On: 13 October 2014, At: 14:37Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street,
Trang 1On: 13 October 2014, At: 14:37
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Applicable Analysis: An International Journal
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/gapa20
Existence of traveling waves to any Lax shock satisfying Oleinik’s criterion in conservation laws
Mai Duc Thanha & Nguyen Huu Hiepb a
Department of Mathematics, International University, Vietnam National University-Ho Chi Minh City, Quarter 6, Linh Trung Ward, Thu Duc District, Ho Chi Minh City, Vietnam
b Faculty of Applied Science, University of Technology, 268 Ly Thuong Kiet str., District 10, Ho Chi Minh City, Vietnam
Published online: 30 Jul 2014
To cite this article: Mai Duc Thanh & Nguyen Huu Hiep (2014): Existence of traveling waves to any
Lax shock satisfying Oleinik’s criterion in conservation laws, Applicable Analysis: An International Journal, DOI: 10.1080/00036811.2014.915520
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content Any opinions and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis The accuracy of the Content should not be relied upon and should be independently verified with primary sources
of information Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content
This article may be used for research, teaching, and private study purposes Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden Terms &
Trang 2and-conditions
Trang 3Existence of traveling waves to any Lax shock satisfying Oleinik’s
criterion in conservation laws
Mai Duc Thanha ∗and Nguyen Huu Hiepb
a Department of Mathematics, International University, Vietnam National University-Ho Chi Minh City, Quarter 6, Linh Trung Ward, Thu Duc District, Ho Chi Minh City, Vietnam; b Faculty of Applied Science, University of Technology, 268 Ly Thuong Kiet str., District 10, Ho Chi Minh City,
Vietnam
Communicated by M Shearer
(Received 9 March 2013; accepted 13 April 2014)
Given any shock wave of a conservation law where the flux function may not
be convex, we want to know whether it is admissible under the criterion of vanishing viscosity/capillarity effects In this work, we show that if the shock satisfies the Oleinik’s criterion and the Lax shock inequalities, then for an arbitrary diffusion coefficient, we can always find suitable dispersion coefficients such that the diffusive-dispersive model admits traveling waves approximating the given shock The paper develops the method of estimating attraction domain for traveling waves we have studied before
Keywords: Conservation laws; traveling wave; shock; diffusion; dispersion;
equilibria; asymptotical stability; Lyapunov function; LaSalle’s invariance principle; attraction domain
AMS Subject Classifications: 35L65; 74N20; 76N10; 76L05
1 Introduction
Let us consider the scalar conservation law
∂ t u (x, t) + ∂ x f (u(x, t)) = 0, x ∈ R I , t > 0, (1.1)
where the flux function f is assumed to be merely differentiable, and may not be convex
or concave, see Figure1
Solutions of (1.1) are understood in the sense of distributions and so are called weak
solutions Weak solutions are in general discontinuous as they may contain shock waves,
which are discontinuous jumps As well known, weak solutions are not unique Moreover, there are several kinds of shock waves, depending on the corresponding admissibility criterion, such as classical shocks and nonclassical shocks Each of these kinds of shock waves can be suitable for a particular application A reasonable mathematical modeling of the application can be made through the inclusion of some diffusion and dispersion terms
in the Equation (1.1) to get a diffusive-dispersive models
∗Corresponding author Email: mdthanh@hcmiu.edu.vn
Trang 4Figure 1 A nonconvex flux function and a shock.
∂ t u (x, t) + ∂ x f (u(x, t)) = β(b(u)u x ) x + γ (c1(u)(c2(u)u x ) x ) x , x ∈ R I , t > 0, (1.2)
where β > 0, γ > 0 are small scales, the functions b(u), c1(u), c2(u), andu ∈ R I are
assumed to be differentiable and positive, and c2(u), u ∈ R I , is twice differentiable The first
and the second terms on the right-hand side of (1.2) represent the diffusion and dispersion coefficients, respectively It has been known that traveling waves of (1.2) connecting a
given left-hand state u−and a right-hand state u+tends to the shock wave connecting these left-hand and right-hand states whenβ and γ tend to zero Thus, the existence of traveling
wave solutions of (1.2) corresponding to a shock wave can justify the admissibility criterion used to select this shock wave
In [1], given a diffusive-dispersive model with constant viscosity and capillarity coeffi-cients, the author proposed a method of estimating attraction domain of an asymptotically stable equilibrium point to establish the existence of a traveling wave associated with a given Lax shock In this paper, we will develop the method and improve the argument in [1] to show that for any given Lax shock of (1.1) satisfying the Oleinik’s condition, there are corresponding traveling waves of (1.2), providing that the diffusion and dispersion coefficients are chosen in a convenient way
Traveling waves for diffusive and/or dispersive terms have attracted many authors Traveling waves were considered earlier for diffusive-dispersive scalar equations by Bona and Schonbek [2], Jacobs, McKinney, and Shearer [3] Traveling waves and admissibility criteria of the hyperbolic-elliptic model of phase transition dynamics were also studied by Slemrod [4,5] and Fan [6,7] Traveling waves corresponding to nonclassical shocks were studied by LeFloch and his collaborators and students, see [8 16] The developments of the method of estimating attraction domain of an asymptotically stable equilibrium point to establish the existence of a traveling wave for various models were carried out in [17–21] See also [21–23] for related works
2 Background on shock waves and traveling waves
First, let us recall that a shock wave of (1.1) is a weak solution of the form
u (x, t) =
u−, x < st,
Trang 5where u−, u+are the left-hand and right-hand states, respectively, and s is the shock speed.
