Electronic Journal of Differential Equations, Vol. 1996(1996), No. 4, pp. 117. ISSN 1072-6691. URL: http://ejde.math.swt.edu or http://ejde.math.unt.

edu ftp (login: ftp) 147.26.103.110 or 129.120.3.113

Weak Solutions to the One-dimensional Non-Isentropic Gas Dynamics by the Vanishing Viscosity Method
Kazufumi Ito

Abstract In this paper we consider the non-isentropic equations of gas dynamics with the entropy preserved. Equations are formulated so that the problem is reduced into the 2 2 system of conservation laws with a forcing term in momentum equation. The method of compensated compactness is then applied to prove the existence of weak solution in the vanishing viscosity method.

Introduction
t + (u)x = 0 (u)t + (u2 + p)x = 0 st + usx = 0.

Consider the one-dimensional gas dynamics equation in the Eulerian coordinate (1.1)

where , u, p and s denote the density, velocity, pressure and entropy. Other relevant quantities are the internal energy e and the temperature T . We assume that the gas is ideal, so that the equation of state is given by p = RT and that it is polytropic, so that e = cv T and (1.2) p = ( 1)es/cv

where = cp /cv > 1 and R = cp cv . Dene by (1.3) 1 = ( 1) es/cv .

1991 Mathematics Subject Classications: 35L65. Key words and phrases: Compensated compactness, Conservation laws, Entropy. c 1996 Southwest Texas State University and University of North Texas. Published June 18, 1996.

Weak Solutions

EJDE1996/04

Then, satises t + u x = 0. Thus, we consider the Cauchy problem (equivalent to (1.1)) t + mx = 0 mt + (m2 / + p)x = 0 t + u x = 0 .

(1.4)

where (1.5) m = u and p = 1 1 ,

with smooth initial data (0 , m0 ) in L (R2 ) that approaches a constant state ( , m ) at innity and satises 0 (x) 1 > 0, and 0 in W 1, (R) that satises (0 )x converges to 0 at innity and (1.6) 0 (x) 2 > 0 and (0 )x (x) 0 (or (0 )x (x) 0).

Consider the conservation form of the gas dynamics t + (u)x = 0 (u)t + (u2 + p)x = 0 1 1 [ ( u2 + e)]t + (u [ u2 + e] + pu)x = 0. 2 2

(1.7)

System (1.7) can be written as the hyperbolic system of conservation laws yt + f (y )x = 0

2 3 where y = y (t, x) = (, u, ( 1 2 u + e)) R and f is a smooth nonlinear map3 3 ping from R to R . System (1.4) is equivalent to system (1.7) when solutions are smooth but not necessarily when solutions are weak (e.g.,[Sm, Chapters 1617]). It is proved in Corollary 3.6 that the viscosity limit of solutions to (1.10) satises t + qx 0 in the sense of distributions, i.e., the third equation of (1.7), the conservation of energy t + qx = 0 is replaced by the non-energy production. We also note that the isentropic solution ( = const) [Di1] is a weak solution of (1.4) but not necessarily of (1.7). In this paper we show the existence of weak solutions to (1.4)-(1.5) using the vanishing viscosity method. The function (, m, ) L () L () W 1, () with = [0, ] R is a weak solution of (1.4)-(1.5) if satises the third equation of (1.4) a.e in and

(1.8)
0

(v t + F (, m, ) x ) dx dt = 0

EJDE1996/04

Kazufumi Ito

for all Cc (; R2 ) where v = (, m) and

(1.9)

F (, m, ) = (m, m2 / + p)

We consider the viscous equation of (1.4) with equal diusion rates t + mx = xx mt + (m2 / + p)x = mxx t + u x = xx .

(1.10)

It will be shown in Theorem 3.5 that the solutions ( , m , ) to (1.10) converge to a locally dened (in time) weak solution (, m, ) of (1.4). Our approach is based on the following observation. Suppose is a constant. Then equation (1.4) reduces to the isentropic gas dynamics. For the isentropic equation it is shown in DiPerna [Di1] that (1.4) has a weak solution by the vanishing viscosity method and using the theory of compensated compactness. The key steps in [Di1] are given as follows. First, if is a constant and = ( 1)/2 then (1.11) m 1 m 1 w = G1 (, m, ) = + and z = G2 (, m, ) = + are the Riemann invariants so that v G1 and v G2 are the two left eigenvectors of the 2 2 matrix 0 1 2m . v F = m2 2 + 2 2 where is assumed to be a positive constant. The method of invariant regions ([CCS],[Sm]) is applied to G1 , G2 to obtain that 0 const, |m / | const. Then, there exist a subsequence of v = ( , m ), still denoted by v and a Young measure t,x such that for each C (R2 ) we have (v ) converges in L () where weak star to t, x) = , = (

(y ) dt,x (y ), a.e. (t, x) .

