Skip to main content

Fourier spectral method for the modified Swift-Hohenberg equation

Abstract

In this paper, we consider the Fourier spectral method for numerically solving the modified Swift-Hohenberg equation. The semi-discrete and fully discrete schemes are established. Moreover, the existence, uniqueness and the optimal error bound are also considered.

1 Introduction

In [1], Doelman et al. studied the modified Swift-Hohenberg equation

u t =k ( 1 + Δ ) 2 u+μub | u | 2 u 3 .
(1)

Setting a=kμ, considering (1) in 1D case, we find that

u t +k u x x x x +2k u x x +au+b | u x | 2 + u 3 =0,(x,t)(0,1)×(0,T).
(2)

On the basis of physical considerations, as usual, Eq. (2) is supplemented with the following boundary value conditions:

u(x,t)= u x x (x,t)=0,x=0,1
(3)

and the initial condition

u(x,0)= u 0 (x),x(0,1),
(4)

where k>0 and a, b are constants. u 0 (x) is a given function from a suitable phase space.

The Swift-Hohenberg equation is one of the universal equations used in the description of pattern formation in spatially extended dissipative systems (see [2]), which arise in the study of convective hydrodynamics [3], plasma confinement in toroidal devices [4], viscous film flow and bifurcating solutions of the Navier-Stokes [5]. Note that the usual Swift-Hohenberg equation [3] is recovered for b=0. The additional term b | u x | 2 , reminiscent of the Kuramoto-Sivashinsky equation, which arises in the study of various pattern formation phenomena involving some kind of phase turbulence or phase transition (see [68]), breaks the symmetry uu.

During the past years, many authors have paid much attention to the Swift-Hohenberg equation (see, e.g., [3, 9, 10]). However, only a few people have been devoted to the modified Swift-Hohenberg equation. It was Doelman et al. [1] who first studied the modified Swift-Hohenberg equation for a pattern formation system with two unbounded spatial directions that are near the onset to instability. Polat [7] also considered the modified Swift-Hohenberg equation. In his paper, the existence of a global attractor is proved for the modified Swift-Hohenberg equation as (2)-(4). Recently, Song et al. [2] studied the long time behavior for the modified Swift-Hohenberg equation in an H k (k0) space. By using an iteration procedure, regularity estimates for the linear semigroups and a classical existence theorem of a global attractor, they proved that problem (2)-(4) possesses a global attractor in the Sobolev space H k for all k0, which attracts any bounded subset of H k (Ω) in the H k -norm.

The spectral methods are essentially discretization methods for the approximate solution of partial differential equations. They have the natural advantage in keeping the physical properties of primitive problems. During the past years, many papers have already been published to study the spectral methods, for example, [1114]. However, for the other boundary condition, can we also use the Fourier spectral method? The answer is ‘Yes’. Choose a good finite dimensional subspace S N (here, we set S N =span{sinkπx;k=0,1,,N}), we can also have the basic results for the orthogonal projecting operator P N .

In this paper, we consider the Fourier spectral method for the modified Swift-Hohenberg equation. The existence of a solution locally in time is proved by the standard Picard iteration, global existence results are obtained by proving a priori estimate for the appropriate norms of u(x,t). Adjusted to our needs, the results are given in the following form.

Theorem 1.1 Assume that u 0 H E 2 (0,1)={v;v H 2 (0,1),v(0,t)=v(1,t)=0} and b 2 8k, then there exists a unique global solution u(x,t) of the problem (2)-(4) for all T0 such that

u(x,t) L ( 0 , T ; H E 2 ( 0 , 1 ) ) L 2 ( 0 , T ; H 4 ( 0 , 1 ) ) .

Furthermore, it satisfies

( u t , v ) + k ( u x x , v x x ) 2 k ( u x , v x ) + γ ( u , v ) + b ( | u x | 2 , v ) + ( u 3 , v ) = 0 , v H E 2 ( 0 , 1 ) , ( u ( , 0 ) , v ) = ( u 0 , v ) , v H E 2 ( 0 , 1 ) .
(5)

This paper is organized as follows. In the next section, we consider a semi-discrete Fourier spectral approximation, prove its existence and uniqueness of the numerical solution and derive the error bound. In Section 3, we consider the full-discrete approximation for problem (2)-(4). Furthermore, we prove convergence to the solution of the associated continuous problem. In the last section, some numerical experiments which confirm our results are performed.

Throughout this paper, we denote L 2 , L p , L , H k norm in Ω simply by , p , and H k .

2 Semi-discrete approximation

In this section, we consider the semi-discrete approximation for problem (2)-(4). First of all, we recall some basic results on the Fourier spectral method which will be used throughout this paper. For any integer N>0, introduce the finite dimensional subspace of H E 2 (0,1)

S N =span{sinkπx;k=1,,N}.

Let P N : L 2 (0,1) S N be an orthogonal projecting operator which satisfies

(u P N u,v)=0,v S N .
(6)

For the operator P N , we have the following result (see [13, 15]):

(B1) P N commutes with derivation on H E 2 (0,1), i.e.,

P N u x x = ( P N u ) x x ,u H E 2 (0,1).

Using the same method as [15, 16], we can obtain the following result (B2) for problem (2)-(4):

(B2) For any real 0μ2, there is a constant c such that

u P N u μ c N μ 2 u x x ,u H E 2 (0,1).

Define the Fourier spectral approximation: Find u N (t)= j = 1 N a j (t)cosjπx S N such that

( u N t , v N ) + k ( u N x x , v N x x ) + 2 k ( u N , v N x x ) + a ( u N , v N ) + b ( | u N x | 2 , v N ) + ( u N 3 , v N ) = 0 , v N S N
(7)

for all T0 with u N (0)= P N u 0 .

Now, we are going to establish the existence, uniqueness etc. of the Fourier spectral approximation solution u N (t) for all T0.

Lemma 2.1 Let u 0 L 2 (0,1) and b 2 8k, then problem (7) has a unique solution u N (t) satisfying the following inequalities:

u N ( t ) 2 c 1 u 0 2 , 0 T u N x x ( τ ) 2 dτ c 1 u 0 2 ,
(8)

where c 1 = e ( | a | + 2 ) T and c 1 = 8 ( | a | + 2 ) T c 1 + 8 8 k b 2 for all T0.