A function of the form (2.1) is a weak solution of the conservation law (1.1), the Rankine– Hugoniot relation
−s(u+− u−) + f (u+) − f (u−) = 0 (2.2) holds This relationship (2.2) means that the shock speed s can be evaluated by
s = s(u−, u+) = f (u+) − f (u−)
u+− u− .
Often, weak solutions are not unique Admissibility criteria have been set to select a unique solution For a single conservation law (1.1), one may use Oleinik’s criterion:
f (u) − f (u−)
u − u− ≥
f (u+) − f (u−)
u+− u− , for any u between u+and u−. (2.3)
The condition (2.3) is also stated as
f (u) − f (u+)
u − u+ ≤
f (u+) − f (u−)
u+− u− , for any u between u+and u−.
Geometrically, the inequality (2.3) means that if u+ < u−, the graph of f is lying below
the straight line() connecting the two points (u±, f (u±)) in the interval [u+, u−]
To deal with simultaneous conservation laws, one can make use of the Lax shock
inequalities, see [25] A shock wave of (1.1) is called a Lax shock if it satisfies the Lax
shock inequalities
f(u−) > s(u−, u+) > f(u+), u−= u+. (2.4)
Next, a traveling wave of (1.2) connecting a left-hand state u−to a right-hand state u+
is a smooth solution of the form u = u(y), y = x − st, where s is a constant, and satisfies
the boundary conditions
lim
y→±∞u (y) = u±,
lim
y→±∞
du
d y = lim
y→±∞
d2u
Substituting the traveling wave u into (1.2), we can see that the traveling wave u satisfies
the ordinary differential equation
−su+ ( f (u))= β(b(u)u)+ γ (c1(u)(c2(u)u)), (2.6) where(.)= d(.)/dy Integrating (2.6) and using the boundary condition (2.5), we obtain
βb(u)u+ γ c1(u)(c2(u)u)= −s(u − u−) + f (u) − f (u−). (2.7) Furthermore, the last equation and (2.5) give us
s= f (u+) − f (u−)
u+− u− ,
which means that u−, u+, and s satisfy the Rankine–Hugoniot relation (2.2) So, these quantities are, respectively, the left-hand state, right-hand state, and shock speed of a shock wave of (2.1)
Trang 6v(y) = c2(u)u.
Then, forγ = 0, the second-order differential equation (2.7) can be re-written as a system
of two differential equations of first order
u= v
c2(u) ,
v= −γ c βb(u)v
1(u)c2(u) +
1
γ c1(u) (−s(u − u−) + f (u) − f (u−)), (2.8)
where u = u(y), v = v(y), y ∈ R I satisfying
lim
y→±∞u (y) = u±, lim
y→±∞v(y) = 0.
