Accepted Manuscript

Blow-up phenomena for a pseudo-parabolic equation with variable

exponents

Huafei Di, Yadong Shang, Xiaoming Peng

PII: S0893-9659(16)30248-8
DOI: http://dx.doi.org/10.1016/j.aml.2016.08.013
Reference: AML 5077

To appear in: Applied Mathematics Letters

Received date : 27 June 2016

Revised date : 17 August 2016
Accepted date : 17 August 2016

Please cite this article as: H. Di, et al., Blow-up phenomena for a pseudo-parabolic equation with
variable exponents, Applied Mathematics Letters (2016),
http://dx.doi.org/10.1016/j.aml.2016.08.013

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form.
Please note that during the production process errors may be discovered which could affect the
content, and all legal disclaimers that apply to the journal pertain.
*Manuscript
Click here to view linked References

Blow-up phenomena for a pseudo-parabolic equation

with variable exponents

Huafei Di∗, Yadong Shang, Xiaoming Peng

School of Mathematics and Information Science, Guangzhou University,
Guangzhou 510006, Guangdong, P R China

Abstract
We consider a pseudo-parabolic equation with nonlinearities of variable exponent type
ut − ν△ut − div(|∇u|m(x)−2 ∇u) = |u|p(x)−2 u, in Ω × (0, T ),
associated with initial and Dirichlet boundary conditions. By means of a differential inequality technique,
we obtain an upper bound for blow-up time if variable exponents p(·), m(·) and the initial data satisfy some
conditions. Also, a lower bound for blow-up time is determined under some other conditions.

Keywords: Pseudo-parabolic equation, Blow up, lower bound, Variable exponent

1. Introduction
In this paper, we consider the initial-boundary value problem

 m(x)−2 ∇u) = |u|p(x)−2 u, in Ω × (0, T ),
ut − ν△ut − div(|∇u|
u(x, t) = 0, on ∂Ω × (0, T ), (1.1)


u(x, 0) = u0 (x), in Ω,

where Ω is a bounded domain of Rn with a smooth boundary ∂Ω, the nonlinear term div(|∇u|m(x)−2 ∇u)
is the so called m(x)-Laplace operator, and the variable exponents p(·), m(·) are two continuous functions
on Ω such that:

2 ≤ m− ≤ m(x) ≤ m+ < p− ≤ p(x) ≤ p+ < ∞. (1.2)

It is well known that nonlinear pseudo-parabolic equations appear in the study of various problems of the
hydrodynamics, thermodynamics and filtration theory etc (see [1]-[3]). Obviously, if ν = 1, m(x) = 2,
p(x) = p, then Eq. (1.1) reduces to the following pseudo-parabolic equation

ut − △u − △ut = |u|p−2 u, in Ω × (0, T ). (1.3)

For Eq. (1.3), many results have been obtained, such as the existence and uniqueness in [1, 4], blow up in
[5]-[7], asymptotic behavior in [5, 8] and so on. Especially, Xu [5] proved that the solutions blow up in finite
time in H01 (Ω)-norm. Luo [6] obtained an upper bound and a lower bound of the blow up rate.
In the absence of the damped term (ν = 0), Eq. (1.1) becomes

ut − div(|∇u|m(x)−2 ∇u) = |u|p(x)−2 u, in Ω × (0, T ), (1.4)

Alaoui, Messaoudi and Khenous [9] proved that any solutions of this equation with nontrivial initial datum
blow up in finite time. For the constant exponents case (m(x) = m, p(x) = p), Eq. (1.4) has been extensively
studied and results concerning existence, nonexistence and asymptotic behavior have been established by
many authors [10]-[12]. For instance, Payne et al. [10, 11] obtained the upper and lower bounds on blow up
time when blow up does occur by applying the differential inequality techniques.
Let us mention that the equations with nonlinearities of variable exponent type are usually referred as
equations with nonstandard growth conditions. This type of equations occurs in the mathematical modelling

∗ Correspondingauthor.
Email address: dihuafei@yeah.net (Huafei Di )

Preprint submitted to Elsevier August 17, 2016

of various physical phenomena, e.g., the flows of electrorheological fluids, nonlinear viscoelasticity, fluids with
temperature-dependent viscosity, processes of filtration through a porous media and image processing and
so on, see [13]-[15] and the further references therein.
Our objective in this paper is to study the blow up phenomenon of solutions of the problem (1.1) in the
framework of the Lebesgue and Sobolev spaces with variable exponents. As far as we known, there is little
information on the bounds for blow up time to problem (1.1). In details, this paper is organized as follows:
in section 2, we introduce the function spaces of Orlicz-Sobolev type and present a brief description of their
main properties. In section 3, we dedicate the upper bound for blow up time to problem (1.1) under suitable
conditions on p(·), m(·) and the initial data. In section 4, the lower bound of blow up time is obtained under
some other conditions.

