 𝑀. Let π‘Š(𝑑) be the fundamental
solution of (64); then, π‘Š(𝑑) is a path in 𝑆𝑝(2𝑁). Let 𝑧(𝑑) be a
nontrivial homoclinic orbit of (64); that is to say, 𝑧(0) =ΜΈ 0 and
satisfies
𝑧 (𝑑) = π‘Š (𝑑) 𝑧 (0) ,
lim π‘Š (𝑑) 𝑧 (0) = 0.
(67)
𝑑→∞
𝑑→∞
(68)
lim π‘Š (𝑑) 𝑧0 = 0,
lim π‘Š (𝑑) 𝐽𝑧0 = 0.
(69)
𝑑→∞
Since π‘Š(𝑑) is a path in 𝑆𝑝(2𝑁), π‘Šπ‘‡ (𝑑)π½π‘Š(𝑑) = 𝐽, for all 𝑑 ∈ R,
thus
0 = lim (π½π‘Š (𝑑) 𝑧0 , π‘Š (𝑑) 𝐽𝑧0 )R2𝑁
𝑑→∞
= − lim (𝑧0 , π‘Šπ‘‡ (𝑑) π½π‘Š (𝑑) 𝐽𝑧0 )R2𝑁
𝑑→∞
By Theorem 8, we have a functional π‘Ž(π‘₯) with π‘₯ ∈ 𝑋, whose
critical points give rise to solutions of (HS).
Lemma 15. (1) π‘Ž satisfies (PS) condition;
(2) π»π‘ž (𝑋, π‘Ž; R) ≅ π›Ώπ‘ž,π‘Ÿ R, π‘ž = 0, 1, . . . for −π‘Ž ∈ R large
enough, where π‘Ÿ = dim(𝐸− (𝑋)) + πœ‡πΉ (𝐾𝐡1 ).
Proof. Assume that there is a sequence {π‘₯𝑛 } ⊂ 𝑋, satisfying
π‘ŽσΈ€  (π‘₯𝑛 ) → 0(𝑛 → ∞). That is,
󡄩󡄩󡄩𝐹𝑧𝑛 − 𝐾∇𝑧 𝑅 (𝑑, 𝑧𝑛 )σ΅„©σ΅„©σ΅„© 󳨀→ 0,
(72)
σ΅„©
󡄩𝐸
where 𝑧𝑛 = 𝑧(π‘₯𝑛 ) defined in Section 2.1. Since 𝑋 is a finite
dimensional space, and from the definition of 𝑧𝑛 , it is enough
to prove that {𝑧𝑛 } is bounded in 𝐸. For each πœ€ ∈ (0, 1), define
𝐢𝑛 ∈ B by
1
2
{
{
{∫0 ∇𝑧 𝑅 (𝑑, 𝑠𝑧𝑛 ) 𝑑𝑠,
𝐢𝑛 (𝑑) = {
{
{
𝐡 (𝑑) ,
{ 1
󡄨 𝑅
󡄨󡄨
󡄨󡄨𝑧𝑛 (𝑑)󡄨󡄨󡄨 ≥ 0 ,
πœ€
𝑅
󡄨
󡄨󡄨
󡄨󡄨𝑧𝑛 (𝑑)󡄨󡄨󡄨 < 0 .
πœ€
(70)
= (𝑧0 , 𝑧0 )R2𝑁 ,
which contradicts 𝑧0 =ΜΈ 0. Since 𝐽 is an isomorphism on R2𝑁,
we have dim 𝑆0 ≤ 𝑁. And from the definition of 𝜐𝐹 (𝑧) in the
last part of Section 2.2, we have completed the proof.
Before the proof of Theorem 2, we need the following
lemma. Since 𝑅 satisfies condition (𝑅1 ), performing on (HS)
(73)
It is easy to verify that {𝐢𝑛 } satisfies
≤ 𝐢𝑛 (𝑑) ≤ 𝐡2 (𝑑) + πœ€ (𝑐 ⋅ 𝐼 − 𝐡2 (𝑑)) ,
Then, we have dim 𝑆 = dim(𝑆0 ). We claim that 𝐽𝑧0 ∉ 𝑆0 if 𝑧0 ∈
𝑆0 and 𝑧0 =ΜΈ 0. We prove it indirectly. Assume that 𝑧0 , 𝐽𝑧0 ∈ 𝑆0
with 𝑧0 =ΜΈ 0; that is to say,
𝑑→∞
𝑋 = 𝑃𝐿 (R, R ) .