Setting
h (u) = −s(u − u−) + f (u) − f (u−),
U = (u, v) T , F(U) =
v
c2(u) , −
βb(u)v
γ c1(u)c2(u)+
1
γ c1(u) h (u)
T
, (2.9)
we can re-write the system (2.8) in the form
dU
It is easy to check that a point U in the (u, v)-phase plane is an equilibrium point of the
autonomous differential equations (2.8) if and only if U has the form U = (u+, 0) for some
constant u+so that the states u±and the shock speed s satisfy the Rankine–Hugoniot relation
(2.2) Consequently, u = u(x, t) defined by (2.1) is a weak solution of the conservation law (1.1) Conversely, a jump of (1.1) of the form (2.1) gives equilibria(u−, 0), (u+, 0) of the
differential equation (2.8)
The Jacobian matrix D F (U) of the system (2.10) at U = (u, v) is given by
D F (U) =
⎛
⎜
⎝
−c2(u)v
c22(u)
1
c2(u)
−βdγ (u)v −c1(u)h(u)
γ c2
1(u) +
f(u) − s
γ c1(u) −
βb(u)
γ c1(u)c2(u)
⎞
⎟
⎠ ,
where d (u) = b (u)
c1(u)c2(u) At u±, using the condition that h (u±) = 0, we have
D F (u±, 0) =
⎛
⎜
⎝
c2(u±)
f(u±) − s
γ c1(u±) −
βb(u±)
γ c1(u±)c2(u±)
⎞
⎟
⎠
The characteristic equation of D F (u±, 0) is then given by
λ2+ a1(u±)λ − a2(u±) = 0,
where
a1(u±) = βb(u±)
γ c1(u±)c2(u±) > 0, a2(u±) =
f(u±) − s
γ c1(u±)c2(u±) .
Trang 7Assume that the shock satisfies the Lax shock inequalities
f(u−) > s(u−, u+) > f(u+), u−= u+.
Then,
a2(u+) < 0, a2(u−) > 0.
Since a2(u−) > 0, the Jacobian matrix at (u−, 0) admits two real eigenvalues having
opposite signs
λ1(u−, 0) = −a1(u−) −
a21(u−) + 4a2(u−)
λ2(u−, 0) = −a1(u−) +
a21(u−) + 4a2(u−)
The point(u−, 0) is thus a saddle point.
As seen above, a2(u+) < 0 So, the Jacobian matrix at (u+, 0) admits two eigenvalues
with negative real parts
λ1,2 (u+, 0) = −a1(u+) ±
a12(u+) + 4a2(u+)
Precisely, if a12(u+) + 4a2(u+) ≥ 0, then λ1,2 (u+, 0) are real and negative Otherwise,
λ1,2 (u+, 0) are complex, conjugate, and have the real negative part −a1(u+)/2 Thus, the
point(u+, 0) is asymptotically stable.
3 Existence of traveling waves
3.1 Estimate of attraction domain of the attracting equilibrium
Let us re-write the system (2.8) in the form
u= v
c2(u) ,
v= − βb(u)v
γ c1(u)c2(u)+
1
where
h (u) = −s(u − u−) + f (u) − f (u−).
We define a Lyapunov function candidate corresponding to the equilibrium point(u+, 0):
L (u, v) = γ1
u+
u
c2(ξ)
c1(ξ) h (ξ)dξ +
v2
Let the shock wave connecting the left-hand state u−with the right-hand state u+with
the shock speed s = s(u−, u+) satisfy the Oleinik’s criterion and the Lax shock inequalities.
For definitiveness, we assume that
u+< u−,
without restriction
Trang 8Th e o r e m 3.1 There always exists a value u∗< u+such that
h (u) > 0, u∗< u < u+,
h (u) < 0, u+< u < u−. (3.3)
Consequently, the function L defined by (3.2) satisfies
L (u+, 0) = 0, L(u, v) > 0, , u∗< u < u−, u = u+,
˙L(u, v) = − β γ c b (u)v2
1(u)c2(u) < 0, for v = 0. (3.4) This means that L is a Lyapunov function on the set
D : u∗≤ u ≤ u− Proof Since h(u+) = −s + f(u+) < 0, by the Lax shock inequalities, and h(u+) = 0,
the continuity implies that h (u) > 0 for u ∈ (u+− ε, u+), for some ε > 0 This establishes
the first statement in (3.3) The second statement of (3.3) follows from the Oleinik criterion:
h (u) = −s(u − u−) + f (u) − f (u−) = (u − u−)
−s + f (u) − f (u−)
u − u−
< 0,
for u+< u < u−
Next, we have
L (u+, 0) = 0,
and
L (u, v) ≥ γ1
u+
u
c2(ξ)
c1(ξ) h (ξ)dξ > 0
by using (3.3) This establishes the statements in the first line of (3.4) Next, the derivative
of L along trajectories of (3.1) is given by
˙L(u, v) = ∇L(u, v)· < du
d y , d v
d y >
= −c2(u)h(u)
γ c1(u)
v
c2(u) +
− βb(u)v
γ c1(u)c2(u) +
h (u)
γ c1(u)
v
= − βb(u)v2
γ c1(u)c2(u) < 0, for v = 0,
Le m m a 3.2
(a) Let u m = (u∗+ u+)/2, where u∗is given by (3.3) Given any continuous function
c1(u) > 0, u ∈ R I , there are always infinitely many choices of a C∞ function
c2(u) > 0, u ∈ R I such that
1
c2
c1 L∞[u+ ,u−]
u m
u∗
c2(u)
c1(u) du >
||h|| L∞[u+ ,u−]
min[u∗ ,u m]h (u) |u+− u−|. (3.5)
Trang 9Figure 2 An estimate β of the attraction domain of the asymptotically stable equilibrium point
(u+, 0).
