This paper considers the long-time behavior for a system of coupled wave equations of higher-order Kirchhoff type with strong damping terms. Using the Hadamard graph transformation method, we obtain the existence of the inertial manifold while such equations satisfy the spectral interval condition.
The concept of inertial manifold proposed by C. Foias, G. R. Sell and R. Temam [
It is well known that early researches on inertial manifold have yielded considerable results. In 1988, the concept of spectral barriers was utilized in the Hilbert space to attempt to refine spectral separation condition by Constantin et al. [
In recent years, there have been many works which focus on using the latter method to study it. Wu Jingzhu and Lin Guoguang introduced the graph transformation method in [
u t t − α Δ u t − Δ u + u 2 k + 1 = f ( x , y ) .
Subsequently, Xu Guigui, Wang Libo, and Lin Guoguang dealt with the existence of inertial manifold for second-order nonlinear wave equation with delays in the literature [
∂ 2 u ∂ t 2 + α ∂ u ∂ t − β Δ ∂ u ∂ t − Δ u + g ( u ) = f ( x ) + h ( t , u t ) .
In addition, Guo Yamei and Li Huahui obtained the existence of inertial manifold for a class of strongly dissipative nonlinear wave equation in [
u t t − α Δ u t + Δ 2 u t − Δ u + Δ 2 u + Δ g ( u ) = f ( x ) .
Chen Ling, Wang Wei and Lin Guoguang discussed the situation of higher-order Kirchhoff equation in [
u t t + ( − Δ ) m u t + ϕ ( ‖ ∇ m u ‖ 2 ) ( − Δ ) m u + g ( u ) = f ( x ) .
In this paper, basing on previous studies, the existence of the inertial manifold for nonlinear Kirchhoff type equations with higher-order strong damping is considered by using the Hadamard graph transformation method. The paper is arranged as follows. In Section 2, some notations, definitions and lemmas are given. In Section 3, in order to acquire the result of the existence of the inertial manifold, we show spectral gap condition.
u t t + M ( ‖ D m u ‖ 2 + ‖ D m v ‖ 2 ) ( − Δ ) m u + β ( − Δ ) m u t + g 1 ( u , v ) = f 1 ( x ) , in Ω × [ 0,+ ∞ ) , (1.1)
v t t + M ( ‖ D m u ‖ 2 + ‖ D m v ‖ 2 ) ( − Δ ) m v + β ( − Δ ) m v t + g 2 ( u , v ) = f 2 ( x ) , in Ω × [ 0,+ ∞ ) , (1.2)
u ( x , 0 ) = u 0 ( x ) , u t ( x , 0 ) = u 1 ( x ) , x ∈ Ω , (1.3)
v ( x , 0 ) = v 0 ( x ) , v t ( x , 0 ) = v 1 ( x ) , x ∈ Ω , (1.4)
∂ i u ∂ n i = 0 , ∂ i v ∂ n i = 0 , i = 0 , 1 , 2 , ⋯ , m − 1 , x ∈ ∂ Ω , t ≥ 0 , (1.5)
where Ω is a bounded domain in R n with smooth boundary ∂ Ω , β > 0 is real number and m ≥ 1 is positive integer, M ( s ) is a nonnegative C 1 function
and satisfies 0 < m 0 ≤ M ( s ) ≤ m 1 ≤ β 2 μ k 4 , g j ( u , v ) and f j ( x ) ( j = 1 , 2 ) are nonlinear terms and external force terms respectively.
For convenience, we need the following notations in subsequent article. Considering a family of Hilbert spaces V a = D ( A a / 2 ) , a ∈ R , whose inner product and norm are given by ( ⋅ , ⋅ ) V a = ( A a / 2 , A a / 2 ) and ‖ ⋅ ‖ V a = ‖ A a / 2 ⋅ ‖ .
Apparently
V 0 = L 2 ( Ω ) , V m = H m ( Ω ) ∩ H 0 1 ( Ω ) , V 2 m = H 2 m ( Ω ) ∩ H 0 1 ( Ω ) .