𝐡1 (𝑑) − πœ€ (𝐡1 (𝑑) + 𝑐 ⋅ 𝐼)
Denote by 𝑆0 the subset of R2𝑁 satisfying
𝑆0 = {𝑧 ∈ R2𝑁 | lim π‘Š (𝑑) 𝑧 = 0} ,
(71)
2𝑁
∀𝑑 ∈ R,
(74)
where 𝑐 is the constant in condition (𝑅1 ) and 𝐼 is the identity
map on R2𝑁. Since 𝐡1 ≤ 𝐡2 , πœ‡πΉ (𝐡1 ) = πœ‡πΉ (𝐡2 ) and 𝜐𝐹 (𝐾𝐡1 ) =
𝜐𝐹 (𝐾𝐡2 ) = 0, we can choose πœ€ small enough, such that for
each 𝑛 ∈ N+ , πœ‡πΉ (𝐾𝐢𝑛 ) = πœ‡πΉ (𝐾𝐡1 ) and 𝜐𝐹 (𝐾𝐢𝑛 ) = 0. Thus
𝐹 − 𝐾𝐢𝑛 is reversible on 𝐸, and there is a constant 𝛿 > 0, such
that
σ΅„©
σ΅„©σ΅„©
+
(75)
σ΅„©σ΅„©(𝐹 − 𝐾𝐢𝑛 )𝑧󡄩󡄩󡄩𝐸 ≥ 𝛿‖𝑧‖𝐸 , ∀𝑧 ∈ 𝐸, 𝑛 ∈ N .
On the other hand, for 𝑏 ∈ (0, 1), there is a constant 𝑐 > 0
depending on 𝑏, such that for each 𝑛 ∈ N+ ,
󡄨
󡄨
󡄨𝑏
󡄨󡄨
(76)
󡄨󡄨∇𝑧 𝑅 (𝑑, 𝑧𝑛 (𝑑)) − 𝐢𝑛 𝑧𝑛 (𝑑)󡄨󡄨󡄨 ≤ 𝑐 󡄨󡄨󡄨𝑧𝑛 (𝑑)󡄨󡄨󡄨 , ∀𝑑 ∈ R.
Choose 𝑏 > (1−𝛼)/(1+𝛼) in (76); that is, 1+𝑏 ∈ (2/(1+𝛼), 2),
we have
σ΅„©2
σ΅„©σ΅„©
σ΅„©σ΅„©(𝐹𝑧𝑛 − 𝐾∇𝑧 𝑅 (𝑑, 𝑧𝑛 )) − (𝐹 − 𝐾𝐢𝑛 ) 𝑧𝑛 󡄩󡄩󡄩𝐸
σ΅„©
σ΅„©2
= 󡄩󡄩󡄩𝐾 (∇𝑧 𝑅 (𝑑, 𝑧𝑛 ) − 𝐢𝑛 𝑧𝑛 )󡄩󡄩󡄩𝐸
σ΅„©2
σ΅„©
≤ σ΅„©σ΅„©σ΅„©∇𝑧 𝑅 (𝑑, 𝑧𝑛 ) − 𝐢𝑛 𝑧𝑛 󡄩󡄩󡄩𝐿2
(77)
󡄨󡄨
󡄨
󡄨󡄨∇𝑧 𝑅 (𝑑, 𝑧𝑛 ) − 𝐢𝑛 𝑧𝑛 󡄨󡄨󡄨 󡄨󡄨 󡄨󡄨1+𝑏
≤ 𝑐∫
󡄨󡄨𝑧𝑛 󡄨󡄨 𝑑𝑑
󡄨󡄨 󡄨󡄨𝑏
R
󡄨󡄨𝑧𝑛 󡄨󡄨
σ΅„© σ΅„©1+𝑏
≤ 𝑐󡄩󡄩󡄩𝑧𝑛 󡄩󡄩󡄩𝐿1+𝑏 .
8
Abstract and Applied Analysis
For each π‘˜ ∈ N, we consider the following problem
As we claimed in the introduction, in (76) and (77), the letter
𝑐 denotes different positive constants whose exact value is
irrelevant. Thus, from (72), (75), (77), and Lemma 1, we have
that {𝑧𝑛 } is bounded in 𝐸, and π‘Ž satisfies the (PS) conditions.
And by Lemma 5.1 in Chapter II of [21], we have
π»π‘ž (𝑋, (π‘Ž)𝛼 ; R) ≅ π›Ώπ‘ž,π‘Ÿ R,
π‘ž = 0, 1, . . . ,
(78)
for −𝛼 ∈ R large enough.
From Theorem 11, Lemmas 14 and 15, Theorem 2 is a
direct consequence of Theorem 5.1 and Corollary 5.2 in
Chapter II of [21].
In order to proof Theorems 3 and 4, we need the following
lemma which is similar to Lemma 3.4 in [29] and Lemma 3.3
in [23].
Μƒ = ∇𝑧 π‘…π‘˜ (𝑑, 𝑧) ,
𝐹𝑧
𝑧 (𝑑) 󳨀→ 0,
π‘…π‘˜ (𝑑, 𝑧) ≡ 𝑅 (𝑑, 𝑧) ,
∀𝑑 ∈ R, |𝑧| ≤ π‘€π‘˜ .
(3) For each π‘˜ ∈ N, there exist some πΆπ‘˜ > 0 and a constant
𝛾 with 𝛾𝐼2𝑁 > 𝐡∞ , ]𝐹 (𝐾𝛾𝐼2𝑁) = 0, such that
󡄨󡄨
󡄨
󡄨󡄨∇𝑧 π‘…π‘˜ (𝑑, 𝑧) − 𝛾𝑧󡄨󡄨󡄨 < πΆπ‘˜ ,
∀ (𝑑, 𝑧) ∈ R × R2𝑁,
(81)
where 𝐼2𝑁 is the identity map on R2𝑁.