Using the entropy elds [La] and the div-curl theorem of Murat [Mu] and Tatar [Ta] for bilinear maps in the weak topology, (1.12) , 1 q2 2 q1 = , 1 , q2 , 2 , q1

for all entropy/entropy ux pairs (i , qi ) so that v v F = v q.

Weak Solutions

EJDE1996/04

Then, using the weak entropy pairs (i.e., (0, ) = 0) it is shown that reduces to a point mass, i.e., v converges to v , a.e. in . We will apply the method described above for the non-isentropic equation (1.10). We need to overcome the two major diculties. First, G1 , G2 are no longer the Riemann invariants of the 3 3 matrix M : 0 1 0 m2 2m 2 2 2 21 M = F = 2 + . m 0 0 Next, equation (1.12) should be extended to the 3 3 system. We resolve these diculties by the following steps. Note (see Lemma 2.4) that Gi = Gi (, m, ), i = 1, 2 satises (1.13) 1 2 21 x , i=1 = ((Gi )xx 2 Gi (yx , yx )) (Gi )t + i Gi yx + 2 2 1 x , i = 2 where y = (, m, ) is a solution to (1.10) and 1 = u + , 2 = u are the eigenvalues of the 2 2 matrix v F . Here, Gi , i = 1, 2 are quasi-convex functions of (, m, ) (see Lemma 2.5), i.e., r Gi = 0 implies 2 Gi (r, r) 0. Note that from the third equation of (1.10) that 2 and | | |0 | (see Lemma 2.1) and moreover x satises (x )t + (u x )x = (x )xx
x Observing that if (, m, ) is a solution to (1.10) then = log( ) satises

(1.14)

t + u x = (xx + |(logx )x |2 |(log)x |2 ),

we show that if (0 )x 0 (resp. 0) then x 0 (resp. 0) and |x | c in provided that |(0 )x | c 0 in R (see Lemma 2.3). It thus follows from (1.13) and the quasi-convexity of Gi , i = 1, 2 that maxx G2 (t, x) maxx G2 (0, x) (resp. maxx G1 (t, x) maxx G1 (0, x)) and |2 21 x | c
2 +1

By the maximum principle (see Theorem 2.6), there exists a = c > 0 with c c monotonically decreasing and 0 = such that maxt[0, ], xR G1 (t, x) is less than a constant independent of > 0. Hence, we obtain 0 const, |m / | const and |x | const in .

EJDE1996/04

Kazufumi Ito

Second, in contrast to the isentropic case, the system (1.10) is not endowed with a rich family of entropy-entropy ux pairs. Thus, in order to prove that the Young measures t,x of a weakly star convergent subsequence of ( , m , ) reduce to a point mass, we rst note that { (t, x)} is precompact in L2 loc () and thus converges a.e. in (see Lemma 3.4). Also, we note that dividing the rst two equations (1.10) by , we obtain x x ) 2 (1.15) 1 px mx x )x + 2 = (m m t + (m 2 / + xx + 2 ). 2 m and m = (see Lemma 3.2). This implies that v = ( , m ) where = satises the (viscous) isentropic gas-dynamics with the forcing term px /2 p x in the momentum equation. Since L () uniformly in > 0, thus ( )2 p x 1,q { } >0 is precompact in Hloc (), 1 q < 2. Hence, the method of ( )2 compensated compactness in [Di1],[Ch] can be applied to the functions ( ,m ) to show that t,x is a point mass provided that 1 < 5/3. Regarding work on existence of weak solutions for conservation laws, we refer the reader to an excellent treatise by DiPerna [Di3] and references therein. Concerning basic framework on conservation laws, we refer the reader to [La],[Sm] and for the functional analytic framework of compensated compactness we refer [Mu],[Ta1],[Ev] and [Di2]. For scalar conservation laws the vanishing viscosity method is employed (e.g., in [Ol],[Kr] and references in [Sm]) to dene the unique entropy solution. Also, the vanishing viscosity method is used to develop the viscosity solution to the Hamilton-Jacobi equation in [CL]. The nite-dierence methods (e.g., Lax-Friedrichs and Gudunov schemes) are also used to construct weak solutions to a scalar and 2 2 system of conservation laws (e.g., see [Di2],[Ch] and [Sm]). In the case where the initial data have small total variation, Glimm [Gl] proved the global existence of BV-solutions for a general class of hyperbolic systems as the strong limit of random choice approximations. However, the problem of existence of solutions to (1.7) with large initial data is still unsolved. In [CD] the vanishing viscosity method is applied to the system (1.7) under a special class of constitutive relations in Lagrangian coordinates. t + m x = ( xx + 2