Definition 2.1 [
space X, a subset μ ⊂ X is said to be an inertial manifold if it satisfies the following three properties:
1) μ is a finite-dimensional Lipschitz manifold;
2) μ is positively invariant, i.e., S ( t ) μ ⊂ μ , for all t ≥ 0 ;
3) μ attracts exponentially all orbits of solution , that is, there are constants η > 0 , c > 0 such that
d i s t ( S ( t ) x , μ ) ≤ c e − η t , t ≥ 0 , (2.1)
for every x ∈ X , and the rate of decay in (2.1) is exponential, uniformly for x in bounded sets in X. property 3) implies that the inertial manifold must contain the universal attractor.
In order to describe the spectral interval condition, we firstly consider that the nonlinear term F : X → X is globally bounded and Lipschitz continuous, and has a positive Lipschitz constant l F ; its operator A has several positive real part eigenvalues, and the eigenfunctions expand to the corresponding orthogonal spaces in X.
Lemma 2.1 Let the operator A : X → X have countable positive real part eigenvalues whose eigenfunctions expand to the corresponding orthogonal spaces in X, and F ∈ C b ( X , X ) satisfies the Lipschitz condition:
‖ F ( u ) − F ( v ) ‖ X ≤ l F ‖ u − v ‖ X , u , v ∈ X , (2.2)
and operator A satisfies spectral interval condition related to F, if the point spectrum of the operator A can be divided into the following two parts σ 1 and σ 2 , where σ 1 is finite,
∧ 1 = sup { Re λ | λ ∈ σ 1 } , ∧ 2 = inf { Re λ | λ ∈ σ 2 } , (2.3)
X i = s p a n { w j | λ j ∈ σ i } , ( i = 1 , 2 ) . (2.4)
Then
∧ 2 − ∧ 1 > 4 l F , (2.5)
X = X 1 ⊕ X 2 , (2.6)
hold with continuous orthogonal projection P 1 : X → X 1 , P 2 : X → X 2 .
Lemma 2.2 g i : V m × V m → V m × V m ( i = 1 , 2 ) is uniformly bounded and global Lipschitz continuous functions.
Proof. ∀ ( u ˜ , v ˜ ) , ( u , v ) ∈ V m × V m ,
‖ g 1 ( u ˜ , v ˜ ) − g 1 ( u , v ) ‖ V m × V m = ‖ g 1 u ( u + θ ( u ˜ − u ) , v + θ ( v ˜ − v ) ) ( u ˜ − u ) + g 1 v ( u + θ ( u ˜ − u ) , v + θ ( v ˜ − v ) ) ( v ˜ − v ) ‖ ≤ ‖ g 1 u ( u + θ ( u ˜ − u ) , v + θ ( v ˜ − v ) ) ( u ˜ − u ) ‖ V m × V m + ‖ g 1 v ( u + θ ( u ˜ − u ) , v + θ ( v ˜ − v ) ) ( v ˜ − v ) ‖ V m × V m ≤ l ( ‖ u ˜ − u ‖ V m + ‖ v ˜ − v ‖ V m ) .
‖ g 2 ( u ˜ , v ˜ ) − g 2 ( u , v ) ‖ V m × V m ≤ l ( ‖ u ˜ − u ‖ V m + ‖ v ˜ − v ‖ V m ) .
Lemma 2.3 Let eigenvalues λ k ± ( k ≥ 1 ) be arranged in non-decreasing order, then for m ∈ N , there is N ≥ m such that λ N − and λ N + 1 − are adjacent values.
Equations (1.1)-(1.2) are equivalent to the following first-order evolution equation
U t + A ∗ U = F ( U ) , (3.1)
with
A ∗ = ( 0 − I 0 0 M ( s ) ( − Δ ) m β ( − Δ ) m 0 0 0 0 0 − I 0 0 M ( s ) ( − Δ ) m β ( − Δ ) m ) , (3.2)
F ( U ) = ( 0 f 1 ( x ) − g 1 ( u , v ) 0 f 2 ( x ) − g 2 ( u , v ) ) . (3.3)
D ( A ∗ ) = { ( u , v ) ∈ V m × V m | ( u , v ) ∈ V 0 × V 0 , ( ( − Δ ) m u , ( − Δ ) m v ) ∈ V 0 × V 0 } × V 0 × V 0 ,
X = V m × V 0 × V m × V 0 .