Proof Set v N =cosjπx in (7) for each j (1jN) to obtain

d d t a j (t)= f j ( a 1 ( t ) , a 2 ( t ) , , a N ( t ) ) ,j=1,2,,N,
(9)

where all f j : R N R (1jN) are smooth and locally Lipschitz continuous. Noticing that u N (0)= P N u 0 , then

a j (0)=( u 0 ,cosjπx),j=1,2,,N.
(10)

Using the theory of initial-value problems of the ordinary differential equations, there is a time T N >0 such that the initial-value problem (9)-(10) has a unique smooth solution ( a 1 (t),, a N (t)) for t[0, T N ].

Setting v N = u N in (7), we have

1 2 d d t u N 2 +k u N x x 2 + u N 4 4 |a| u N 2 +2k u N x 2 +|b| ( | u N x | 2 , u N ) .
(11)

Noticing that

|b| ( | u N x | 2 , u N ) = | b | 2 ( u N 2 , u N x x ) u N 4 4 + b 2 16 u N x x 2

and

2k u N x 2 =2k( u N , u N x x ) k 2 u N x x 2 +2k u N 2 .

Summing up, we get

d d t u N 2 + ( k b 2 8 ) u N x x 2 2 ( | a | + 2 k ) u N 2 .
(12)

Using Gronwall’s inequality, we deduce that

u N 2 e 2 ( | a | + 2 k ) t u N ( 0 ) 2 e 2 ( | a | + 2 k ) T u 0 2 = c 1 u 0 2 .

Integrating (12) from 0 to t, we derive that

u N 2 u N ( 0 ) 2 + ( k b 2 8 ) 0 T u N x x 2 dt ( 2 | a | + 4 k ) 0 T u N 2 dt.

Hence

( k b 2 8 ) 0 T u N x x 2 dt2 ( | a | + 2 k ) c 1 T u 0 2 + u 0 2 .

From the above inequality, we obtain the second inequality of (8) immediately. Therefore, Lemma 2.1 is proved. □

Lemma 2.2 Let u 0 H 0 1 (0,1) and b 2 8k, then the solution u N (t) of problem (7) satisfying

u N x ( t ) 2 c 2 , 0 T u n x x x ( τ ) 2 dτ c 2
(13)

for all T0, where c 2 and c 2 are positive constants depending only on k, a, b, T and u 0 H 1 .

Proof Setting v N = u N x x in (7), we obtain

1 2 d d t u N x 2 + k u N x x x 2 + a u N x 2 = 2 k u N x x 2 + b ( | u N x | 2 , u N x x ) + ( u N 3 , u N x x ) .
(14)

Notice that

( u N 3 , u N x x ) = ( 3 u N 2 u N x , u N x ) =3 u N u N x 2 0

and

2k u N x x 2 =2k( u N x , u N x x x ) k 4 u N x x x 2 +4 u N x 2 .

On the other hand, by Nirenberg’s inequality, we have

u N x 4 C u N x x x 5 12 u N 7 12 ,

where C is a positive constant independent of N. Hence

| b | ( | u N x | 2 , u N x x ) | b | 2 u N x x 2 + | b | 2 u N x 4 4 k 8 u N x x x 2 + | b | 2 2 k u N x 2 + C 4 u N x x x 5 3 u N 7 3 k 4 u N x x x 2 + | b | 2 2 k u N x 2 + c ( u N ) k 4 u N x x x 2 + | b | 2 2 k u N x 2 + c ( a , k , T , u 0 ) .
(15)

Summing up, we get

d d t u N x 2 +k u N x x x 2 ( 2 | a | + 8 + b 2 k ) u N x 2 +2c ( a , k , T , u 0 ) .
(16)

Using Gronwall’s inequality, we immediately obtain

u N x 2 e ( 2 | a | + 8 + b 2 k ) t u N x ( 0 ) 2 + 2 k c ( a , k , T , u 0 ) 2 | a | k + b 2 + 8 k e ( 2 | a | + 8 + b 2 k ) T u x 0 2 + 2 k c ( a , k , T , u 0 ) 2 | a | k + b 2 + 8 k c 2 ( k , a , b , T , u 0 H 1 ) .

Integrating (16) from 0 to t, we obtain

0 T u N x x x ( τ ) 2 d τ 1 k ( ( 2 | a | + 8 + b 2 k ) c 2 T + 2 c T + u x 0 2 ) = c 2 ( k , a , b , T , u 0 H 1 ) .

Then Lemma 2.2 is proved. □

Lemma 2.3 Let u 0 H E 2 (0,1) and b 2 8k, then the solution u N (t) of problem (7), satisfying

u N x x ( t ) 2 c 3 , 0 T u N x x x x ( τ ) 2 dτ c 3
(17)

for all T0, where c 3 and c 3 are positive constants, depending only on k, a, b, T and u 0 H 2 .

Proof Setting v N = u N x x x x in (7), we obtain

1 2 d d t u N x x 2 + k u N x x x x 2 = 2 u N x x x 2 a u N x x 2 b ( | u N x | 2 , u N x x x x ) ( u N 3 , u N x x x x ) .
(18)

Using Nirenberg’s inequality, we obtain

u N 6 C u N x x x x 1 12 u N 11 12 , u N x 4 C u N x x x x 1 12 u N x 11 12 ,

where C>0 is a constant depending only on the domain. Therefore

| b | ( | u N x | 2 , u N x x x x ) k 10 u N x x x x 2 + 5 b 2 2 k u N x 4 4 k 10 u N x x x x 2 + k 10 u N x x x x 2 + c ( a , k , b , T , u 0 H 1 )

and

( u N 3 , u N x x x x ) k 10 u N x x x x 2 + 5 2 k u N 6 6 k 10 u N x x x x 2 + k 10 u N x x x x 2 + c ( a , k , T , u 0 ) .

On the other hand, we have

2k u N x x x 2 =2k( u N x x , u N x x x x ) k 10 u N x x x x 2 +10 u N x x 2 .

Summing up, we get

d d t u N x x 2 +k u N x x x x 2 ( 20 + 2 | a | ) u N x x 2 +4c ( a , k , b , T , u 0 H 1 ) .
(19)

Using Gronwall’s inequality, we have

u N x x 2 e ( 20 + 2 | a | ) t u N x x ( 0 ) 2 + 2 c ( a , k , b , T , u 0 H 1 ) 10 + | a | e ( 20 + 2 | a | ) T u x x 0 2 + 2 c ( a , k , b , T , u 0 H 1 ) 10 + | a | = c 3 ( a , k , b , T , u x x 0 ) .
(20)

Integrating (19) from 0 to t, we obtain

0 T u N x x x x ( τ ) 2 d τ 1 k ( ( 20 + 2 | a | ) c 3 T + 4 T c ( a , k , b , T , u 0 H 1 ) + u N x x ( 0 ) 2 ) = c 3 ( k , a , b , T , u 0 H 2 ) .