2. Function spaces
We recall some known results about the Lebesgue and Sobolev spaces with variable exponents (see[16]-
[18]) which will be needed in the course of this paper.
Let p(·) : Ω → [1, ∞) be a measurable function, where Ω is a domain of Rn . We denoted by p− =
ess inf x∈Ω p(x) and p+ = ess supx∈Ω p(x). The variable exponent Lebesgue space Lp(·) (Ω) consists of all
measurable functions u defined on Ω for which
∫
ρp(·) = |u(x)|p(x) dx < ∞.
Ω

The set Lp(·) (Ω)

equipped with the Luxembourg norm ∥u∥p(·) = inf{λ > 0 : ρp(·) (u/λ) ≤ 1} is a Banach
space (see[18]). The variable exponent Sobolev space W 1,p(·) (Ω) is defined by
{
W 1,p(·) (Ω) = {u ∈ Lp(·) (Ω) : |∇u|p(x) ∈ L1 (Ω)},
∥u∥W 1,p(·) (Ω) = ∥u∥1,p(·) = ∥∇u∥p(·) + ∥u∥p(·) .
1,p(·) ′
W0 (Ω) is defined as the closure in W 1,p(·) (Ω) of C0∞ (Ω). W 1,p (·) (Ω) is the dual space of W 1,p(·) (Ω)
1
where p′ (·) is the function such that p(·) + p′1(·) = 1.
Next, we present some useful lemmas which will be used later. Let the variable exponent p(·) satisfies
the Zhikov-Fan conditions:
A
|p(x) − p(y)| ≤ 1
, for all x, y ∈ Ω with |x − y| < δ, (2.1)
log( |x−y| )

where A > 0 and 0 < δ < 1.

Lemma 2.1. (see [16, 18])
(1) If Ω has a finite measure and p(·), q(·) are variable exponents so that p(x) ≤ q(x) almost everywhere in
Ω, then there is a continuous embedding from Lq(·) (Ω) ,→ Lp(·) (Ω).
1,p(·)
(2) Let the variable exponent p(·) satisfies (2.1), then ∥u∥p(·) ≤ C∥∇u∥p(·) for all u ∈ W0 (Ω), where Ω
is bounded.
(3) Let the variable exponents p(·) ∈ C(Ω), q : Ω → [1, ∞) is a measurable function and satisfy
{ np(x)
, if p(x) < n,
ess inf (p∗ (x) − q(x)) > 0, where p∗ (x) = n−p(x)
x∈Ω +∞, if p(x) ≥ n.
1,p(·)
Then, the Sobolev embedding W0 (Ω) ,→ Lq(·) (Ω) is continuous and compact.

3. Upper bound for blow-up time

In this section, we obtain an upper bound for blow-up time to the problem (1.1) if the variable exponents
p(·), m(·) and initial data satisfy some conditions. Hereafter, for simplicity, we set ν = 1. Firstly, we start
with a local existence result for the problem (1.1) which can be obtained by Faedo-Galerkin methods.
1,m(·)
Theorem 3.1. Assume that (1.2) and (2.1) hold. Then for any u0 ∈ W0 (Ω) ∩ Lp(·) (Ω), there exists
a number T0 ∈ (0, T ] such that the problem (1.1) has a unique local solution
1,m(·)
u ∈ L∞ ([0, T0 ]; W0 (Ω) ∩ Lp(·) (Ω)), ut ∈ L2 ([0, T0 ]; W01,2 (Ω)),
satisfying
1,m(·)
(ut , v) + (∇ut , ∇v) + (|∇u|m(x)−2 ∇u, ∇v) = (|u|p(x)−2 u, v), for all v ∈ W0 (Ω) ∩ Lp(·) (Ω). (3.1)