To determine characteristic values of operator A ∗ , we consider the graph norm on X, which induced by the scale product
( U , V ) X = ( M ( s ) D m u , D m x ¯ ) + ( y ¯ , p ) + ( M ( s ) D m v , D m z ¯ ) + ( w ¯ , q ) , (3.4)
where U = ( u , p , v , q ) , V = ( x , y , z , w ) , x ¯ , y ¯ , z ¯ , w ¯ represent the conjugation of x , y , z , w respectively. Moreover, the operator A ∗ defined in (3.2) is monotone. Indeed, for U ∈ D ( A ∗ ) ,
( A ∗ U , U ) X = − ( M ( s ) D m p , D m u ¯ ) + ( p ¯ , M ( s ) ( − Δ ) m u + β ( − Δ ) m p ) − ( M ( s ) D m q , D m v ¯ ) + ( q ¯ , M ( s ) ( − Δ ) m v + β ( − Δ ) m q ) = − ( M ( s ) D m p , D m u ¯ ) + ( p ¯ , M ( s ) ( − Δ ) m u ) + ( p ¯ , β ( − Δ ) m p ) − ( M ( s ) D m q , D m v ¯ ) + ( q ¯ , M ( s ) ( − Δ ) m v ) + ( q ¯ , β ( − Δ ) m q ) = β ( ‖ D m p ‖ 2 + ‖ D m q ‖ 2 ) ≥ 0 , (3.5)
therefore, ( A ∗ U , U ) X is a non-negative real number.
To further determine the eigenvalues of A ∗ , we consider the following characteristic equation
A ∗ U = λ U , U = ( u , p , v , q ) ∈ X . (3.6)
That is
{ − p = λ u , M ( s ) ( − Δ ) m u + β ( − Δ ) m p = λ p , − q = λ v , M ( s ) ( − Δ ) m v + β ( − Δ ) m q = λ q . (3.7)
Substituting the first and third equations of (3.7) into the second and fourth equations, thus u , v satisfy the problem of eigenvalues
{ λ 2 u − λ β ( − Δ ) m u + M ( s ) ( − Δ ) m u = 0 , λ 2 v − λ β ( − Δ ) m v + M ( s ) ( − Δ ) m v = 0 , ∂ i u ∂ n i | ∂ Ω = ∂ i v ∂ n i | ∂ Ω = 0 , i = 0 , 1 , 2 , ⋯ , m − 1 , (3.8)
taking the inner product of u , v on both sides of the first and second equations of (3.8) respectively, we acquire
{ λ 2 ‖ u ‖ 2 − λ β ‖ D m u ‖ 2 + M ( s ) ‖ D m u ‖ 2 = 0 , λ 2 ‖ v ‖ 2 − λ β ‖ D m v ‖ 2 + M ( s ) ‖ D m v ‖ 2 = 0 , (3.9)
that is to say
λ 2 ( ‖ u ‖ 2 + ‖ v ‖ 2 ) − λ β ( ‖ D m u ‖ 2 + ‖ D m v ‖ 2 ) + M ( s ) ( ‖ D m u ‖ 2 + ‖ D m v ‖ 2 ) = 0. (3.10)
(3.10) is a quadratic equation about λ , bringing u k , v k to the position of u , v , for any positive integer k, the equation (3.6) has paired eigenvalues
λ k ± = β μ k ± ( β μ k ) 2 − 4 M ( μ k ) μ k 2 , (3.11)
where μ k is the characteristic value of ( − Δ ) m in V m × V m , then μ k = λ 1 k m n .
If
( β μ k ) 2 ≥ 4 M ( μ k ) μ k ,
that is
β ≥ 2 M ( μ k ) μ k or μ k ≥ 4 M ( μ k ) β 2 ,
then the eigenvalues of the operator A ∗ are all real numbers, and the corresponding characteristic functions are
U k ± = ( u k , − λ k ± u k , v k , − λ k ± v k ) .
For convenience, we note that for any k ≥ 1 ,
‖ D m u k ‖ 2 + ‖ D m v k ‖ 2 = μ k , ‖ u k ‖ 2 + ‖ v k ‖ 2 = 1 , ‖ D − m u k ‖ 2 + ‖ D − m v k ‖ 2 = 1 μ k .