Therefore, Lemma 2.3 is proved. □

Remark 2.1 Basing on the above Lemmas 2.1-2.3, we can get the H 2 -norm estimate for problem (7). Then, by Sobolev’s embedding theorem, we immediately conclude that

sup x [ 0 , 1 ] | u N ( x , t ) | = u N ( x , t ) c 4 ( k , a , b , T , u 0 H 1 ) ,
(21)
sup x [ 0 , 1 ] | u N x ( x , t ) | = u N x ( x , t ) c 5 ( k , a , b , T , u 0 H 2 ) .
(22)

Now, we give the following theorem.

Theorem 2.1 Suppose that u 0 H E 2 (0,1) and b 2 8k. Suppose further that u(x,t) is the solution of problem (2)-(4) and u N (x,t) is the solution of semi-discrete approximation (7). Then there exist a constant c depending on k, a, b, T and u 0 H 2 such that

u ( x , t ) u N ( x , t ) c ( N 2 + u 0 u N ( 0 ) ) .

Proof Denote η N =u(t) P N u(t) and e N = P N u(t) u N (t). From (2) and (7), we get

( e N t , v N ) + k ( e N x x , v N x x ) 2 k ( e N x , v N x ) + a ( e N , v N ) + b ( | u x | 2 | u N x | 2 , v N ) + ( u 3 u N 3 , v N ) = 0 , v N S N .
(23)

Set v N = e N in (23), we derive that

1 2 d d t e N 2 + k e N x x 2 = 2 k e N x 2 a e N 2 b ( | u x | 2 | u N x | 2 , e N ) ( u 3 u N 3 , e N ) .

By Theorem 1.1, we have

sup x [ 0 , 1 ] | u ( x , t ) | c 6 ( k , a , b , u 0 H 1 ) .

Then

( u 3 u N 3 , e N ) = ( ( e N + η N ) ( u 2 + u u N + u N 2 ) , e N ) sup x [ 0 , 1 ] ( | u | 2 + | u u N | + | u N | 2 ) ( e N 2 + η N e N ) ( c 4 2 + c 4 c 6 + c 6 2 ) ( e N 2 + 1 2 e N 2 + 1 2 η N 2 ) .

By Theorem 1.1, we have

sup x [ 0 , 1 ] | u x ( x , t ) | c 7 ( k , a , b , u 0 H 2 ) , u x x ( x , t ) 2 c 8 ( k , a , b , u 0 H 2 ) .
(24)

Using Sobolev’s embedding theorem, we have

sup x [ 0 , 1 ] | e N x | C e N H 2 C e N x x ,
(25)

where C and C are positive constants depending only on the domain. Then, using the method of integration by parts, we have

b ( | u x | 2 | u N x | 2 , e N ) = b ( ( e N x + η N x ) ( u x + u N x ) , e N ) = b ( ( e N + η N ) ( u x + u N x ) , e N x ) + b ( ( e N + η N ) ( u x x + u N x x ) , e N ) .
(26)

Hence, by (24)-(26) and Hölder’s inequality, we get

b ( | u x | 2 | u N x | 2 , e N ) | b | sup x [ 0 , 1 ] | u x + u N x | ( e N e N x + η N e N x ) + | b | sup x [ 0 , 1 ] | e N | e N + η N u x x + u N x x | b | sup x [ 0 , 1 ] | u x + u N x | ( e N e N x + η N e N x ) + C | b | e N x e N + η N u x x + u N x x ( c 7 + c 5 ) | b | ( e N x 2 + 1 2 e N 2 + 1 2 η N 2 ) + 2 C ( c 3 + c 8 ) | b | ( ε e N x 2 + 1 2 ε e N 2 + 1 2 ε η N 2 ) ( c 7 + c 5 ) | b | ( ε e N x x 2 + ( 1 2 + 1 4 ε ) e N 2 + 1 2 η N 2 ) + 2 C ( c 3 + c 8 ) | b | ( ε e N x x 2 + ( 1 2 ε + ε 4 ) e N 2 + 1 2 ε η N 2 ) ,

where ε R + is a constant. Summing up, we get

d d t e N 2 + 2 ( k [ ( c 5 + c 7 ) | b | + 2 C ( c 3 + c 8 ) | b | ] ε ) e N x x 2 4 k e N x 2 + 2 c 9 e N 2 + 2 c 10 η N 2 = 4 k ( e N , e N x x ) + 2 c 9 e N 2 + 2 c 10 η N 2 ε e N x x 2 + ( 4 k 2 ε + 2 c 9 ) e N 2 + 2 c 10 η N 2 ,
(27)

where

c 9 = | γ | + 3 2 ( c 4 2 + c 4 c 6 + c 6 2 ) + ( 1 2 + 1 4 ε ) ( c 5 + c 7 ) | b | + ( 1 ε + ε 2 ) C ( c 3 + c 8 ) | b | , c 10 = 1 2 ( c 4 2 + c 4 c 6 + c 6 2 ) + 1 2 ( c 7 + c 5 ) | b | + 1 ε C ( c 3 + c 8 ) | b | .

From Theorem 1.1 and (B2), we have

η N c N 2 u x x c 11 ( k , a , b , T , u 0 H 2 ) N 2 .
(28)

Then a simple calculation shows that

d d t e N 2 + [ 2 ( k [ ( c 5 + c 7 ) | b | + 2 C ( c 3 + c 8 ) | b | ] ε ) ε ] e N x x 2 ( 4 k 2 ε + 2 c 9 ) e N 2 + 2 c 10 η N 2 ( 4 k 2 ε + 2 c 9 ) e N 2 + 2 c 10 c 11 2 N 4 ,
(29)

where ε is small enough, it satisfies 2(k[( c 5 + c 7 )|b|+2C( c 3 + c 8 )|b|]ε)ε>0. Therefore, by Gronwall’s inequality, we deduce that

e N 2 e ( 2 ε + 2 c 9 ) T e N ( 0 ) 2 + c 10 c 11 2 ( 1 + c 9 ) ε N 4 .
(30)

Hence, the proof is completed. □

3 Fully discrete scheme

In this section, we set up a full-discretization scheme for problem (2)-(4) and consider the fully discrete scheme which implies the pointwise boundedness of the solution.