2
 Next, we seek the upper bound for the blow up of the problem (1.1).
Replacing v by ut in the Eq. (3.1), we have
∫ ∫ ∫
d 1 d 1
[|ut |2 + |∇ut |2 ]dx + |∇u|m(x) dx = |u|p(x) dx. (3.2)
Ω dt Ω m(x) dt Ω p(x)
We then define the energy by
∫ [ ]
1 1
E(t) = |∇u|m(x) − |u|p(x) dx. (3.3)
Ω m(x) p(x)
Hence, by (3.2) and (3.3), we have
∫
E ′ (t) = − [|ut |2 + |∇ut |2 ]dx ≤ 0. (3.4)
Ω

1,m(·)
Theorem 3.2. Assume that (1.2) and (2.1) hold. Let u0 ∈ W0 (Ω) ∩ Lp(·) (Ω) such that ∥u0 ∥H 1 > 0
0
and
∫ [ ]
1 1
|u0 |p(x) − |∇u0 |m(x) dx ≥ 0. (3.5)
Ω p(x) m(x)

Then, the solution u of the problem (1.1) blow up in finite time T ∗ in H01 (Ω)-norm. Moreover, an upper
bound for blow up time is given by
m−
2(F (0))1− 2
T∗ ≤ , (3.6)
(m− − 2)β

where β is a suitable positive constant given later and F (0) = ∥u0 ∥2H 1 .
0

Proof. We define an auxiliary function

∫ ∫
F (t) = u2 dx + |∇u|2 dx. (3.7)
Ω Ω

Multiplying u on two sides of the problem (1.1), and integrating by part, we have
∫ ∫ ∫ ∫
uut dx + ∇u · ∇ut dx = − |∇u|m(x)+1 dx + |u|p(x)+1 dx. (3.8)
Ω Ω Ω Ω

By differentiating F (t) with respect to t, we get

∫ ∫ ∫ ∫
F ′ (t) = 2 uut dx + 2 ∇u · ∇ut dx = 2 |u|p(x) dx − 2 |∇u|m(x) dx
Ω Ω Ω Ω
∫ [ ] ∫ [ ]
|u|p(x) |∇u|m(x) 1 1
=2 p(x) − dx + 2 p(x) − |∇u|m(x) dx. (3.9)
Ω p(x) m(x) Ω m(x) p(x)

Since E ′ (t) ≤ 0, so we nave

∫ [ ] ∫ [ ]
|u|p(x) |∇u|m(x) |u0 |p(x) |∇u0 |m(x)
p(x) − dx ≥ p(x) − dx
Ω p(x) m(x) Ω p(x) m(x)
∫ [ ]
|u0 |p(x) |∇u0 |m(x)
≥ p− − dx ≥ 0. (3.10)
Ω p(x) m(x)

By (3.9) and (3.10), we see

∫ [ ] ∫
1 1
F ′ (t) ≥ 2 p− − |∇u|m(x) dx = C0 |∇u|m(x) dx. (3.11)
Ω m+ p− Ω
 
We define the sets Ω+ = {x ∈ Ω |∇u| ≥ 1} and Ω− = {x ∈ Ω |∇u| < 1}, so we get
(∫ ∫ )
F ′ (t) ≥ C0 |∇u|m+ dx + |∇u|m− dx
Ω− Ω+
( )m+ /2 (∫ )m− /2 
∫
≥ C1  2
|∇u| dx + 2
|∇u| dx , (3.12)
Ω− Ω+

3
using the fact that ∥∇u∥2 ≤ C∥∇u∥r for all r ≥ 2. This implies that
∫ ∫
(F ′ (t))2/m+ ≥ C2 |∇u|2 dx ≥ 0 and (F ′ (t))2/m− ≥ C3 |∇u|2 dx ≥ 0. (3.13)
Ω− Ω+

The Poincáre inequality gives ∥∇u∥2 ≥ λ1 ∥u∥2 , where λ1 is the first eigenvalue of the problem
{
△w + λw = 0, in Ω,
w = 0, on ∂Ω.