The Viscosity Method

In this section we establish the uniform L bound of y = ( , m , ). Lemma 2.1 If C 1,2 ([0, ] R) satises t + u x = xx then min 0 (x) (t, x) max 0 (x).
x x

Weak Solutions

EJDE1996/04

Proof: Using the same arguments as in the proof of Theorem 2.6, we can show that maxx (t, x) maxx 0 (x) and minx (t, x) minx 0 (x). 2 It will be shown in Section 3 (see (3.4) and Lemma 3.1) that the normalized mechanical energy (2.1) satises

E (, u, ) =

1 1 1 ( ) )1 (u u )2 + ( 2

(2.2)

E ((t, x), u(t, x), (t, x)) dx

E (0 (x), u0 (x), 0 (x)) dx.

The following lemma shows the lower bound of . Lemma 2.2 If C 1,2 ([0, ] R) satises (2.3) t + (u )x = xx

with (0, ) 0 and u C 1 (), then (t, ) 0. Moreover, if (0, ) > 0 and

(2.4)
0

|u u0 |2 dx dt const,

then (t, ) ( , ) > 0 on (0, ). Proof: Choose = min ((t, x), 0). Then we have

1 | (t, x)|2 + 2

t 0

( |x |2 u x ) dx ds = 0.

By the H older inequality, we obtain

| (t, 0)|2

|u| 2

t 0

| |2 dx ds

where |u| = sup(t,x)(0, )R |u(t, x)|, and the Gronwalls inequality implies = 0. Thus, 0. Next, we prove (t, ) = ( , ) > 0 if (0, ) > 0 by using the Stampacchias lemma (e.g., see [FI],[Tr]), i.e., suppose (c) is a nonnegative, non-increasing function on [c0 , ), and there exist positive constants K, s and t such that ( c) Kcs ( c c)s (c)1+t then (c ) = 0 for c = 2c0 (1 + 2 for all c > c c0 ,
1+2t t2

1+t st

(c0 )

1+t s

).

EJDE1996/04

Kazufumi Ito

First, we establish a priori bound. We consider the class K [Di1] of strictly convex C 2 functions h with following properties: h( ) = h ( ) = 0, h() = on (0, /2) for some 0 < < 1.

Premultiplying the rst equation of (1.10) by h () we obtain h()t + (h ()u)x h ()x u = (h()xx h ()2 x) Integration of this over (0, t) R yields
t

h((t, x)) h(0 (x)) dx +

0 t

h ()2 x dx dt

=
0

h ()x (u u ) dx dt .

Note that h ()x (u u ) 2 h ()2 x+ 1 h ()2 (u u )2 . 2

Since there exists some constant > 0 such that /2 M 2 h () for 2 h () h() for 0 < < /2 it follows that 2 h ()(u u )2 ( (u u )2 + h()). Hence,
t

h((t, x)) h(0 (x)) dx + 2

t 0

h ()2 x dx dt

(u u )2 + h() dx dt .

and it follows from (2.4) and Gronwalls inequality that

(2.5)

h((t, x)) dx const on [0, ].

Set = 1/. Then from (2.3) satises t + u x ux = (xx 2|x |2 ).

We further introduce = e t with > 0 to be determined later. The equation for becomes (2.6) t + +u x ux = ( xx 2| x |2 ).