Let Δt be the time-step. The full-discretization spectral method for problem (2)-(4) is read as: find u N j S N (j=0,1,2,,N) such that for any v N S N , the following holds:

( u N j + 1 u N j Δ t , v N ) + k ( u ¯ N x x j + 1 2 , v N x x ) 2 k ( u ¯ N x j + 1 2 , v N x ) + a ( u ¯ N j + 1 2 , v N ) + b ( ( u ¯ N x j + 1 2 ) 2 , v N ) + ( ( u ¯ N j + 1 2 ) 3 , v N ) = 0 ,
(31)

with u N (0)= P N u 0 , where u ¯ N j + 1 2 = 1 2 ( u N j + u N j + 1 ).

The solution u N j has the following property.

Lemma 3.1 Assume that u 0 H E 2 (0,1) and b 2 8k. Suppose that u N j is a solution of problem (31), then there exist positive constants c 12 , c 13 , c 14 , c 15 , c 16 depending only on k, a, b, T and u 0 H 2 such that

u N j c 12 , u N x j c 13 , u N x x j c 14 .

Furthermore, we have

sup x [ 0 , 1 ] | u N j | c 15 , sup x [ 0 , 1 ] | u N x j | c 16 .

Proof It can be proved the same as Lemmas 2.1-2.3. Since the proof is so easy, we omit it. □

In the following, we analyze the error estimates between numerical solution u N j and exact solution u( t j ). According to the properties of the projection operator P N , we only need to analyze the error between P N u( t j ) and u N j . Denoted by u j =u( t j ), e j = P N u j u N j and η j = u j P N u j . Therefore

u j u N j = η j + e j .

If no confusion occurs, we denote the average of the two instant errors e n and e n + 1 by e ¯ n + 1 2 , where e ¯ n + 1 2 = e n + e n + 1 2 . On the other hand, we let η ¯ j + 1 2 = η j + η j + 1 2 .

Firstly, we give the following error estimates for the full discretization scheme.

Lemma 3.2 For the instant errors e j + 1 and e j , we have

e j + 1 2 e j 2 + 2 Δ t ( u t ( t j + 1 2 ) u N j + 1 u N j Δ t , e ¯ j + 1 2 ) + 1 320 ( Δ t ) 4 t j t j + 1 u t t t 2 d t + Δ t e ¯ j + 1 2 2 .
(32)

Proof Applying Taylor’s expansion about t j + 1 2 , using Hölder’s inequality, we can prove the lemma immediately. Since the proof is the same as [11], we omit it. □

Taking the inner product of (2) with e ¯ j + 1 2 , and letting t= t j + 1 2 , we obtain

( u t j + 1 2 , e ¯ j + 1 2 ) + ( k u x x j + 1 2 , e ¯ x x j + 1 2 ) 2 k ( u x j + 1 2 , e ¯ x j + 1 2 ) + a ( u j + 1 2 , e ¯ j + 1 2 ) + b ( ( u x j + 1 2 ) 2 , e ¯ j + 1 2 ) + ( ( u j + 1 2 ) 3 , e ¯ j + 1 2 ) = 0 .

Taking v N = e ¯ n + 1 2 in (31), we obtain

( u N j + 1 u N j Δ t , e ¯ j + 1 2 ) + k ( u ¯ N x x j + 1 2 , e ¯ x x j + 1 2 ) 2 k ( u ¯ N x j + 1 2 , e ¯ x j + 1 2 ) + a ( u ¯ N j + 1 2 , e ¯ j + 1 2 ) + b ( ( u ¯ N x j + 1 2 ) 2 , e ¯ j + 1 2 ) + ( ( u ¯ N j + 1 2 ) 3 , e ¯ j + 1 2 ) = 0 .

Comparing the above two equations, we get

( u t j + 1 2 u N j + 1 u N j Δ t , e ¯ j + 1 2 ) = k ( u x x j + 1 2 u ¯ N x x j + 1 2 , e ¯ x x j + 1 2 ) + 2 k ( u x j + 1 2 u ¯ N x j + 1 2 , e ¯ x j + 1 2 ) a ( u j + 1 2 u ¯ N j + 1 2 , e ¯ j + 1 2 ) b ( ( u x j + 1 2 ) 2 ( u ¯ N x j + 1 2 ) 2 , e ¯ j + 1 2 ) ( ( u j + 1 2 ) 3 ( u ¯ N j + 1 2 ) 3 , e ¯ j + 1 2 ) .

So, we investigate the error estimates of the five items on the right-hand side of the previous equation.

Lemma 3.3 Suppose that u 0 H E 2 (0,1) and b 2 8k, u is the solution for problem (2)-(4) and u N j is the solution for problem (31), then

k ( u x x j + 1 2 u ¯ N x x j + 1 2 , e ¯ x x j + 1 2 ) k 2 e ¯ x x j + 1 2 2 + k ( Δ t ) 3 192 t j t j + 1 u x x t t 2 dt.

Proof Using Taylor’s expansion, we obtain

u j = u j + 1 2 Δ t 2 u t j + 1 2 + t j t j + 1 2 ( t t j ) u t t d t , u j + 1 = u j + 1 2 + Δ t 2 u t j + 1 2 + t j + 1 2 t j + 1 ( t j t ) u t t d t .

Hence

1 2 ( u j + u j + 1 ) u j + 1 2 = 1 2 ( t j t j + 1 2 ( t t j ) u t t d t + t j + 1 2 t j + 1 ( t j t ) u t t d t ) .

By Hölder’s inequality, we have

u x x j + 1 2 1 2 ( u x x j + u x x j + 1 ) 2 = 1 4 ( t j t j + 1 2 ( t t j ) u t t d t + t j + 1 2 t j + 1 ( t j t ) u t t d t ) x x 2 ( Δ t ) 3 96 t j t j + 1 u x x t t 2 d t .