Proof. Define πœ‚ : [0, ∞) → R by
0,
0 ≤ 𝑠 < 1,
{
{
{
{
{
{
1
{2
3
4
πœ‚ (𝑠) = { 9 (𝑠 − 1) − 9 (𝑠 − 1) , 1 ≤ 𝑠 < 2,
{
{
{
{
128
{
{1 −
,
2 ≤ 𝑠 < ∞.
9 (12 + 𝑠2 )
{
𝛾
π‘…π‘˜ (𝑑, 𝑧) = (1 − πœ‚π‘˜ (|𝑧|)) 𝑅 (𝑑, 𝑧) + πœ‚π‘˜ (|𝑧|) |𝑧|2 ,
2
π‘˜ ∈ N.
(83)
As in [23, 29], we can check that π‘…π‘˜ satisfies (79)–(81) for each
π‘˜ ∈ N.
(84)
𝛽
𝑃𝛽 = ∫ π‘‘πΈπœ† ,
−𝛽
2
(85)
2𝑁
𝑋𝛽 = 𝑃𝛽 𝐿 (R, R ) .
Thus, for each π‘˜ and such a 𝛽 fixed, by Theorem 8, we have a
functional
π‘Žπ‘˜,𝛽 (π‘₯) ,
π‘₯ ∈ 𝑋𝛽 ,
(86)
whose critical points give rise to solutions of (HS)π‘˜ . Similarly
we have a functional
π‘Žπ›Ύ,𝛽 (π‘₯) ,
π‘₯ ∈ 𝑋𝛽 ,
(87)
whose critical points give rise to solutions of the following
systems (HS)𝛾
Μƒ = 𝛾𝑧,
𝐹𝑧
𝑧 (𝑑) 󳨀→ 0,
𝑧󸀠 (𝑑) 󳨀→ 0,
𝑑 󳨀→ ∞.
(HS)𝛾
For notational simplicity, we denote π‘Žπ‘˜ , π‘Žπ›Ύ for π‘Žπ‘˜,𝛽 and π‘Žπ›Ύ,𝛽 .
Define
∞
−∞
It is easy to see that πœ‚ ∈ 𝐢2 ([0, ∞), R). Choose a sequence
{π‘€π‘˜ } of positive numbers such that 𝑅0 < 𝑀1 < 𝑀2 < ⋅ ⋅ ⋅ <
π‘€π‘˜ < ⋅ ⋅ ⋅ → ∞ as π‘˜ → ∞. For each π‘˜ ∈ N, let πœ‚π‘˜ (𝑠) =
πœ‚(𝑠/π‘€π‘˜ ) and
(HS)π‘˜
Let
Φπ‘˜ (𝑧) = ∫
(82)
𝛽 ∉ 𝜎 (𝐴 0 ) .
𝛽 > max {2 (𝐢 + 1) , 2 (𝛾 + 1)} ,
(79)
(2) For each π‘˜ ∈ N, there is a 𝐢 > 0 independent of π‘˜, such
that
󡄨
󡄨󡄨 2
󡄨󡄨∇𝑧 π‘…π‘˜ (𝑑, 𝑧)󡄨󡄨󡄨 ≤ 𝐢, ∀𝑑 ∈ R, 𝑧 ∈ R2𝑁,
󡄨
󡄨
(80)
2
∇𝑧 π‘…π‘˜ (𝑑, 𝑧) ≥ 𝐡∞ , ∀𝑑 ∈ R, |𝑧| ≥ 𝑅0 .
𝑑 󳨀→ ∞,
where π‘…π‘˜ is given in Lemma 16. Performing on (HS)π‘˜ the
saddle point reduction, we choose the number 𝛽 which is
used in the projection for the saddle point reduction in
Section 2.2. First, we choose
+
) hold; then, there
Lemma 16. Assume that (𝑅1 ), (𝑅0 ), and (𝑅∞
2
exists a sequence of functions π‘…π‘˜ ∈ 𝐢 (R × R2𝑁, R), π‘˜ ∈ N,
satisfying the following properties.
(1) There exists an increasing sequence of real numbers
π‘€π‘˜ → ∞ (π‘˜ → ∞) such that
𝑧󸀠 (𝑑) 󳨀→ 0,
π‘…π‘˜ (𝑑, 𝑧) 𝑑𝑑.
(88)
For the functional π‘Žπ‘˜ , similar to Lemma 15, we have the following lemma.
Lemma 17. (1) π‘Žπ‘˜ satisfies (PS) condition, and the critical
point set of π‘Žπ‘˜ is compact;
(2) π»π‘ž (𝑋𝛽 , (π‘Žπ‘˜ )π›Όπ‘˜ ; R) ≅ π›Ώπ‘ž,π‘Ÿπ›½ R, π‘ž = 0, 1, . . . for −π›Όπ‘˜ ∈ R
large enough, where π‘Ÿπ›½ = dim(𝐸− (𝑋𝛽 )) + πœ‡πΉ (𝐾𝛾𝐼2𝑁).