Theorem 3.1 Suppose 0 < β < 2 M ( μ k ) μ k , and l be the Lipschitz constant of g i ( u , v ) ( i = 1 , 2 ) in (3.1) , set N 1 ∈ N be so large such that if N > N 1 ,
β ( μ N + 1 − μ N ) ≥ 16 l . (3.12)
Then the operator A ∗ satisfies the spectral interval condition of Definition 1.2.
Proof. We firstly estimate the Lipschitz property of F, from (3.1) and (3.4), we have
‖ F ( U ) − F ( V ) ‖ X = ‖ g 1 ( u , v ) − g 1 ( u ˜ , v ˜ ) ‖ + ‖ g 2 ( u , v ) − g 2 ( u ˜ , v ˜ ) ‖ ≤ 2 l ( ‖ u ˜ − u ‖ V m + ‖ v ˜ − v ‖ V m ) . (3.13)
That is l F ≤ 2 l . Next it can be known from (3.11) that λ k ± to be real numbers if and only if β ≥ 2 M ( μ k ) μ k . By assumption M ( s ) > 0 , A ∗ has at most number N 0 for finite real eigenvalues, and when N 0 = 0 , β < 2 M ( μ k ) μ k , ∧ 0 = max { λ k ± | k ≤ N 0 } . The eigenvalues are complex, and
Re λ k ± = β μ k 2 , (3.14)
therefore, there exists N 1 ≥ N 0 + 1 making Re λ k ± > ∧ 0 , k > N 1 .
Let N ≥ N 1 be such that (3.12) holds, decomposing the point spectrum of A ∗
σ 1 = { λ k ± | k ≤ N } , σ 2 = { λ k ± | k ≥ N + 1 } , (3.15)
meanwhile, define the corresponding subspaces of X
X 1 = s p a n { U k ± | k ≤ N } , X 2 = s p a n { U k ± | k ≥ N + 1 } , (3.16)
there is no k such that λ k − ∈ σ 1 and λ k + ∈ σ 2 , i.e., it is impossible to have U k − ∈ X 1 and U k + ∈ X 2 , vice versa, so X 1 and X 2 are orthogonal subspaces of X. From (2.3) and (3.14), we have ∧ 1 = Re λ N + , ∧ 2 = Re λ N + 1 − ,
∧ 2 − ∧ 1 = Re ( λ N + 1 − − λ N + ) = β ( μ N + 1 − μ N ) 2 . (3.17)
Thus, (3.12) implies that A ∗ satisfies the spectral interval inequality (2.5), in conclusion, A ∗ satisfies the spectral interval condition.
The proof of Theorem 3.1 is completed.
Theorem 3.2 Suppose l be the Lipschitz constant of g i ( u , v ) ( i = 1 , 2 ) in (3.1), assume μ k ≥ 4 m 1 β 2 ≥ 4 M ( μ k ) β 2 , N 1 ∈ N be large enough, when N ≥ N 1 , the following inequalities hold, either
( μ N + 1 − μ N ) ( β − β 2 μ 1 − 4 M ( s ) ) 2 ≥ 8 2 l β 2 μ 1 − 4 M ( s ) + 1 , (3.18)
| R ( N ) − R ( N + 1 ) + ( μ N + 1 − μ N ) β 2 μ 1 − 4 M ( s ) | < 2. (3.19)
Or
( μ N + 1 − μ N ) β + 1 2 ≥ 8 2 l β 2 μ 1 − 4 M ( s ) + 1 , (3.20)
| R ( N ) − R ( N + 1 ) + μ N − μ N + 1 | < 2 , (3.21)
where
R ( N ) = β 2 μ N 2 − 4 M ( μ N ) μ N . (3.22)
Then in either case, the operator A ∗ satisfies the spectral interval condition (2.5).
Proof. Due to μ k ≥ 4 M ( μ k ) β 2 , the eigenvalues of A ∗ are all real numbers, and we know that both { λ k − } k ≥ 1 and { λ k + } k ≥ 1 are monotonically increasing sequences.