Conversely, given any positive continuous function c2(u), u ∈ R I , there are always infinitely many choices of a C∞function c
1(u), u ∈ R I satisfying (3.5).
(b) Let c1, c2 be a pair of positive continuous functions satisfy (3.5) Then, for any
positive number 0 < β < L(u−, 0), the set
β := {(u, v) ∈ D| L(u, v) ≤ β} (3.6)
(see Figure2) is a compact set, positively invariant with respect to (3.1), and has
the point (u+, 0) as an interior point.
Proof
(a) Observe that a C∞function has the derivative up to any order Setα to be the value
of the right-hand side of the inequality (3.5):
α = ||h|| L∞[u+ ,u−]
min[u∗ ,u m]h (u) |u+− u−|. (3.7)
Obviously,α > 0 Let us take an arbitrary positive continuous function ω(u), u ∈ R I ,
and define a function
c (u) =
⎧
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎩
2α
u m − u∗max{c1(u), ω(u)}, if u ∈ [u∗, u m ],
1
2min{c1(u), ω(u)}, if u ∈ [u+, u−],
c (u m ) + c (u+) − c(u m )
u+− u m (u − u m ), if u ∈ [u m , u+],
2α
u m − u∗max{c1(u∗), ω(u∗)} = c(u∗), if u ≤ u∗,
1
2min{c1(u−), ω(u−)} = c(u−), if u ≥ u−.
(3.8)
It is easy to see that the function c defined by (3.8) is positive and continuous on R I ,
and
Trang 10c (u) ≥ 2α
u m − u∗c1(u), u ∈ [u∗, u m ],
c (u) ≤ 1
2c1(u), u ∈ [u+, u−]. (3.9)
Then, we will consider the mollification c ε in the interval I = (u∗− 1, u−+ 1) of
c defined by (3.8) Recall that a standard mollifier is defined by
η(u) =
C exp
1
|u|2 −1
, if |u| < 1,
where
1
C =
1
−1exp
1
|u|2− 1
du
Then, for eachε > 0, the function
η ε (u) = 1
ε η(u/ε), u ∈ R I ,
is C∞, has the support in(−ε, ε), and satisfies
R I
η ε (u)du = 1.
A mollification of c on I is defined by
c ε (u) = (c ∗ η ε )(u) =
u−+1
u∗−1 η ε (u − v)c(v)dv =
ε
−ε η ε (v)c(u − v)dv (3.10)
As well known, c ε ∈ C∞(I ε ), where I ε = (u∗−1+ε, u−+1−ε) It is not difficult
to check that
c ε (u) = c(u∗), u ∈ (u∗− 1 + ε, u∗− ε),
c ε (u) = c(u−), u ∈ (u−+ ε, u−+ 1 − ε),
for 0< ε < 1/2 Thus, a natural extension of c ε to be constant outside I εmakes sure
that it is of C∞(R I ) Moreover, since c is continuous, c ε → c as ε → 0 uniformly
on compact subsets of the interval I = (u∗− 1, u−+ 1) Thus, since c(u), u ∈ R I ,
is positive, there existsε0> 0 such that
2c (u) > c ε (u) > c(u)/2, u ∈ [u+, u−] ∪ [u∗, u m ], 0 < ε < ε0.
From (3.9) and the last inequalities, considering c ε as c2, we can estimate the left-hand side of (3.5) by
1
c ε
c1 L∞[u+ ,u−]
u m
u∗
c ε (u)
c1(u) du >
1
2c
c1 L∞[u+,u−]
u m
u∗
c (u)
2c1(u) du
2(1/2)
u m
u∗
2α
2(u m − u∗) du
= α, 0 < ε < ε0,
... (ξ)dξ > 0by using (3.3) This establishes the statements in the first line of (3.4) Next, the derivative
of L along trajectories of (3.1) is given by
˙L(u,... 9
Figure An estimate β of the attraction domain of the asymptotically stable equilibrium point
(u+,... s(u−, u+) satisfy the Oleinik’s criterion and the Lax shock inequalities.
For definitiveness, we assume that
u+< u−,