Weak Solutions

EJDE1996/04

Dene = c = max (0, c) with c c0 = 1 . Pre-multiplication of (2.6) by 3 and integration over R yields (2.7) 1 3 ( ( 4 )t + ( 4 + c 3 ) + |( 2 )x |2 ) 4 4

u (5 3 + 3c 2 )x dx.

Note that

5u 3 x

5 2 2 u ( )x dx 2 4

|( 2 )x |2 dx+

25 2 |u| 4

| |4 dx

To estimate the second term on the right hand side of (2.7), we dene Ic (t) = {x R : (t, x) > 0} = {x R : (t, x) > c et }. Then, using the H older inequality, we have

3c 2 x dx =

3c u ( 2 )x dx 2
1/2 1/6

3c |u| 2

|( 2 )x |2 dx 9c2 2 |u| 4

| |6 dx
1/3

|Ic (t)|1/3 |Ic (t)|2/3 . 25 2 |u| , and integrat4

|( 2 )x |2 dx +

| |6 dx +

Substituting these estimates into (2.7), choosing = ing on the interval [0, t], we obtain
t t

| |4 dx +
0

(| 2 |2 + |( 2 )x |2 ) dx ds
1/3

(2.8) Since

9c2

|u|2 2

| |6 dx

|Ic (s)|2/3 ds.

1/2

| 2 | we have

(| 2 |2 + |( 2 )x |2 ) dx

1/3

| |6 dx

21/6

1/2

(| 2 |2 + |( 2 )x |2 ) dx

Thus, from (2.8),

t

| |4 dx +
0

(| 2 |2 + |( 2 )x |2 ) dx ds
t

21/3

81c4 4 |u| 4

|Ic (s)|4/3 ds Kc4 (c)4/3

0

EJDE1996/04

Kazufumi Ito

where we dene (c) = sup |Ic (t)| and K = 21/3

t[0, ]

81 . 4

Clearly (c) is nonnegative, non-increasing on [c0 , ). Moreover, for c > c,

| |4 dx
Ic

| |4 dx ( c c)4 Ic (t).

Hence,

4 Ic c c)4 (c)4/3 (t) K c (

and by taking the sup over t we obtain ( c) Kc4 ( c c)4 (c)4/3 . It follows from (2.5) that (2/ ) < and thus from Stampacchias lemma that (c ) = 0 for some c c0 and hence (t, x) c et where c depends on and . and (t, x) (c )1 et ,

The following lemma shows that |x (t, )| is uniformly bounded by (t, ) for every t [0, ] under assumption (1.7). Lemma 2.3 Assume that 0 satises (1.7) and |(0 )x | c 0 in R. Then |x (t, )| c (t, x) in R for every t [0, ]. Proof: If the initial condition (0 , m0 , 0 ) is suciently smooth and (0 )x 3 then the solution to (1.10) satises , u, C 3 () and x (t, ) > 0. Note that x satises (2.9) (x )t + (u x )x = (x )xx .

x Then, it is not dicult to show that if we dene = log( ) then satises

t + u x = (xx + |(logx )x |2 |(log)x |2 ) Suppose (t, x0 ) = maxx (t, x). Then x (t, x0 ) = (log x )x (t, x0 ) (log )x (t, x0 ) = 0 and xx (t, x0 ) 0. Thus, t (maxx (t, x)) 0, which implies the lemma. Since the solution to (2.9) continuously depends on the initial data (0 )x the estimate holds for when (0 )x 0. 2 The following lemmas provide the technical properties of the functions Gi (t), i = 1, 2 dened by (1.11).

10

Weak Solutions

EJDE1996/04

Lemma 2.4 If y = (, m, ) C 1,2 ((0, ) R)3 is a solution to (1.10), then (1.13) holds. Proof: First, note that the 3 3 matrix M = F has the eigenvalues 1 = m m m + , 2 = and and that v Gi , i = 1, 2 are the left eigenvectors of the sub-matrix v F corresponding to i . Thus, (G1 )t +1 (Gi yx + 1 x )( 2 21 21 +u 1 )x = G1 yxx .