The three steps to prove Theorem 3.2 are as follows:
Step1 Setting
Z 0 = { 1 ≤ k ≤ N | β ≥ 2 M ( μ k ) μ k } , Z 1 = { k ∈ N | 0 < β < 2 M ( μ k ) μ k } . (3.23)
If k ∈ Z 0 , then λ k ± ∈ R ; and if k ∈ Z 1 , then λ k ± ∈ C .
In addition, if k ∈ Z 0 ,
0 < λ 1 − < ⋯ < λ N 0 + 1 − < β μ k / 2 < λ N 0 + 1 + < ⋯ < λ 1 + , (3.24)
where N 0 = sup Z 0 , Re λ k ± = β μ k 2 , ∀ k > N 0 .
If N 0 ≥ N , let
σ 1 = { λ j − | 1 ≤ j ≤ N } , σ 2 = { λ j + , λ k ± | 1 ≤ j ≤ N ≤ k } . (3.25)
Step 2 Consider the corresponding decomposition of X
X 1 = s p a n { U j − | 1 ≤ j ≤ N } , X 2 = s p a n { U j + , U k ± | 1 ≤ j ≤ N ≤ k } , (3.26)
the equivalent inner product ( ( U , V ) ) X on X will be given below so that X 1 , X 2 are orthogonal. Given
{ X 2 = X C ⊕ X R , X C = s p a n { U 1 + , ⋯ , U N + } , X R = s p a n { U k ± | k ≥ N + 1 } , (3.27)
and set X N = X 1 ⊕ X C .
Now we introduce two functions Φ : X N → R , Ψ : X R → R , defined as
Φ ( U , V ) = − 4 M ( s ) ( u , x ¯ ) + 2 β 2 ( D m u , D m x ¯ ) + 2 β ( ( − Δ ) − m 2 y ¯ , ( − Δ ) m 2 u ) + 2 β ( ( − Δ ) − m 2 p ¯ , ( − Δ ) m 2 x ) + 4 ( ( − Δ ) − m 2 y ¯ , ( − Δ ) − m 2 p ) − 4 M ( s ) ( v , z ¯ ) + 2 β 2 ( D m v , D m z ¯ ) + 2 β ( ( − Δ ) − m 2 w ¯ , ( − Δ ) m 2 v ) + 2 β ( ( − Δ ) − m 2 q ¯ , ( − Δ ) m 2 z ) + 4 ( ( − Δ ) − m 2 w ¯ , ( − Δ ) − m 2 q ) . Ψ ( U , V ) = 2 β 2 ( D m u , D m x ¯ ) + β ( ( − Δ ) − m 2 y ¯ , ( − Δ ) m 2 u ) + β ( ( − Δ ) − m 2 p ¯ , ( − Δ ) m 2 x ) + 4 ( ( − Δ ) − m 2 y ¯ , ( − Δ ) − m 2 p ) + 2 β 2 ( D m v , D m z ¯ ) + β ( ( − Δ ) − m 2 w ¯ , ( − Δ ) m 2 v ) + β ( ( − Δ ) − m 2 q ¯ , ( − Δ ) m 2 z ) + 4 ( ( − Δ ) − m 2 w ¯ , ( − Δ ) − m 2 q ) . (3.28)
With U = ( u , p , v , q ) , V = ( x , y , z , w ) ∈ X N or X R .