Proof. The proof is similar to Lemma 15. From Theorem 11,
we have
σ΅„©σ΅„©σ΅„©π‘ŽσΈ€  (π‘₯) − π‘ŽσΈ€ σΈ€  (0) π‘₯σ΅„©σ΅„©σ΅„© = 󡄩󡄩󡄩𝑃 𝐾(∇ Φ (𝑧 (π‘₯)) − 𝛾π‘₯)σ΅„©σ΅„©σ΅„©
σ΅„©σ΅„© π‘˜
󡄩󡄩𝐿2 σ΅„©σ΅„© 𝛽
󡄩󡄩𝐿2
𝑧 π‘˜ π‘˜
𝛾
σ΅„©
σ΅„©
≤ 󡄩󡄩󡄩󡄩𝑃𝛽 (∇𝑧 π‘…π‘˜ (π‘§π‘˜ (π‘₯)) − π›Ύπ‘§π‘˜ (π‘₯))󡄩󡄩󡄩󡄩𝐿2
σ΅„©
σ΅„©
≤ σ΅„©σ΅„©σ΅„©(∇𝑧 π‘…π‘˜ (π‘§π‘˜ (π‘₯)) − π›Ύπ‘§π‘˜ (π‘₯))󡄩󡄩󡄩𝐿2 .
(89)
Abstract and Applied Analysis
9
Similar to (76), for 𝑏 ∈ (0, 1), there is some 𝑐 > 0, such that
󡄨
󡄨
󡄨󡄨
󡄨
(90)
󡄨󡄨∇𝑧 π‘…π‘˜ (𝑑, π‘§π‘˜ ) − π›Ύπ‘§π‘˜ 󡄨󡄨󡄨 ≤ 𝑐 σ΅„¨σ΅„¨σ΅„¨π‘§π‘˜ (𝑑)󡄨󡄨󡄨 , ∀𝑑 ∈ R.
for 𝛽 large enough, we have the functional π‘Ž∞,𝛽 (denote by
π‘Ž∞ for simplicity) and the function 𝑧(π‘₯); since 𝜐𝐹 (𝐾𝐡∞ ) = 0,
we have the following decomposition:
Choose 𝑏 ∈ ((1 − 𝛼)/(1 + 𝛼), 1), similar to (77), and we have
𝑋𝛽 = 𝑋𝛽+ + 𝑋𝛽− ,
σ΅„©σ΅„©
σ΅„©2
σ΅„© σ΅„©1+𝑏
σ΅„©σ΅„©∇𝑧 π‘…π‘˜ (𝑑, π‘§π‘˜ ) − π›Ύπ‘§π‘˜ 󡄩󡄩󡄩𝐿2 ≤ π‘σ΅„©σ΅„©σ΅„©π‘§π‘˜ 󡄩󡄩󡄩𝐿1+𝑏 .
(91)
From Lemma 1, Remark 10, and (91),
σ΅„©σ΅„©
σ΅„©2
σ΅„© σ΅„©1+𝑏
σ΅„©σ΅„©∇𝑧 π‘…π‘˜ (𝑑, π‘§π‘˜ ) − π›Ύπ‘§π‘˜ 󡄩󡄩󡄩𝐿2 ≤ π‘σ΅„©σ΅„©σ΅„©π‘§π‘˜ 󡄩󡄩󡄩𝐸
≤ 𝑐 (𝛽) β€–π‘₯β€–1+𝑏
𝐿2 .
(92)
From (89) and (92), we have
σ΅„©σ΅„© σΈ€ 
σ΅„©2
σ΅„©σ΅„©π‘Žπ‘˜ (π‘₯) − π‘Žπ›ΎσΈ€ σΈ€  (0) π‘₯σ΅„©σ΅„©σ΅„© 2 ≤ 𝑐 (𝛽) β€–π‘₯β€–1+𝑏
𝐿2 .
σ΅„©
󡄩𝐿
(93)
Now, for each π‘˜ ∈ N, we assume that {π‘₯π‘š } ⊂ 𝑋𝛽 satisfying
β€–π‘Žπ‘˜σΈ€  (π‘₯π‘š )β€– → 0. By ]𝐹 (𝐾𝛾𝐼2𝑁) = 0, we have that π‘Žπ›ΎσΈ€ σΈ€  (0) is
invertible on 𝑋𝛽 , since 𝑏 < 1 the sequence {π‘₯π‘š } must be
bounded. Thus the (PS) condition for π‘Žπ‘˜ holds. From the same
reason, we have the compactness of the critical point set of π‘Žπ‘˜ .
And by Lemma 5.1 in Chapter II of [21], we have
π»π‘ž (𝑋𝛽 , (π‘Žπ‘˜ )π›Όπ‘˜ ; R) ≅ π›Ώπ‘ž,π‘Ÿπ›½ R,
π‘ž = 0, 1, . . . ,
(94)
for −π›Όπ‘˜ ∈ R large enough.
((𝐹̃ − 𝐡∞ (𝑑)) π‘₯, π‘₯)𝐿2 ≤ −𝛼‖π‘₯β€–2𝐿2 ,
∀π‘₯ ∈ 𝑋𝛽− .