Noticing that ( η ¯ x x j + 1 2 , e ¯ x x j + 1 2 )=0. Therefore

( u x x j + 1 2 u ¯ N x x j + 1 2 , e ¯ x x j + 1 2 ) = ( u x x j + 1 2 u x x j + u x x j + 1 2 , e ¯ x x j + 1 2 ) ( u x x j + 1 + u x x j 2 u N x x j + 1 + u N x x j 2 , e ¯ x x j + 1 2 ) u x x j + 1 2 u x x j + u x x j + 1 2 e ¯ x x j + 1 2 ( η ¯ x x j + 1 2 + e ¯ x x j + 1 2 , e ¯ x x j + 1 2 ) ( ( Δ t ) 3 96 t j t j + 1 u x x t t 2 d t ) 1 2 e ¯ x x j + 1 2 ( η ¯ x x j + 1 2 , e ¯ x x j + 1 2 ) e ¯ x x j + 1 2 2 ( Δ t ) 3 192 t j t j + 1 u x x t t 2 d t 1 2 e ¯ x x j + 1 2 2 .

Then Lemma 3.3 is proved. □

Lemma 3.4 Suppose that u 0 H E 2 (0,1) and b 2 8k, u is the solution for problem (2)-(4) and u N j is the solution for problem (31), then

2 k ( u x j + 1 2 u ¯ N x j + 1 2 , e ¯ x j + 1 2 ) k 16 e ¯ x x j + 1 2 2 + 192 k e ¯ j + 1 2 2 + 192 k c 11 N 4 + k ( Δ t ) 3 2 t j t j + 1 u t t 2 d t ,

where c 11 is the same constant as (28).

Proof Noticing that η ¯ j + 1 2 c N 2 . Hence

2 ( u x j + 1 2 u ¯ N x j + 1 2 , e ¯ x j + 1 2 ) = 2 ( u j + 1 2 u ¯ N j + 1 2 , e ¯ x x j + 1 2 ) = 2 ( u j + 1 2 u ¯ j + 1 2 , e ¯ x x j + 1 2 ) 2 ( u j + 1 + u j 2 u N j + 1 + u N j 2 , e ¯ x x j + 1 2 ) 2 u j + 1 2 u j + 1 + u j 2 e ¯ x x j + 1 2 + 2 u j + 1 + u j 2 u N j + 1 + u N j 2 e ¯ x x j + 1 2 2 t j t j + 1 2 ( t t j ) u t t d t + t j + 1 2 t j + 1 ( t j + 1 t ) u t t d t e ¯ x x j + 1 2 + 2 e ¯ j + 1 2 + η ¯ j + 1 2 e ¯ x x j + 1 2 2 ( ( Δ t ) 3 96 t j t j + 1 u t t 2 d t ) 1 2 e ¯ x x j + 1 2 + 4 e ¯ j + 1 2 e ¯ x x j + 1 2 + 4 η ¯ j + 1 2 e ¯ x x j + 1 2 3 ε e ¯ x x j + 1 2 2 + 4 ε ( e ¯ j + 1 2 2 + c 11 N 4 ) + ( Δ t ) 3 96 ε t j t j + 1 u t t 2 d t .

In the above inequality, setting ε= 1 48 , we get the conclusion. □

Lemma 3.5 Suppose that u 0 H E 2 (0,1) and b 2 8k, u is the solution for problem (2)-(4) and u N j is the solution for problem (31), then

a ( u j + 1 2 u ¯ N j + 1 2 , e ¯ j + 1 2 ) 4 | a | e ¯ j + 1 2 2 + | a | c 11 N 4 + | a | ( Δ t ) 3 384 t j t j + 1 u t t 2 d t ,

where c 11 is the same constant as (28).

Proof We have

a ( u j + 1 2 u ¯ N j + 1 2 , e ¯ j + 1 2 ) = a ( u j + 1 2 u j + 1 + u j 2 , e ¯ j + 1 2 ) + a ( u j + 1 u j 2 u ¯ N j + 1 2 , e ¯ j + 1 2 ) | a | u j + 1 2 u j + 1 + u j 2 e ¯ j + 1 2 + | a | u j + 1 u j 2 u ¯ N j + 1 2 e ¯ j + 1 2 | a | t j t j + 1 2 ( t t j ) u t t d t + t j + 1 2 t j + 1 ( t j + 1 t ) u t t d t e ¯ j + 1 2 + | a | e ¯ j + 1 2 + η ¯ j + 1 2 e ¯ j + 1 2 | a | ( ( Δ t ) 3 96 t j t j + 1 u t t 2 d t ) 1 2 e ¯ j + 1 2 + 2 | a | ( e ¯ j + 1 2 2 + η ¯ j + 1 2 e ¯ j + 1 2 ) 4 | a | e ¯ j + 1 2 2 + | a | c 11 N 4 + | a | ( Δ t ) 3 384 t j t j + 1 u t t 2 d t .

Then Lemma 3.5 is proved. □

Lemma 3.6 Suppose that u 0 H E 2 (0,1) and b 2 8k, u is the solution for problem (2)-(4) and u N j is the solution for problem (31), then

b ( ( u x j + 1 2 ) 2 ( u ¯ N x j + 1 2 ) 2 , e ¯ j + 1 2 ) k 16 e ¯ x x j + 1 2 2 + c 7 | b | ( Δ t ) 3 96 t j t j + 1 u x t t 2 d t + c 17 c 11 N 4 + c 18 e ¯ j + 1 2 2 ,

where C R + , c 17 = c 7 |b|+ c 16 |b|+C|b|( c 8 + c 14 ) and c 18 = c 7 |b|+ c 17 + 8 c 17 k .

Proof Notice that

sup x [ 0 , 1 ] | u x ( x , t ) | c 7 , u x x ( x , t ) 2 c 8 , u N x x j c 14 , sup x [ 0 , 1 ] | u N x j | c 16 .

Hence

b ( ( u x j + 1 2 ) 2 ( u ¯ N x j + 1 2 ) 2 , e ¯ j + 1 2 ) = b ( ( u x j + 1 2 + u x j + 1 + u x j 2 ) ( u x j + 1 2 u x j + 1 + u x j 2 ) , e ¯ j + 1 2 ) b ( ( u x j + 1 + u x j 2 + u ¯ N x j + 1 2 ) ( u x j + 1 + u x j 2 u ¯ N x j + 1 2 ) , e ¯ j + 1 2 ) = b ( ( u x j + 1 2 + u x j + 1 + u x j 2 ) ( u x j + 1 2 u x j + 1 + u x j 2 ) , e ¯ j + 1 2 ) + b ( ( u x j + 1 + u x j 2 + u ¯ N x j + 1 2 ) ( u j + 1 + u j 2 u N j + 1 + u N j 2 ) , e ¯ x j + 1 2 ) + b ( ( u x x j + 1 + u x x j 2 + u N x x j + 1 + u N x x j 2 ) ( u j + 1 + u j 2 u N j + 1 + u N j 2 ) , e ¯ j + 1 2 ) .
(33)