Thus, we have

1 λ1 λ1
∥∇u∥2 = ∥∇u∥2 + ∥∇u∥2 ≥ ∥u∥2H 1 . (3.14)
1 + λ1 1 + λ1 1 + λ1 0

It follows from (3.13) and (3.14) that

λ1 (C3 + C2 )
(F ′ (t))2/m+ + (F ′ (t))2/m− ≥ (C3 + C2 )∥∇u∥2 ≥ ∥u∥2H 1 = C4 F (t), (3.15)
1 + λ1 0

or
( )
2( 1 − 1 )
(F ′ (t))2/m− 1 + (F ′ (t)) m+ m− C4 F (t). (3.16)

By (3.15) and the fact that F (t) ≥ F (0) > 0 (F ′ (t) ≥ 0), we have either

C4 C4 C4 C4
(F ′ (t))2/m+ ≥ F (t) ≥ F (0) or (F ′ (t))2/m− ≥ F (t) ≥ F (0), (3.17)
2 2 2 2
which implies that

(F ′ (t)) ≥ C5 (F (0))m+ /2 or (F ′ (t)) ≥ C5 (F (0))m− /2 . (3.18)

Therefore, we have that F ′ (t) ≥ α, where α = min{C5 (F (0))m+ /2 , C5 (F (0))m− /2 }. Furthermore, from
1
m
− m1 ≤ 0 and (3.16), we get
+ −

′
F (t) ≥ β(F (t))2/m− , (3.19)
( )2/m−
C4
where the constant β = 2( 1 − 1 )
. Integrating the inequality (3.19) from 0 to t, we see
m+ m−
1+α

m− m− (2 − m− )βt
(F (t))1− 2 ≤ (F (0))1− 2 + ,
2
which implies that
1
F (t) ≥ . (3.20)
[ m−
(m− −2)βt ] m− −2
2
(F (0))1− 2 + 2
m−
1−
2(F (0)) 2
Thus, (3.20) shows that F (t) blows up at some finite time T ∗ ≤ (m− −2)β
, so the solution u blows up
in H01 (Ω)-norm in finite time.

Remark 3.1. From the time estimate (3.6), we see that the larger F (0) is, the quicker the blow-up phe-
nomena takes place.

4. Lower bound for blow-up time

In this section, our aim is to determine a lower bound for blow-up time of the problem (1.1).

4
Theorem 4.1. Suppose that (1.2) and (2.1) hold. Furthermore assume that 2 < p+ < ∞ if n ≤ 2,
2n 1,m(·)
2 < p+ ≤ n−2 if n ≥ 3, u0 ∈ W0 (Ω) ∩ Lp(·) (Ω) and the solution u of the problem (1.1) becomes
unbounded at finite time T in H0 (Ω)-norm, then an lower bound T ∗ for blow up time is given by
∗ 1

∫ ∞
dη
T∗ ≥ p+ p /2 p . (4.1)
F (0) 2B+ η + + 2B−− η p− /2

where B+ , B− are the Sobolev embedding constants for H01 (Ω) ,→ Lp+ (Ω), H01 (Ω) ,→ Lp− (Ω), respectively
and F (0) = ∥u0 ∥2H 1 .
0

Proof. We define the function F (t) the same as (3.7), and compute as in the previous section,
∫ ∫
F ′ (t) = 2 uut dx + 2 ∇u · ∇ut dx
Ω Ω
∫ ∫ ∫
=2 |u|p(x) dx − 2 |∇u|m(x) dx ≤ 2 |u|p(x) dx. (4.2)
Ω Ω Ω
 
Defining the sets Ω+ = {x ∈ Ω |u| ≥ 1} and Ω− = {x ∈ Ω |u| < 1}, so we get
∫ ∫ ∫ ∫ ∫
|u|p(x) dx ≤ |u|p+ dx + |u|p− dx ≤ |u|p+ dx + |u|p− dx
Ω Ω+ Ω− Ω Ω
(∫ )p+ /2 (∫ )p− /2
p p
≤B++ |∇u|2 dx + B−− |∇u|2 dx , (4.3)
Ω Ω

where B+ , B− are the optimal constants satisfying the Sobolev embedding inequalities ∥u∥Lp+ ≤ ∥∇u∥2
and ∥u∥Lp− ≤ ∥∇u∥2 , respectively. Thus, the combination of (4.2) and (4.3) implies that
(∫ )p+ /2 (∫ )p− /2
p p p p
F ′ (t) ≤ 2B++ |∇u|2 dx + 2B−− |∇u|2 dx = 2B++ (F (t))p+ /2 + 2B−− (F (t))p− /2 . (4.4)
Ω Ω

Integrating the inequality (4.4) from 0 to t, we get

∫ F (t) dη
p p ≤ t. (4.5)
F (0) 2B++ η p+ /2 + 2B−− η p− /2

If u blows up in H01 (Ω)-norm, then we obtain a lower bound for T ∗ given by

∫ ∞
dη
T∗ ≥ p+ p /2 p . (4.6)
F (0) 2B+ η + + 2B−− η p− /2

Clearly, the integral is bound since exponents p+ ≥ p− > 2. This completes the proof of theorem 4.1.