For U = ( u , p , v , q ) ∈ X N , then
Φ ( U , U ) = − 4 M ( s ) ( u , u ¯ ) + 2 β 2 ( D m u , D m u ¯ ) + 2 β ( ( − Δ ) − m 2 p ¯ , ( − Δ ) m 2 u ) + 2 β ( ( − Δ ) − m 2 p ¯ , ( − Δ ) m 2 u ) + 4 ( ( − Δ ) − m 2 p ¯ , ( − Δ ) − m 2 p ) − 4 M ( s ) ( v , v ¯ ) + 2 β 2 ( D m v , D m v ¯ ) + 2 β ( ( − Δ ) − m 2 q ¯ , ( − Δ ) m 2 v ) + 2 β ( ( − Δ ) − m 2 q ¯ , ( − Δ ) m 2 v ) + 4 ( ( − Δ ) − m 2 q ¯ , ( − Δ ) − m 2 q ) ≥ − 4 M ( s ) ( ‖ u ‖ 2 + ‖ v ‖ 2 ) + 2 β 2 ( ‖ D m u ‖ 2 + ‖ D m v ‖ 2 ) − 4 β ( ‖ D − m p ‖ ‖ D m u ‖ + ‖ D − m q ‖ ‖ D m v ‖ ) + 4 ( ‖ D − m p ‖ 2 + ‖ D − m q ‖ 2 )
≥ − 4 M ( s ) ( ‖ u ‖ 2 + ‖ v ‖ 2 ) + 2 β 2 ( ‖ D m u ‖ 2 + ‖ D m v ‖ 2 ) + 4 ( ‖ D − m p ‖ 2 + ‖ D − m q ‖ 2 ) − 4 ( ‖ D − m p ‖ 2 + ‖ D − m q ‖ 2 ) − β 2 ( ‖ D m u ‖ 2 + ‖ D m v ‖ 2 ) = β 2 ( ‖ D m u ‖ 2 + ‖ D m v ‖ 2 ) − 4 M ( s ) ( ‖ u ‖ 2 + ‖ v ‖ 2 ) ≥ ( β 2 μ 1 − 4 M ( s ) ) ( ‖ u ‖ 2 + ‖ v ‖ 2 ) . (3.29)
For any k, there is β 2 μ k ≥ 4 M ( μ k ) , and according to the initial hypothesis 0 < m 0 ≤ M ( s ) ≤ m 1 ≤ β 2 μ k 4 , that is Φ ( U , U ) ≥ 0 , Φ is positive definite.
Analogously, for U = ( u , p , v , q ) ∈ X R , then
Ψ ( U , V ) = 2 β 2 ( D m u , D m u ¯ ) + β ( ( − Δ ) − m 2 p ¯ , ( − Δ ) m 2 u ) + β ( ( − Δ ) − m 2 p ¯ , ( − Δ ) m 2 u ) + 4 ( ( − Δ ) − m 2 p ¯ , ( − Δ ) − m 2 p ) 2 β 2 ( D m v , D m v ¯ ) + β ( ( − Δ ) − m 2 q ¯ , ( − Δ ) m 2 v ) + β ( ( − Δ ) − m 2 q ¯ , ( − Δ ) m 2 v ) + 4 ( ( − Δ ) − m 2 q ¯ , ( − Δ ) − m 2 q ) ≥ 2 β 2 ( ‖ D m u ‖ 2 + ‖ D m v ‖ 2 ) − 4 ( ‖ D − m p ‖ 2 + ‖ D − m q ‖ 2 ) − β 2 ( ‖ D m u ‖ 2 + ‖ D m v ‖ 2 ) + 4 ( ‖ D − m p ‖ 2 + ‖ D − m q ‖ 2 ) ≥ β 2 μ 1 ( ‖ u ‖ 2 + ‖ v ‖ 2 ) . (3.30)
that is Ψ ( U , U ) ≥ 0 , Ψ is positive definite.
Specify the inner product of X:
( ( U , V ) ) X = Φ ( P N U , P N V ) + Ψ ( P R U , P R V ) , (3.31)
where P N and P R are projections of X to X N and X R respectively, for briefly, (3.31) can be abbreviated as the following
( ( U , V ) ) X = Φ ( U , V ) + Ψ ( U , V ) .