We have used the method of integration by parts in (33). Then

b ( ( u x j + 1 2 ) 2 ( u ¯ N x j + 1 2 ) 2 , e ¯ j + 1 2 ) | b | sup x [ 0 , 1 ] | u x j + 1 2 + u x j + 1 + u x j 2 | u x j + 1 2 u x j + 1 + u x j 2 e ¯ j + 1 2 + | b | sup x [ 0 , 1 ] | u x j + 1 + u x j 2 + u ¯ N x j + 1 2 | u j + 1 + u j 2 u N j + 1 + u N j 2 e ¯ x j + 1 2 + | b | sup x [ 0 , 1 ] | e ¯ j + 1 2 | u x x j + 1 + u x x j 2 + u N x x j + 1 + u N x x j 2 u j + 1 + u j 2 u N j + 1 + u N j 2 2 c 7 | b | t j t j + 1 2 ( t t j ) u x t t d t + t j + 1 2 t j + 1 ( t j + 1 t ) u x t t d t e ¯ j + 1 2 + ( c 7 | b | + c 16 | b | + C | b | ( c 8 + c 14 ) ) e ¯ j + 1 2 + η ¯ j + 1 2 e ¯ x j + 1 2 c 7 | b | ( Δ t ) 3 96 t j t j + 1 u x t t 2 d t + c 7 | b | e ¯ j + 1 2 2 + ε e ¯ x x j + 1 2 2 + c 17 ( e ¯ j + 1 2 2 + η ¯ j + 1 2 2 ) + c 17 2 ε e ¯ j + 1 2 2 ε e ¯ x x j + 1 2 2 + c 7 | b | ( Δ t ) 3 96 t j t j + 1 u x t t 2 d t + ( c 7 | b | + c 17 + c 17 2 ε ) e ¯ j + 1 2 2 + c 17 c 11 N 4 .

Setting ε= k 16 in the above inequality, we obtain the conclusion. □

Lemma 3.7 Suppose that u 0 H E 2 (0,1) and b 2 8k, u is the solution for problem (2)-(4) and u N j is the solution for problem (31), then

( ( u j + 1 2 ) 3 ( u ¯ N j + 1 2 ) 3 , e ¯ j + 1 2 ) ( Δ t ) 3 96 t j t j + 1 u t t 2 dt+ c 19 e ¯ j + 1 2 2 + c 20 c 11 N 4 ,

where c 19 = 9 c 6 2 4 +3( c 6 2 + c 6 c 12 + c 12 2 ), c 20 = c 6 2 + c 6 c 12 + c 12 2 .

Proof Notice that

sup x [ 0 , 1 ] | u j | c 6 , sup x [ 0 , 1 ] | u N j | c 12 .

Hence

( ( u j + 1 2 ) 3 ( u ¯ N j + 1 2 ) 3 , e ¯ j + 1 2 ) = ( ( u j + 1 2 ) 3 ( u j + 1 + u j 2 ) 3 , e ¯ j + 1 2 ) ( ( u j + 1 + u j 2 ) 3 ( u ¯ N j + 1 2 ) 3 , e ¯ j + 1 2 ) sup x [ 0 , 1 ] | ( u j + 1 2 ) 2 + u j + 1 2 u j + 1 + u j 2 + ( u j + 1 + u j 2 ) 2 | u j + 1 2 u j + 1 + u j 2 e ¯ j + 1 2 + sup x [ 0 , 1 ] | ( u j + 1 + u j 2 ) 2 + u ¯ N j + 1 2 u j + 1 + u j 2 + ( u ¯ N j + 1 2 ) 2 | u j + 1 + u j 2 u ¯ N j + 1 2 e ¯ j + 1 2 3 c 6 2 t j t j + 1 2 ( t t j ) u t t d t + t j + 1 2 t j + 1 ( t j + 1 t ) u t t d t e ¯ j + 1 2 + ( c 6 2 + c 6 c 12 + c 12 2 ) e ¯ j + 1 2 + η ¯ j + 1 2 e ¯ j + 1 2 ( Δ t ) 3 96 t j t j + 1 u t t 2 d t + ( 9 c 6 2 4 + 3 ( c 6 2 + c 6 c 12 + c 12 2 ) ) e ¯ j + 1 2 2 + ( c 6 2 + c 6 c 12 + c 12 2 ) η ¯ j + 1 2 2 ( Δ t ) 3 96 t j t j + 1 u t t 2 d t + c 19 e ¯ j + 1 2 2 + c 20 c 11 N 4 .

Then Lemma 3.7 is proved. □

Now, we obtain the following theorem.

Theorem 3.1 Suppose that u 0 H E 2 (0,1) and b 2 8k, u(x,t) is the solution for problem (2)-(4) satisfying

u L ( 0 , T ; H 2 ( 0 , 1 ) ) , u t t L 2 ( 0 , T ; H 2 ( 0 , 1 ) ) , u t t t L 2 ( 0 , T ; L 2 ( 0 , 1 ) ) .

Suppose further that u N j is the solution for problem (31). Then if Δt is sufficiently small, there exist positive constants c 21 depending on k, a, b, T, u 0 H 2 and c 22 depending on k, a, b, T, u 0 H 2 , 0 T u t t H 2 2 dt and 0 T u t t t 2 dt such that, for j=0,1,2,,N,

e j + 1 c 21 ( N 2 + e 0 ) + c 22 ( Δ t ) 2 .

Proof By Lemmas 3.2-3.7, we obtain

e j + 1 2 e j 2 + Δ t c 23 ( e j + 1 2 + e j 2 + N 4 ) + ( Δ t ) 4 c 24 t j t j + 1 ( u t t 2 + u x t t 2 + u x x t t 2 + u t t t 2 ) d t ,

where c 23 and c 24 are positive constants depending only on k, a, b, T and u 0 H 2 . For Δt being sufficiently small such that c 24 Δt 1 2 , setting c 25 =2( c 23 + c 24 ), we get

e j + 1 2 (1+ c 25 Δt) e j 2 + c 25 ( Δ t N 4 + ( Δ t ) 4 B j ) ,

where

B j = t j t j + 1 ( u t t 2 + u x t t 2 + u x x t t 2 + u t t t 2 ) dt.

Using Gronwall’s inequality for the discrete form, we have

e j + 1 2 e c 25 ( j + 1 ) Δ t ( e 0 2 + c 25 ( j Δ t N 4 + ( Δ t ) 4 i = 0 j B i ) ) .