(99)
From the uniform boundary of ∇𝑧2 π‘…π‘˜ (𝑑, 𝑧) and Remark 10, we
can choose 𝛽 large enough, such that
𝛼
σ΅„©
σ΅„©σ΅„© 2
σ΅„©σ΅„©(∇𝑧 π‘…π‘˜ (𝑑, π‘§π‘˜ ) π‘§π‘˜σΈ€ ± (π‘₯π‘˜ ) π‘₯, π‘₯σ΅„©σ΅„©σ΅„© 2 ≤ β€–π‘₯‖𝐿2 ,
󡄩𝐿
σ΅„©
4
∀π‘₯ ∈ 𝐿2 ,
(100)
where π‘₯π‘˜ = 𝑃𝛽 π‘§π‘˜ , π‘§π‘˜ (π‘₯π‘˜ ) = π‘§π‘˜ (𝑑) defined in Theorem 8.
Choose πœ€ > 0 small enough and πœ€ < 𝛼/4, such that
πœ‡πΉ (𝐾𝐡∞ ) = πœ‡πΉ (𝐾(𝐡∞ −πœ€⋅𝐼𝑑)). Since 𝑋𝛽− is finite dimensional
space, choose π‘Ÿ large enough, such that
((∇𝑧2 π‘…π‘˜ (𝑑, π‘§π‘˜ ) − 𝐡∞ ) π‘₯, π‘₯)𝐿2 (𝐼𝑐 ) ≥ −πœ€(π‘₯, π‘₯)𝐿2 ,
∀π‘₯ ∈ 𝑋𝛽− ,
(101)
Proof. We prove it indirectly. Assume there exist π‘…π‘˜ and π‘§π‘˜
satisfying the conditions, and β€–π‘§π‘˜ ‖𝐿∞ → ∞; that is, ‖𝑧‖𝐹̃ →
∞. Since |∇𝑧 π‘…π‘˜ (𝑑, 𝑧)| < 𝑐|𝑧|, for all 𝑑 ∈ R, 𝑧 ∈ R2𝑁, we have
β€–π‘§π‘˜ ‖𝐹̃ ≤ π‘β€–π‘§π‘˜ ‖𝐿2 . Denote that π‘¦π‘˜ = π‘§π‘˜ /β€–π‘§π‘˜ ‖𝐹̃ ; then, we have
π‘¦π‘˜ → 𝑦 in 𝐿2 for some 𝑦 ∈ 𝐿2 with ‖𝑦‖𝐿2 > 0, and
π‘…π‘˜σΈ€  (𝑑, π‘§π‘˜ )
σ΅„©σ΅„© σ΅„©σ΅„© .
σ΅„©σ΅„©π‘§π‘˜ 󡄩󡄩𝐹̃
(95)
Then, for any π‘Ÿ > 0, there exists πΆπ‘Ÿ > 0, satisfying
󡄨
󡄨
󡄨󡄨 Μ‡
󡄨
σ΅„¨σ΅„¨π‘¦π‘˜ (𝑑)󡄨󡄨󡄨 ≤ πΆπ‘Ÿ σ΅„¨σ΅„¨σ΅„¨π‘¦π‘˜ (𝑑)󡄨󡄨󡄨 , 𝑑 ∈ πΌπ‘Ÿ ,
1σ΅„© σ΅„©
σ΅„©σ΅„© σ΅„©σ΅„©
󡄩󡄩𝑦󡄩󡄩𝐿2 (−π‘Ÿ,π‘Ÿ) > 󡄩󡄩󡄩𝑦󡄩󡄩󡄩𝐿2 > 0,
2
∀π‘Ÿ > π‘Ÿ0 .
(96)
(97)
Then, from the similar argument in [23], there is a subsequence throught which we may assume that {π‘¦π‘˜ } converges
in uniform norm to 𝑦, and 𝑦(𝑑) =ΜΈ 0, for all 𝑑 ∈ πΌπ‘Ÿ . Therefore
|π‘§π‘˜ (𝑑)| → ∞ uniformly on πΌπ‘Ÿ , and there is 𝐾(π‘Ÿ) depending
on π‘Ÿ, such that |π‘§π‘˜ (𝑑)| ≥ 𝑅0 , for any 𝑑 ∈ πΌπ‘Ÿ and π‘˜ ≥ 𝐾(π‘Ÿ).
Performing the saddle point reduction on the following
systems:
Μƒ = 𝐡∞ 𝑧,
𝐹𝑧
𝑧󸀠 (𝑑) 󳨀→ 0,
𝑑 󳨀→ ∞,
where πΌπ‘Ÿπ‘ = R \ πΌπ‘Ÿ , and from the definition of π‘…π‘˜ ,
∇𝑧2 π‘…π‘˜ (𝑑, π‘§π‘˜ (𝑑)) ≥ 𝐡∞ (𝑑) ,
𝑑 ∈ πΌπ‘Ÿ , π‘˜ ≥ 𝐾 (π‘Ÿ) ;
(HS)∞
(102)
that is,
((∇𝑧2 π‘…π‘˜ (𝑑, π‘§π‘˜ ) − 𝐡∞ ) π‘₯, π‘₯)𝐿2 (𝐼 ) ≥ 0,
π‘Ÿ
∀π‘₯ ∈ 𝑋𝛽− , π‘˜ ≥ 𝐾 (π‘Ÿ) .