In the inner product of X, to prove that X 1 and X 2 are orthogonal, as long as X 1 and X C are proved to be orthogonal, i.e.,
( ( U j − , U j + ) ) X = Φ ( U j − , U j + ) = 0 ( U j − ∈ X 1 , U j + ∈ X C ) . (3.32)
Recalling (3.28)
Φ ( U j − , U j + ) = − 4 M ( μ j ) ( u j , u ¯ j ) + 2 β 2 ( D m u j , D m u ¯ j ) − 2 β λ j + ( D − m u ¯ j , D m u j ) − 2 β λ j − ( D − m u ¯ j , D m u j ) + 4 λ j − λ j + ( D − m u ¯ j , D − m u j ) − 4 M ( μ j ) ( v j , v ¯ j ) + 2 β 2 ( D m v j , D m v ¯ j ) − 2 β λ j + ( D − m v ¯ j , D m v j ) − 2 β λ j − ( D − m v ¯ j , D m v j ) + 4 λ j − λ j + ( D − m v ¯ j , D − m v j )
= − 4 M ( μ j ) ( ‖ u j ‖ 2 + ‖ v j ‖ 2 ) + 2 β 2 ( ‖ D m u j ‖ 2 + ‖ D m v j ‖ 2 ) − 2 β ( λ j + + λ j − ) ( ‖ u j ‖ 2 + ‖ v j ‖ 2 ) + 4 λ j − λ j + ( ‖ D − m u j ‖ 2 + ‖ D − m v j ‖ 2 ) = − 4 M ( μ j ) + 2 β 2 μ j − 2 β ( λ j + + λ j − ) + 4 λ j − λ j + ⋅ 1 μ j , (3.33)
according to (3.10)
λ j + + λ j − = β μ j , λ j + λ j − = M ( μ j ) μ j ,
thus, (3.33) is equivalent to
Φ ( U j − , U j + ) = − 4 M ( μ j ) + 2 β 2 μ j − 2 β ( λ j + + λ j − ) + 4 λ j − λ j + ⋅ 1 μ j = 0.
Step3 The orthogonal decomposition (2.6) has now been established. Let us prove that A ∗ satisfies the spectral interval condition (2.5) and its equivalent norm on X is shown in (3.31), for this, we must estimate Lipschitz constant l F of F in (2.2).
Recalling that
F ( U ) = ( 0 , f 1 ( x ) − g 1 ( u , v ) , 0 , f 2 ( x ) − g 2 ( u , v ) ) T ,
g i ( u , v ) : V m × V m → V m × V m is Lipschitz continuous. Assume P 1 , P 2 be the orthogonal maps of X → X 1 , X → X 2 respectively, P 1 , P 2 are their corresponding mappings on V m × V m and V 0 × V 0 , from (3.29) and (3.30), for
U = ( u , p , v , q ) ∈ X , U 1 = ( u 1 , p 1 , v 1 , q 1 ) ∈ P 1 U , U 2 = ( u 2 , p 2 , v 2 , q 2 ) ∈ P 2 U ,
then
P 1 u = u 1 , P 1 v = v 1 , P 2 u = u 2 , P 2 v = v 2 .
‖ U ‖ X 2 = Φ ∗ ( P 1 U , P 1 U ) + Ψ ∗ ( P 2 U , P 2 U ) ≥ ( β 2 μ 1 − 4 M ( s ) ) ( ‖ P 1 u ‖ 2 + ‖ P 1 v ‖ 2 ) + β 2 μ 1 ( ‖ P 2 u ‖ 2 + ‖ P 2 v ‖ 2 ) ≥ ( β 2 μ 1 − 4 M ( s ) ) ( ‖ u ‖ 2 + ‖ v ‖ 2 ) . (3.34)
Given U = ( u , p , v , q ) , V = ( u ˜ , p ˜ , v ˜ , q ˜ ) ∈ X , we get
‖ F ( U ) − F ( V ) ‖ X = ‖ g 1 ( u , v ) − g 1 ( u ˜ , v ˜ ) ‖ V m × V m + ‖ g 2 ( u , v ) − g 2 ( u ˜ , v ˜ ) ‖ V m × V m ≤ 2 l ( ‖ u − u ˜ ‖ V m + ‖ v − v ˜ ‖ V m ) ≤ 2 2 l β 2 μ 1 − 4 M ( s ) ‖ U − V ‖ X ,
thus
l F ≤ 2 2 l β 2 μ 1 − 4 M ( s ) , (3.35)
by (3.35), then the spectral interval condition (2.5) holds if
∧ 2 − ∧ 1 = λ N + 1 − − λ N − > 8 2 l β 2 μ 1 − 4 M ( s ) . (3.36)
Recalling (3.22), we have
λ N + 1 − − λ N − = R ( N ) − R ( N + 1 ) 2 + β ( μ N + 1 − μ N ) 2 , (3.37)
and
lim N → + ∞ ( R ( N ) − R ( N + 1 ) + β 2 μ 1 − 4 M ( s ) ( μ N + 1 − μ N ) ) = 0. (3.38)
For formula (3.38), in fact, setting
R ′ ( N ) = β 2 μ N − 4 M ( μ N ) μ N ( β 2 μ 1 − 4 M ( s ) ) ,
we compute
R ( N ) − R ( N + 1 ) + β 2 μ 1 − 4 M ( s ) ( μ N + 1 − μ N ) = β 2 μ 1 − 4 M ( s ) ( μ N + 1 ( 1 − R ′ ( N + 1 ) ) − μ N ( 1 − R ′ ( N ) ) ) , (3.39)
lim N → + ∞ μ N ( 1 − R ′ ( N ) ) = 0.