Since G1 yxx = (G1 )xx 2 G1 (yx , yx ) we obtain (1.13) for G1 . The same calculation applies to G2 . 2 Lemma 2.5 If > 0, > 0 then Gi , i = 1, 2, are quasi-convex. m 1 Proof: We prove G1 = + is quai-convex. The same proof applies to G2 . Note that m 2 + 1 1 G1 = 1 2m 1 3 + ( 1) 2 2 1 1 1 2 G1 = . 0 0 2 1 1 0 ( + 1) 2 If r = (X, Y, Z ) satises r G1 then Y = m + X + +1 1 Z . Thus,

2 Gi (r, r) = ( + 1) (2 X 2 2 11 XZ + 2 Z 2 ) = ( + 1)2 2 ( X Z )2 0. 2 We now state the main result of this section that establishes the uniform L -bound of ( , m , x ) in > 0. Theorem 2.6 Suppose 0 satises (1.7) and |(0 )x | c0 in R. Then, there exists a = c > 0 with c c monotonically decreasing and 0 = such that 0 const, | m | const and |x | const in = [0, ] R. Proof: Suppose that (0 )x 0. Then, it follows from Lemmas 2.3-2.6 that m 1 maxx G2 (t, x) maxx G2 (0, x) = A. Hence, + A 0. Set G = A + G1 . It then follows from Lemma 2.4 that Gt + 1 G yx + 1 2 21 x = (Gxx 2 G(yx , yx ))

EJDE1996/04 Let Gk = G(kh), h > 0. Then,

Kazufumi Ito

11

k 1 2 k k k Gk Gk1 +1 Gk yx + (k )2 (k )21 (k )x = (Gk xx G (yx , yx ))+(h)

where (h)/h 0 as h 0+ . Suppose Gk (x0 ) = maxx Gk (x). Then, Gk x (x0 ) = k (Gk yx )(x0 ) = 0 and Gk xx (x0 ) 0. It follows from Lemmas 2.3 and 2.5 that if (t) = maxx G(t, x) then (kh) ((k 1)h) (h) c c ( (x0 ))2+1 (2+1)/ (kh)(2+1)/ .

Taking the limit h 0+ , we obtain

t

(t) (0)
0

c (2+1)/ ( )(2+1)/ d

and thus (2.10) In fact, if

t

(t)

(0)(+1)/ c +1 (2 +1)/ 1 ( ) (0)(+1)/ t

/( +1)

s(t) = (0) +
0

c (2+1)/ ( )(2+1)/ d s , which implies (2.10). Since G2 m G1 ,

then (t) s(t) and s 0

(2 +1)/ (2 +1)/

1 1 (G1 + G2 ) 2

and

the lemma follows from (2.9).

Compensated Compactness

In this section we show that the sequence {( , m , )} >0 has a subsequence that converges to a weak solution of (1.10) a.e in using the method of compensated compactness. First note that the mechanical energy (3.1) = 1 m2 1 + +1 2 ( 1)

and the corresponding entropy-ux (3.2) q= 1 m +1 m 3 ( ) + 2 1

12

Weak Solutions

EJDE1996/04

form an entropy pair, i.e., (3.3) M = q

In order to treat solutions approaching a nonzero state at innity, we consider a normalized entropy pair = (y ) (( v , )) v (( v , ))(v v ), q = q (y ) q (( v , )) v (( v , ))F (y ) where v = (, m), v = ( , m ) and y = (v, ). Premultiplying (1.10) by , we obtain t + q x = ( xx 2 (yx , yx )). Integration over yields an energy estimate
t

(3.4)

(t, x) dx +
0

2 (yx , yx ) dx dt =

(0, x) dx.

The following lemma implies the energy estimate (2.2) where (y ) = E (, u, ). Lemma 3.1 For > 0, > 0, 2 is non-negative. Proof: Note that m2 + 2 +1 3 m 2 = 2 1 Thus, (3.5) 2 (yx , yx ) = 1 m ( x mx )2 + 2 1 ( x x )2 0 m 2 1 0 1 0
1

for yx = (x , mx , x ).

The following lemma establishes the viscosity estimate which is essential for the method of compensated compactness. Lemma 3.2 Assume that 1 < 2 and is a solution of (1.10)
0

(0, x) dx < . Then, if (, m, )

(|x (t, x)|2 + |mx (t, x)|2 ) dx dt const

where > 0 is dened in Theorem 2.6

EJDE1996/04

Kazufumi Ito

13

Proof: From (2.9) and Lemma 2.2 we have

|x (t, x)| dx =

|x (0, x)| dx ,

t [0, ].