(103)
From (101) and (103),
(∇𝑧2 π‘…π‘˜ (𝑑, π‘§π‘˜ ) π‘₯, π‘₯)𝐿2 ≥ (𝐡∞ π‘₯, π‘₯)𝐿2 − πœ€ (π‘₯, π‘₯)𝐿2 ,
where πΌπ‘Ÿ = [−π‘Ÿ, π‘Ÿ]. Since ‖𝑦‖𝐿2 > 0, there is a π‘Ÿ0 > 0, such that
𝑧 (𝑑) 󳨀→ 0,
σΈ€ σΈ€ 
where π‘Ž∞
(0) is positive definite on 𝑋𝛽+ and negative definite
on 𝑋𝛽− . From Remark 10, and 𝜐𝐹 (𝐾𝐡∞ ) = 0, there exists 𝛼 > 0,
such that for 𝛽 large enough,
π‘Ÿ
Lemma 18. There exist 𝑐 > 0, such that for any π‘˜ ∈ N, and 𝑧 ∈
𝐿2 satisfies the systems (HS)π‘˜ , if πœ‡πΉ (𝑧) ≤ πœ‡πΉ (𝐾𝐡∞ ) − 1, one has
‖𝑧‖𝐿∞ ≤ 𝑐.
Μƒ π‘˜=
𝐹𝑦
(98)
(104)
for π‘˜ large enough. Thus we have
(π‘Žπ‘˜σΈ€ σΈ€  (π‘₯π‘˜ ) π‘₯, π‘₯)𝐿2
= ((𝐹̃ − ∇𝑧2 π‘…π‘˜ (𝑑, π‘§π‘˜ )) π‘₯, π‘₯)𝐿2
− (∇𝑧2 (π‘…π‘˜ (𝑑, π‘§π‘˜ )) (π‘§π‘˜σΈ€ + + π‘§π‘˜σΈ€ − ) π‘₯, π‘₯)𝐿2
(105)
𝛼
≤ ((𝐹̃ − 𝐡∞ ) π‘₯, π‘₯)𝐿2 + β€–π‘₯β€–2𝐿2 + πœ€β€–π‘₯β€–2𝐿2
2
𝛼
≤ − β€–π‘₯β€–2𝐿2 .
4
That is, π‘šπ‘Ž−π‘˜ (π‘₯) ≥ π‘šπ‘Ž−∞ (0), from Theorem 11, π‘šπ‘Ž−π‘˜ (π‘₯π‘˜ ) =
dim(𝐸− (H0 )) + πœ‡πΉ (𝑧π‘₯ (π‘₯π‘˜ )), π‘šπ‘Ž−∞ (0) = dim(𝐸− (H0 )) +
πœ‡πΉ (𝐾𝐡∞ ), and thus πœ‡πΉ (𝑧π‘₯ ) ≥ πœ‡πΉ (𝐾𝐡∞ ), which contradicts
the assumption.
10
Abstract and Applied Analysis
Proof of Theorem 3. As claimed in Remark 13, we can only
+
). Note that 𝑧 = 0 is a critical point of
consider the case of (𝑅∞
π‘Žπ‘˜ , and the morse index of 0 for π‘Žπ‘˜ is π‘šπ‘Ž−π‘˜ (0) = dim(𝐸− (H0 ))+
πœ‡πΉ (𝐾𝐡0 ), since 𝛾 ⋅ 𝐼2𝑁 > 𝐡∞ , we have
πœ‡πΉ (𝐾𝛾 ⋅ 𝐼2𝑁) ≥ πœ‡πΉ (𝐾𝐡∞ ) ≥ πœ‡πΉ (𝐾𝐡0 ) .
(106)
From proposition (2) in Lemma 17, use the (π‘šπ‘Ž−π‘˜ (0))th and
(π‘šπ‘Ž−π‘˜ (0) + 1)th Morse inequalities, and π‘Žπ‘˜ has a nontrivial
critical point π‘₯π‘˜ with its morse index π‘šπ‘Ž−π‘˜ (π‘₯π‘˜ ) ≤ π‘šπ‘Ž−π‘˜ (0) + 1;
that is, πœ‡πΉ (π‘§π‘˜ ) ≤ πœ‡πΉ (𝐾𝐡0 ) + 1 ≤ πœ‡πΉ (𝐾𝐡∞ ) − 1, then from
Lemma 18, we have that {π‘§π‘˜ } is bounded in 𝐿∞ . Thus π‘§π‘˜ is a
nontrivial solution of (HS) for π‘˜ large enough.
The proof of Theorem 4 is similar to the proof of
Theorem 3. Instead of Morse theory we make use of minimax
arguments for multiplicity of critical points.
Let 𝑋 be a Hilbert space and assume πœ™ ∈ 𝐢2 (𝑋, R) is an
even functional, satisfying the (PS) condition and πœ™(0) = 0.
Denote 𝑆𝑐 = {𝑒 ∈| 𝑋‖𝑒‖ = 𝑐}.
Lemma 19 (see [30, Corollary 10.19]). Assume that π‘Œ and 𝑍
are subspaces of 𝑋 satisfying dim π‘Œ = 𝑗 > π‘˜ = codim 𝑍. If
there exist 𝑅 > π‘Ÿ > 0 and 𝛼 > 0 such that
inf πœ™ (π‘†π‘Ÿ ∩ 𝑍) ≥ 𝛼,
sup πœ™ (𝑆𝑅 ∩ π‘Œ) ≤ 0,
(107)
then πœ™ has 𝑗−π‘˜ pairs of nontrivial critical points {±π‘₯1 , ±π‘₯2 , . . . ,
±π‘₯𝑗−π‘˜ }, so that πœ‡(π‘₯𝑖 ) ≤ π‘˜ + 𝑖, for 𝑖 = 1, 2, . . . , 𝑗 − π‘˜.