Direct computation shows that

i = 0 j B i 0 t j + 1 ( u t t 2 + u x t t 2 + u x x t t 2 + u t t t 2 ) dt.

Thus, Theorem 3.1 is proved. □

Furthermore, we have the following theorem.

Theorem 3.2 Suppose that u 0 H E 2 (0,1) and b 2 8k, u(x,t) is the solution for problem (2)-(4) satisfying

u L ( 0 , T ; H 2 ( 0 , 1 ) ) , u t t L 2 ( 0 , T ; H 2 ( 0 , 1 ) ) , u t t t L 2 ( 0 , T ; L 2 ( 0 , 1 ) ) .

Suppose further that u N j S N (j=0,1,2,) is the solution for problem (31) and the initial value u N 0 satisfies e 0 = P N u 0 u N 0 c N 2 u x x . Then there exist positive constants c depending on k, a, b, T, u 0 H 2 and c depending on k, a, b, T, u 0 H 2 , 0 T u t t H 2 2 dt, 0 T u t t t 2 dt such that

u ( x , t j ) u N j c N 2 + c ( Δ t ) 2 ,j=0,1,2,,N.

4 Numerical results

In this section, using the spectral method described in (31), we carry out some numerical computations to illustrate out results in previous section. The full-discretization spectral method is read as: For v l =sinlπx, l=1,,N, find

u N n = i = 1 N α i n siniπx,i=1,,N,

such that (31) holds.

Noticing that under the inner product (,), {siniπx,i=1,2,} is the system of orthogonal functions, then

0 1 sin i 1 πxsin i 2 πxdx= { 0 , i 1 i 2 , 1 2 , i 1 = i 2 .
(34)

Therefore, the terms of (31) are

( u N j + 1 u N j Δ t , v l ) = α l j + 1 α l j 2 Δ t , k ( u ¯ N x x j + 1 2 , v l x x ) = k 2 ( u N x x j + 1 + u N x x j , v l x x ) = k 4 ( l π ) 4 ( α l j + 1 + α l j ) , 2 k ( u ¯ N x j + 1 2 , v l x ) = k ( u N j + 1 + u N j , v l x x ) = k 2 ( l π ) 2 ( α l j + 1 + α l j ) , a ( u ¯ N j + 1 2 , v l ) = a 2 ( u N j + 1 + u N j , v l ) = a 4 ( α l j + 1 + α l j ) , b ( ( u ¯ N x j + 1 2 ) 2 , v l ) = b 4 ( ( u N x j + 1 + u N x j ) 2 , v l ) b ( ( u ¯ N x j + 1 2 ) 2 , v l ) = b 4 p 1 , p 2 = 1 N p 1 p 2 π ( α p 1 j + 1 α p 2 j + 1 + α p 1 j α p 2 j + 2 α p 1 j + 1 α p 2 j ) ξ p 1 p 2 l

and

( ( u ¯ N j + 1 2 ) 3 , v l ) = 1 8 p 1 , p 2 , p 3 = 1 N ( α p 1 j + 1 α p 2 j + 1 α p 3 j + 1 + 3 α p 1 j + 1 α p 2 j + 1 α p 3 j + 3 α p 1 j + 1 α p 2 j α p 3 j + α p 1 j α p 2 j α p 3 j ) η p 1 p 2 p 3 l ,

where

ξ p 1 p 2 l = 0 1 cos p 1 π x cos p 2 π x sin l π x d x , η p 1 p 2 p 3 l = 0 1 sin p 1 π x sin p 2 π x sin p 3 π x sin l π x d x .

Thus, (31) can be transformed as

α l j + 1 α l j 2 Δ t + ( k 4 ( l π ) 4 k 2 ( l π ) 2 + a 4 ) ( α l j + 1 + α l j ) + b 4 ρ l + 1 8 σ l =0,
(35)

where l=1,,N, and

ρ l = p 1 , p 2 = 1 N p 1 p 2 π 2 ( α p 1 j + 1 α p 2 j + 1 + α p 1 j α p 2 j + 2 α p 1 j + 1 α p 2 j ) ξ p 1 p 2 l , σ l = p 1 , p 2 , p 3 = 1 N ( α p 1 j + 1 α p 2 j + 1 α p 3 j + 1 + 3 α p 1 j + 1 α p 2 j + 1 α p 3 j + 3 α p 1 j + 1 α p 2 j α p 3 j + α p 1 j α p 2 j α p 3 j ) η p 1 p 2 p 3 l .

If α k n (k=1,2,,N) is known, there exists an N variable nonlinear system of equations for α l j + 1 (l=1,2,,N) which can be seen as

F ( α j + 1 ) = ( f 1 ( α j + 1 ) f 2 ( α j + 1 ) f N ( α j + 1 ) ) =0.

We use the simple Newton method to seek the solutions. Initialization yields

α ( 0 ) j + 1 = ( α 1 j , α 2 j , , α N j ) T .
(36)

The iterative formulation is as follows:

α ( k + 1 ) j + 1 = α ( k ) j + 1 + Δ α ( k ) j + 1 , F ( α ( 0 ) j + 1 ) Δ α ( k ) j + 1 + F ( α ( k ) j + 1 ) = 0 , k = 0 , 1 , 2 , , }
(37)

where F ( α ( 0 ) j + 1 ) is the N×N order Jacobi matrix for F( α j + 1 ) when α j + 1 = α ( 0 ) j + 1 ,

F ( α ( 0 ) j + 1 ) = ( 1 f 1 ( α ( 0 ) j + 1 ) , , N f 1 ( α ( 0 ) j + 1 ) 1 f N ( α ( 0 ) j + 1 ) , , N f N ( α ( 0 ) j + 1 ) ) .
(38)

Give accuracy ϵ>0, when α ( k + 1 ) j + 1 α ( k ) j + 1 <ϵ, stop the iteration, α j + 1 α ( k + 1 ) j + 1 .

As an example, we choose k=2, a=1, b=1, u 0 = ( 1 x ) 5 x 5 , Δt=0.0005, N=32, and get the solution which evolves from t=0 to t=0.025 (cf. Figure 1).

Figure 1
figure1

Expanded property of solution when N=32 , Δt=0.0005 .

Now, we consider the variation of error. Since there is no exact solution for (2)-(4) known to us, we make a comparison between the solution of (31) on coarse meshes and a fine mesh.