It thus follows from Theorem 2.6 that

0

|x (t, x)|2 dx const.

Since 0 < (t, x), (t, x) const in it follows from (3.5) that 2 (yx (t, x), yx (t, x)) + |x (t, x)|2 c1 |yx (t, x)|2 for some c1 > 0. Hence, the lemma follows from (3.4).

We apply the method of compensated compactness for the function v dened by m v = ( ,m )=( , ) The function v satises the 2 2 viscous conservation law (1.15) with the forcing term which is in L (). Based on this observation we have Lemma 3.3 Assume that the conditions in Theorem 2.6 are satised and that (0, x) dx < . Then, for 1 < 2, the measure set ( v )t + q ( v )x lies in a compact subset of m of v F , where v =( , ).
1 Hloc ()

for all weak entropy/entropy ux pair (, q )

Proof: Suppose (, m, ) is a solution to (1.10). Then, dividing the rst two m equations of (1.10) by , we obtain (1.15) for = and m = . Let (, q ) be a weak entropy/entropy ux pair, i.e., (3.6) v F = q and (0, ) = 0.

It can be shown that for 0 < const, | m | const (3.7) where = | | const and |2 (r, r)| const 2 (r, r) 1 m 2 1 ( ) + 2 ( 1)

is the mechanical energy, r is any vector in R2 and constant is independent of r. Premultiplying (1.15) by , we obtain ( v )t + q ( v )x = ( ( v )xx 2 ( vx , v x )) + ( v )A

14

Weak Solutions

EJDE1996/04

where A=2

x x mx x , 2 2

0,

px 2 > 0. It follows

p x L () uniformly in ( )2 from Lemma 3.2 and Theorem 2.6 that It follows from Theorem 2.6 that
1/2

x x mx x , ( )2 ( )2
>0

L2 ()

uniformly in > 0. Thus, { (v )A } Since

0

1,q is precompact in Wloc (), 1 q < 2.

|v x (t, x)|2 dx dt const

The set { v x } >0 is precompact in L2 () and so is { ( v )xx } >0 in H 1 (). Hence, the lemma follows from the fact that if set S is compact in W 1,q (U ) and bounded in W 1,r (U ) then S is compact in H 1 (U ) for 1 q < 2 < r and any bounded and open set U in R2 . [Ev] 2 In the next lemma we prove that the sequence { } L2 loc (). Lemma 3.4 For > 0 and > 0 dened in Theorem 2.6
0 >0

is precompact in

(|t |2 + |x |2 ) dx dt const. is compact in L2 (U ) for any bounded rectangle

Thus, the family { (t, x)} U = (0, ) (L, L).

>0

Proof: Premultiplying (1.10) by xx and integrating in (0, ) R, we obtain 1 2

|x (, x)|2 dx + 1 2

|xx |2 dx dt
0

|x (0, x)|2 dx +

1 |u|2 2

|x |2 dx dt.

where |u| = sup(t,x)(0, )R |u(t, x)|. Thus,

0 t

| xx |2 dx dt |u|2

|x |2 dx dt +

|x (0, x)|2 dx

and
0

|t |2 dx dt 4|u|2

|x |2 dx dt + 2

|x (0, x)|2 dx

EJDE1996/04

Kazufumi Ito

15

which proves the lemma. Now, we state the main result of the paper. Theorem 3.5 Assume that the conditions in Theorem 2.6 are satised and (0, x) dx < . Then, for 1 < 5/3, there exists a subsequence of ( , m , ) such that (3.8) ( (t, x), m (t, x), (t, x)) ((t, x), m(t, x), (t, x)) a.e. in = [0, ] R.
1, where the triple (, m, ) L () is a weak solution to + () L () W (1.4).

Proof: It follows from Lemma 3.3 that there exists a subsequence of ( ,m ) such that ( (t, x), m (t, x)) ( (t, x), m (t, x)) a.e. in . by applying the results of [Di1] and [Ch]. It follows from Lemma 3.4 that using a standard diagonal process, there is a subsequence of (t, x) that converges a.e. in , weakly in H 1 () and weakly-star in W 1, () to . Dene (t, x) = (t, x)(t, x), m(t, x) = m (t, x)(t, x) a.e. (t, x) . Then, the statement (3.8) holds. It follows from the rst two equations of (1.10) that
0

(( , m ) (t xx ) + F ( , m , ) x ) dx dt = 0

(; R2 ). It thus follows from (3.8) and the dominated convergence for all Cc theorem that (1.8) is satised. It follows from the third equation of (1.10) that 0

((t + u x ) + x x ) dx dt = 0

for all Cc (; R). Since u u in L2 (U ) for any bounded rectangle U = [0, ] [L, L] and weakly in H 1 () it follows that 0

(t + ux ) dx dt = 0

for all Cc (; R). Hence satises (1.4) a.e. in .