+
), since 𝑅 is even, we
First, we consider the case of (𝑅∞
have that π‘…π‘˜ is also even and satisfies Lemma 16. Let π‘Œ = 𝑋𝛽− ,
and 𝑍 the positive space of π‘Žπ‘˜σΈ€ σΈ€  (0) in 𝑋𝛽 , and we have and
dim π‘Œ = 𝐸− (𝑋𝛽 ) + πœ‡πΉ (𝐾𝐡∞ ), codim 𝑍 = 𝐸− (𝑋𝛽 ) + πœ‡πΉ (𝐾𝐡0 ),
dim π‘Œ > codim𝑍. So, π‘Žπ‘˜ has 𝑙 := πœ‡πΉ (𝐾𝐡∞ ) − πœ‡πΉ (𝐾𝐡0 ) pairs
of nontrivial critical points
{±π‘₯1 , ±π‘₯2 , . . . , ±π‘₯𝑙 } ,
(108)
and 𝑙 − 1 pairs of them satisfy
π‘š− (π‘₯𝑖 ) ≤ πœ‡πΉ (𝐾𝐡0 ) + 𝑖 < πœ‡πΉ (𝐾𝐡∞ ) ,
𝑖 = 1, 2, . . . , 𝑙 − 1.
(109)
Then, we can complete the proof. In order to prove the case
−
), we need the following lemma.
of (𝑅∞
Lemma 20 (see [21, Corollary II 4.1]). Assume that π‘Œ and 𝑍
are subspaces of 𝑋 satisfying dim π‘Œ = 𝑗 > π‘˜ = codim𝑍. If
there exist π‘Ÿ > 0, and 𝛼 > 0 such that
inf πœ™ (𝑍) > −∞,
sup πœ™ (π‘†π‘Ÿ ∩ π‘Œ) ≤ −𝛼,
(110)
then πœ™ has 𝑗 − π‘˜ pairs of nontrivial critical points ±π‘’1 , ±π‘’2 , . . . ,
±π‘’𝑗−π‘˜ so that πœ‡(𝑒𝑖 ) + ](𝑒𝑖 ) ≥ π‘˜ + 𝑖 − 1 for 𝑖 = 1, 2, . . . , 𝑗 − π‘˜.
+
), and we omit it
The proof is similar to the case of (𝑅∞
here.
Acknowledgment
This work is partially supported by NFSC (11126154, 11126146,
and 11201196).
References
[1] G. Arioli and A. Szulkin, “Homoclinic solutions of Hamiltonian
systems with symmetry,” Journal of Differential Equations, vol.
158, no. 2, pp. 291–313, 1999.
[2] G. Chen and S. Ma, “Homoclinic orbits of superlinear Hamiltonian systems,” Proceedings of the American Mathematical Society, vol. 139, no. 11, pp. 3973–3983, 2011.
[3] V. C. Zelati, I. Ekeland, and E. Séré, “A variational approach
to homoclinic orbits in Hamiltonian systems,” Mathematische
Annalen, vol. 288, no. 1, pp. 133–160, 1990.
[4] Y. Ding, “Multiple homoclinics in a Hamiltonian system with
asymptotically or super linear terms,” Communications in Contemporary Mathematics, vol. 8, no. 4, pp. 453–480, 2006.
[5] Y. Ding and M. Girardi, “Infinitely many homoclinic orbits
of a Hamiltonian system with symmetry,” Nonlinear Analysis:
Theory, Methods & Applications, vol. 38, no. 3, pp. 391–415, 1999.
[6] Y. Ding and M. Willem, “Homoclinic orbits of a Hamiltonian
system,” Zeitschrift für Angewandte Mathematik und Physik, vol.
50, no. 5, pp. 759–778, 1999.
[7] H. Hofer and K. Wysocki, “First order elliptic systems and the
existence of homoclinic orbits in Hamiltonian systems,” Mathematische Annalen, vol. 288, no. 3, pp. 483–503, 1990.
[8] E. Séré, “Looking for the Bernoulli shift,” Annales de l’Institut
Henri Poincaré. Analyse Non Linéaire, vol. 10, no. 5, pp. 561–590,
1993.
[9] E. Séré, “Existence of infinitely many homoclinic orbits in
Hamiltonian systems,” Mathematische Zeitschrift, vol. 209, no.
1, pp. 27–42, 1992.
[10] J. Sun, J. Chu, and Z. Feng, “Homoclinic orbits for first order
periodic Hamiltonian systems with spectrum point zero,” Discrete and Continuous Dynamical Systems, vol. 33, no. 8, pp. 1241–
1257, 2013.
[11] A. Szulkin and W. Zou, “Homoclinic orbits for asymptotically
linear Hamiltonian systems,” Journal of Functional Analysis, vol.
187, no. 1, pp. 25–41, 2001.