Choose Δt=0.001,0.001× 1 2 ,0.001× 1 4 ,0.001× 1 8 ,0.001× 1 10 ,0.001× 1 16 , respectively, to solve (31). Set u N min (x,0.1) as the solution for Δ t min =0.001× 1 32 . Denote

err(0.1,Δt)= ( 0 1 ( u N k ( x , 0.1 ) u N min ( x , 0.1 ) ) 2 d x ) 1 2 ,k=1,2,,6.
(39)

Then the error is showed in Table 1 at t=0.1.

Table 1 Errors of different time steps at t=0.1

On the other hand, choose N=24,28,32,36,40, Δ t 0 =0.001× 1 16 , respectively, to solve (31). Then the error is showed in Table 2 at t=0.1.

Table 2 Errors of different basic function numbers at t=0.1

It is easy to see that the third column err ( 0.1 , Δ t ) ( Δ t ) 2 of Table 1 is monotone decreasing along with the time step’s waning, the third column err ( 0.1 , Δ t 0 ) N 2 of Table 2 is monotone decreasing along with N’s magnifying. Hence, we can find positive constants C 1 =1.6808, C 2 =1.32× 10 7 such that

err ( 0.1 , Δ t ) ( Δ t ) 2 C,k=1,2,,6

and

err ( 0.1 , Δ t 0 ) N 2 C,N=24,28,32,36,40.

Thus, the order of error estimates is O( ( Δ t ) 2 + N 2 ) proved in Theorem 3.2.

References

  1. 1.

    Doelman A, Standstede B, Scheel A, Schneider G: Propagation of hexagonal patterns near onset. Eur. J. Appl. Math. 2003, 14: 85-110.

    MATH  Article  Google Scholar 

  2. 2.

    Song L, Zhang Y, Ma T:Global attractor of a modified Swift-Hohenberg equation in H k space. Nonlinear Anal. 2010, 72: 183-191. 10.1016/j.na.2009.06.103

    MATH  MathSciNet  Article  Google Scholar 

  3. 3.

    Swift J, Hohenberg PC: Hydrodynamics fluctuations at the convective instability. Phys. Rev. A 1977, 15: 319-328. 10.1103/PhysRevA.15.319

    Article  Google Scholar 

  4. 4.

    La Quey RE, Mahajan PH, Rutherford PH, Tang WM: Nonlinear saturation of the trapped-ion mode. Phys. Rev. Lett. 1975, 34: 391-394. 10.1103/PhysRevLett.34.391

    Article  Google Scholar 

  5. 5.

    Shlang T, Sivashinsky GL: Irregular flow of a liquid film down a vertical column. J. Phys. France 1982, 43: 459-466. 10.1051/jphys:01982004303045900

    Article  Google Scholar 

  6. 6.

    Kuramoto Y: Diffusion-induced chaos in reaction systems. Prog. Theor. Phys. Suppl. 1978, 64: 346-347.

    Article  Google Scholar 

  7. 7.

    Polat M: Global attractor for a modified Swift-Hohenberg equation. Comput. Math. Appl. 2009, 57: 62-66. 10.1016/j.camwa.2008.09.028

    MATH  MathSciNet  Article  Google Scholar 

  8. 8.

    Sivashinsky GL: Nonlinear analysis of hydrodynamic instability in laminar flames. Acta Astron. 1977, 4: 1177-1206. 10.1016/0094-5765(77)90096-0

    MATH  MathSciNet  Article  Google Scholar 

  9. 9.

    Lega J, Moloney JV, Newell AC: Swift-Hohenberg equation for lasers. Phys. Rev. Lett. 1994, 73: 2978-2981. 10.1103/PhysRevLett.73.2978

    Article  Google Scholar 

  10. 10.

    Peletier LA, Rottschäfer V: Large time behavior of solution of the Swift-Hohenberg equation. C. R. Math. Acad. Sci. Paris, Sér. I 2003, 336: 225-230. 10.1016/S1631-073X(03)00021-9

    MATH  Article  Google Scholar 

  11. 11.

    Chai S, Zou Y, Gong C: Spectral method for a class of Cahn-Hilliard equation with nonconstant mobility. Commun. Math. Res. 2009, 25: 9-18.

    MATH  MathSciNet  Google Scholar 

  12. 12.

    He Y, Liu Y: Stability and convergence of the spectral Galerkin method for the Cahn-Hilliard equation. Numer. Methods Partial Differ. Equ. 2008, 24: 1485-1500. 10.1002/num.20328

    MATH  Article  Google Scholar 

  13. 13.

    Ye X, Cheng X: The Fourier spectral method for the Cahn-Hilliard equation. Appl. Math. Comput. 2005, 171: 345-357. 10.1016/j.amc.2005.01.050

    MATH  MathSciNet  Article  Google Scholar 

  14. 14.

    Yin L, Xu Y, Huang M: Convergence and optimal error estimation of a pseudo-spectral method for a nonlinear Boussinesq equation. J. Jilin Univ. Sci. 2004, 42: 35-42.

    MATH  MathSciNet  Google Scholar 

  15. 15.

    Canuto C, Hussaini MY, Quarteroni A, Zang TA: Spectral Methods in Fluid Dynamics. Springer, New York; 1988.

    Google Scholar 

  16. 16.

    Xiang X: The Numerical Analysis for Spectral Methods. Science Press, Beijing; 2000. (in Chinese)

    Google Scholar 

Download references

Acknowledgements

This work is supported by the Graduate Innovation Fund of Jilin University (Project 20121059). The authors would like to express their deep thanks for the referee’s valuable suggestions about the revision and improvement of the manuscript.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Xiaopeng Zhao.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

XZ wrote the first draft, PZ made the figure of numerical solution and results on errors of different time steps, WZ made the results on errors of different basic function numbers, BL and FL corrected and improved the final version. All authors read and approved the final draft.

Authors’ original submitted files for images

Below are the links to the authors’ original submitted files for images.

Authors’ original file for figure 1

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.

Reprints and Permissions

About this article

Cite this article

Zhao, X., Liu, B., Zhang, P. et al. Fourier spectral method for the modified Swift-Hohenberg equation. Adv Differ Equ 2013, 156 (2013). https://doi.org/10.1186/1687-1847-2013-156

Download citation

Keywords

  • Fourier Spectral Methods
  • Swift-Hohenberg Equation
  • Fully-discrete Scheme
  • Variable Nonlinear Systems
  • Viscous Film Flow