Consequently, (3.38) is obtained.
From the condition (3.19), it can be determined that N 1 > 0 such that for all N ≥ N 1 , and with (3.37)
∧ 2 − ∧ 1 = λ N + 1 − − λ N − ≥ μ N + 1 − μ N 2 ( β − β 2 μ 1 − 4 M ( s ) ) − 1 , (3.40)
this shows that (3.36) is established by the conditions (3.18), (3.37), and (3.40), that the Theorem 3.2 is certified completely under the previous hypothesis.
At this point, we continue to use the latter hypothesis to prove, setting
R 1 ( N ) = − R ( N ) μ N 4 ,
then (3.37) is equivalent to
∧ 2 − ∧ 1 = λ N + 1 − − λ N − = β ( μ N + 1 − μ N ) 2 + 1 2 ( μ N + 1 2 R 1 ( N + 1 ) − μ N 2 R 1 ( N ) ) , (3.41)
arranging
R 2 ( N ) = μ N 2 R 1 ( N ) ,
then (3.41) means
∧ 2 − ∧ 1 = λ N + 1 − − λ N − = β ( μ N + 1 − μ N ) 2 + R 2 ( N + 1 ) − R 2 ( N ) 2 , (3.42)
making
R 3 ( N ) = 1 − R 2 ( N ) μ N ,
this implies
R 2 ( N + 1 ) − R 2 ( N ) + μ N − μ N + 1 = μ N R 3 ( N ) − μ N + 1 R 3 ( N + 1 ) , (3.43)
from (3.43), we easily get
lim N → + ∞ R 2 ( N + 1 ) − R 2 ( N ) + μ N − μ N + 1 = 0.
Namely
| R ( N ) − R ( N + 1 ) + μ N − μ N + 1 | < 2 ,
to be specific
− 2 < R ( N ) − R ( N + 1 ) + μ N − μ N + 1 < 2 ,
− 2 − ( μ N − μ N + 1 ) < R ( N ) − R ( N + 1 ) < 2 − ( μ N − μ N + 1 ) ,
therefore
∧ 2 − ∧ 1 = λ N + 1 − − λ N − = β ( μ N + 1 − μ N ) 2 + R ( N ) − R ( N + 1 ) 2 > β 2 ( μ N + 1 − μ N ) − 1 2 ( μ N − μ N + 1 ) − 1 > β + 1 2 ( μ N + 1 − μ N ) − 1 ≥ 8 2 l β 2 μ 1 − 4 M ( s ) ≥ 4 l F . (3.44)
Under the latter assumption, Theorem 3.2 is proved completely.
Theorem 3.3 In the conclusions of Theorem 3.1 and Theorem 3.2, initial boundary value problems (1.1)-(1.5) admits an inertial manifold μ in X of the form
μ = g r a p h ( Γ ) = { ξ + Γ ( ξ ) : ξ ∈ X 1 } , (3.45)
where Γ : X 1 → X 2 is Lipschitz continuous with the Lipschitz constant l F , and g r a p h ( Γ ) represents the diagram of Γ .
The authors express their sincere thanks to the anonymous referee for his/her careful reading of the paper, giving valuable suggestions and comments, which have greatly improved the presentation of this paper.
Lin, G.G. and Hu, L.J. (2018) The Inertial Manifold for a Class of Nonlinear Higher-Order Coupled Kirchhoff Equations with Strong Linear Damping. International Journal of Modern Nonlinear Theory and Application, 7, 35-47. https://doi.org/10.4236/ijmnta.2018.72003