Corollary 3.6 Suppose the entropy pair (, q ) is dened by (3.1)-(3.2). Then

(3.9)
0

( t + q x ) dx dt 0

for all Cc (; R) satisfying 0. That is, the third equation of (1.1) is replaced by the inequality t + qx 0 in the sense of distributions.

16

Weak Solutions

EJDE1996/04

Proof: It follows from (3.3) that

0

( (t xx ) + q x ) dx dt =
0

2 (y , y ) dx dt

for all Cc (; R2 ) satisfying 0. It follows from Lemma 3.1 that the right hand side of this equality is nonnegative. Thus, by taking the limit as 0+ we obtain (3.9) 2

References [CCS] K. Chueh, C. Conley and J. Smoller, Positive invariant regions for systems of nonlinear diusions equations, Ind. U. Math. J., 26 (1979), 373-393. [CD] G. Q. Chen and C. M. Dafermos, The vanishing viscosity method in onedimensional thermo elasticity, LCDS Rep. 94-1, Brown University, RI. [Ch] G. Q. Chen, Convergence of the Lax-Friedrichs scheme for isentropic gas dynamics (III), Acta Math. Sci. 6 (1986), 75-120. [CL] M. G. Crandall and P. L. Lions, Viscosity solutions of Hamilton-Jacobi equations, Trans. Amer. Math. Soc, 277 (1983), 1-42. [Di1] R. J. DiPerna, Convergence of the viscosity method for isentropic gas dynamics, Commun. Math. Physics, 91 (1983), 1-30. [Di2] R. J. DiPerna, Convergence of approximate solutions to conservation laws, Arch. Rat. Mech. Anal., 82 (1983), 27-30. [Di3] R. J. DiPerna, Measure-valued solutions to conservation laws, Arch. Rat. Mech. Anal., 88 (1985), 223-270. [Ev] L. C. Evans, Weak Convergence Methods for Nonlinear Partial Dierential Equations, American Mathematical Society, Providence, 1990. [FI] W. Fang and K. Ito, Global solutions of the time-dependent drift-diusion semiconductor equations, J. Di. Eqns, to appear. [Gl] J. Glimm, Solutions in the large for nonlinear hyperbolic systems of equations, Commu. Pure Appl. Math. 18 (1965), 697-715. [Kr] N. Kruzkov, First order quasilinear equations in several independent variables, Math. USSR Sb., 10 (1970), 217-243. [La] P. D. Lax, Hyperbolic Systems of Conservation Laws and Mathematical Theory of Shock Waves, Conf. Board Math. Sci., 11, SIAM, 1973.

EJDE1996/04

Kazufumi Ito

17

[Mu] F. Murat, Compacit e par compensation, Ann. Scuola Norm. sup. Pisa,5 (1978), 489-507. [Ol] O. Oleinik, Discontinuous solutions of nonlinear dierential equations, Uspekhi Mat. Nauk 12 (1957), 3-73, (AMS Transl., Ser. 2, 26, (1963), 95-172). [Sm] J. Smoller, Shock Waves and Reaction-Diusion Equations, Springer-Verlag, New York, 1983. [Ta1] L. Tatar, Compensated compactness and applications to partial dierential equations, Heriot-Watt Sympos., Vol 4, R.J.Knops ed. Pitman Press, New York, 1979. [Ta2] L. Tatar, The compensated compactness method applied to systems of conservation laws, Systems of Nonlinear PDE, J.M.Ball ed., Reidel, Dordrecht, 1983. [Tr] G. M. Troianiello, Elliptic Dierential Equations and Obstacle Problems, Plenum Press, New York, 1987. Kazufumi Ito Center for Research in Scientific Computation North Carolina State University Raleigh, North Carolina 27695-8205 E-mail: kito@eos.ncsu.edu