[12] K. Tanaka, “Homoclinic orbits in a first order superquadratic
Hamiltonian system: convergence of subharmonic orbits,” Journal of Differential Equations, vol. 94, no. 2, pp. 315–339, 1991.
[13] Y. H. Ding and S. J. Li, “Homoclinic orbits for first order Hamiltonian systems,” Journal of Mathematical Analysis and Applications, vol. 189, no. 2, pp. 585–601, 1995.
[14] Y. Ding and L. Jeanjean, “Homoclinic orbits for a nonperiodic
Hamiltonian system,” Journal of Differential Equations, vol. 237,
no. 2, pp. 473–490, 2007.
[15] Y. Ding and C. Lee, “Existence and exponential decay of homoclinics in a nonperiodic superquadratic Hamiltonian system,”
Journal of Differential Equations, vol. 246, no. 7, pp. 2829–2848,
2009.
[16] W. Qin, J. Zhang, and F. Zhao, “Homoclinic orbits for a class
of nonperiodic Hamiltonian systems,” Abstract and Applied
Analysis, vol. 2012, Article ID 769232, 20 pages, 2012.
[17] J. Sun, H. Chen, and J. J. Nieto, “Homoclinic orbits for a class of
first-order nonperiodic asymptotically quadratic Hamiltonian
systems with spectrum point zero,” Journal of Mathematical
Analysis and Applications, vol. 378, no. 1, pp. 117–127, 2011.
Abstract and Applied Analysis
[18] C. N. Chen and X. Hu, “Maslov index for homoclinic orbits
of Hamiltonian systems,” Annales de l’Institut Henri Poincaré.
Analyse Non Linéaire, vol. 24, no. 4, pp. 589–603, 2007.
[19] C. Zhu and Y. Long, “Maslov-type index theory for symplectic
paths and spectral flow. I,” Chinese Annals of Mathematics B, vol.
20, no. 4, pp. 413–424, 1999.
[20] H. Amann, “Saddle points and multiple solutions of differential
equations,” Mathematische Zeitschrift, vol. 169, no. 2, pp. 127–
166, 1979.
[21] K. C. Chang, Infinite-Dimensional Morse Theory and Multiple
Solution Problems, Progress in Nonlinear Differential Equations
and their Applications, Birkhäauser, Basel, Switzerland, 1993.
[22] Y. Long, Index Theory for Symplectic Paths with Applications, vol.
207 of Progress in Mathematics, Birkhäuser, Basel, Switzerland,
2002.
[23] Z. Liu, J. Su, and Z. Q. Wang, “A twist condition and periodic
solutions of Hamiltonian systems,” Advances in Mathematics,
vol. 218, no. 6, pp. 1895–1913, 2008.
[24] M. F. Atiyah, V. K. Patodi, and I. M. Singer, “Spectral asymmetry
and Riemannian geometry. III,” Mathematical Proceedings of the
Cambridge Philosophical Society, vol. 79, no. 1, pp. 71–99, 1976.
[25] S. E. Cappell, R. Lee, and E. Y. Miller, “On the Maslov index,”
Communications on Pure and Applied Mathematics, vol. 47, no.
2, pp. 121–186, 1994.
[26] A. Floer, “A relative Morse index for the symplectic action,”
Communications on Pure and Applied Mathematics, vol. 41, no.
4, pp. 393–407, 1988.
[27] J. Robbin and D. Salamon, “The Maslov index for paths,” Topology, vol. 32, no. 4, pp. 827–844, 1993.
[28] J. Robbin and D. Salamon, “The spectral flow and the Maslov
index,” The Bulletin of the London Mathematical Society, vol. 27,
no. 1, pp. 1–33, 1995.
[29] Z. Liu, J. Su, and Z. Q. Wang, “Solutions of elliptic problems
with nonlinearities of linear growth,” Calculus of Variations and
Partial Differential Equations, vol. 35, no. 4, pp. 463–480, 2009.
[30] N. Ghoussoub, Duality and Perturbation Methods in Critical
Point Theory, vol. 107 of Cambridge Tracts in Mathematics, Cambridge University Press, Cambridge, UK, 1993.
11
Advances in
Operations Research
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Advances in
Decision Sciences
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Mathematical Problems
in Engineering
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Journal of
Algebra
Hindawi Publishing Corporation
http://www.hindawi.com
Probability and Statistics
Volume 2014
The Scientific
World Journal
Hindawi Publishing Corporation
http://www.hindawi.com
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
International Journal of
Differential Equations
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Volume 2014
Submit your manuscripts at
http://www.hindawi.com
International Journal of
Advances in
Combinatorics
Hindawi Publishing Corporation
http://www.hindawi.com
Mathematical Physics
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Journal of
Complex Analysis
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
International
Journal of
Mathematics and
Mathematical
Sciences
Journal of
Hindawi Publishing Corporation
http://www.hindawi.com
Stochastic Analysis
Abstract and
Applied Analysis
Hindawi Publishing Corporation
http://www.hindawi.com
Hindawi Publishing Corporation
http://www.hindawi.com
International Journal of
Mathematics
Volume 2014
Volume 2014
Discrete Dynamics in
Nature and Society
Volume 2014
Volume 2014
Journal of
Journal of
Discrete Mathematics
Journal of
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com
Applied Mathematics
Journal of
Function Spaces
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Optimization
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Download