Acknowledgements
This work is supported by the NSF of China(11401122) and the Foundation for Fostering the Scientific
and Technical Innovation of Guangzhou University.

References
[1] A.B. Al’shin, M.O. Korpusov, A.G. Siveshnikov, Blow up in nonlinear Sobolev type equations, De
Gruyter,1 edition, Berlin, 2011.

[2] E.S. Dzektser, A generalization of equations of motion of underground water with free surface, Dokl.
Akad. Nauk SSSR 202 (5) (1972) 1031-1033.

[3] M.O. Korpusov, A.G. Sveshnikov, Three-dimensional nonlinear evolution equations of pseudo-parabolic
type in problems of mathematicial physics, Comp. Math. Math. Phys. 43 (2003) 1765-1797.

[4] R.E. Showalter, Existence and representation theorem for a semilinear Sobolev equation in Banach
space, SIAM J. Math. Anal. 3 (1972) 527-543.

5
 [5] R.Z. Xu, J. Su, Global existence and finite time blow-up for a class of semilinear pseudo-parabolic
equations, J. Funct. Anal. 264 (2013) 2732-2763.

[6] P. Luo, Blow-up phenomena for a pseudo-parabolic equation, Math. Meth. Appl. Sci. 38 (2015) 2636-
2641.

[7] X.M. Peng, Y.D. Shang, X.X. Zheng, Blow-up phenomena for some nonlinear pseudo-parabolic equa-
tions, Appl. Math. Lett. 56 (2016) 17-22.

[8] Y. Liu, W.S. Jiang, F.L. Huang, Asymptotic behaviour of solutions to some pseudo-parabolic equations,
Appl. Math. Lett. 25 (2012) 111-114.

[9] M.K. Alaoui, S.A. Messaoudi, H.B. Khenous, A blow-up result for nonlinear generalized heat equation,
Comput. Math. Appl. 68 (2014) 1723-1732.

[10] L.E. Payne, G.A. Philippin, P.W. Schaefer, Blow-up phenomena for some nonlinear parabolic problems,
Nonlinear Anal. 69 (2008) 3495-3502.

[11] L.E. Payne, P.W. Schaefer, Lower bounds for blow-up time in parabolic problems under Dirichlet
conditions, J. Math. Anal. Appl. 328 (2007) 1196-1205.

[12] K. Baghaei, M.B. Ghaemi, M. Hesaaraki, Lower bounds for blow-up time in a semilinear parabolic
problem involing variable source, Appl. Math. Lett. 27 (2014) 49-52.

[13] S.N. Antontsev, J.I. Dı́az, S. Shmarev, Energy methods for free boundary problems: applications to
nonlinear PDEs and fluid mechanics, Progress in Nonlinear Differential Equations and Their Applica-
tions, Vol. 48, Bikhäuser, Boston, 2002.

[14] K. Rajagopal, M. Rúz̆ic̆ka, Mathematical modelling of electro-rheo logical fluids, Cont. Mech. Therm.
13 (2001) 59-78.

[15] M. Rúz̆ic̆ka, Electrorheological fluids, modeling and mathematical theory, in: Lecture Notes in Mathe-
matics, Vol. 1748, Springer, Berlin, 2000.

[16] X.L. Fan, J.S. Shen, D. Zhao, Sobolev embedding theorems for spaces W k,p(x) (Ω), J. Math. Anal.
Appl. 263 (2001) 749-760.

[17] J. Musielak, Orlicz Spaces and Modular Spaces, Lecture Note in Mathematics, Vol. 1034, Springer-
Verlag, Berlin, 1983.

[18] L. Diening, P. Hästo, P. Harjulehto, M. M. Rúz̆ic̆ka, Lebesgue and Sobolev Spaces with Variable
Exponents, in: Springer Lecture Notes, Vol. 2017, Springer-Verlag, Berlin, 2011.