f Mathe

al o ma
rn ti
ou

cs
J
Int. J. Math. And Appl., 6(3)(2018), 77–88

al

An
rnation

d its Applic
ISSN: 2347-1557

nte
• I
Available Online: http://ijmaa.in/

at
57

i
on
s • 5
-1
ISSN: 2347

International Journal of Mathematics And its Applications

Existence and Continuous Dependence of the Solutions of

the Benjamin-Bona-Mahony-Peregrine-Burger’s Equation
on the Circle

R. Gokilam1,∗

1 Department of Mathematics, Periyar University, Salem, Tamil Nadu, India.

Abstract: In this paper, we show the existence and continuous dependence of the solutions of the Benjamin-Bona-Mahony-Peregrine-
Burger’s(BBMPB) equation in Sobolev spaces H s , for s > 32 . We employ a Galerkin approximation argument to show
the existence of solutions of BBMPB equation.
Keywords: Continuous dependence, Sobolev space, Galerkin approximation.

c JS Publication. Accepted on: 20.06.2018

1. Introduction

Consider the initial value problem for the Benjamin-Bona-Mahony-Peregrine-Burgers(BBMPB) equation

ut − uxxt − αuxx + γux + θuux + βuxxx = 0

u(x, 0) = u0 (x)

where α is a positive constant, θ and β are nonzero real numbers. The BBMPB equation can be (and is more conveniently)
written in the following non-local form

ut + θuux = ∂x (1 − ∂x2 )−1 (−θuuxx + αux − γu − βuxx + θu2x )

The non-local form can be obtained from BBMPB equation as follows.

ut − uxxt − αuxx + γux + θuux + βuxxx = 0

adding and subtracting the terms 3θux uxx and θuuxxx

ut + θuux − uxxt − θuuxxx + θuuxxx − 3θux uxx + 3θux uxx − αuxx + γux + βuxxx = 0

ut + θuux − uxxt − θuuxxx − 3θux uxx = −θuuxxx − 3θux uxx + αuxx − γux − βuxxx

∗
E-mail: gokilamrangaswamy@gmail.com

77
 Existence and Continuous Dependence of the Solutions of the Benjamin-Bona-Mahony-Peregrine-Burger’s Equation on the Circle

(1 − ∂x2 )(ut + θuux ) = −θuuxxx − θux uxx − 2θux uxx + αuxx − γux − βuxxx

(1 − ∂x2 )(ut + θuux ) = ∂x [−θuuxx + αux − γu − βuxx + θu2x ]

multiply bothsides by (1 − ∂x2 )−1 we get

(ut + θuux ) = (1 − ∂x2 )−1 ∂x [−θuuxx + αux − γu − βuxx + θu2x ]

written this way, the BBMPB equation is a special case in the family of nonlinear wave equations of the form

ut + auux = L(u).

2. Preliminaries

Definition 2.1. A Schwarz function j(x)∈S(R) satisfying 0 ≤ ĵ(ξ) ≤ 1 for all ξ∈R, with ĵ(ξ) = 1 for |ξ| ≤ 1 and ĵ(ξ) = 0
1
P inx.
for |ξ| ≥ 2. We then define j (x) = 2π n ĵ(n)e Given j (x), we define Friedrichs mollifier on a test function f by the

convolution j f = j ? f .

Definition 2.2. For any s∈R the operator Λs = (1 − ∂x2 )s/2 is defined by

Λˆs u(k) = (1 + k2 )s/2 û(k)

where û is the fourier transform

Z
û(k) = e−ikx u(x)dx
T

The inverse relation is given by

1 X
u(x) = û(k)eikx
2π
k∈Z

s
Then, for u∈H (T ) we have
1 X
kuk2H s (T ) = (1 + k2 )s |û(k)|2 = kΛs uk2L2 (T ) .
2π
k∈Z

where Λ−2 = (1 − ∂x2 )−1 .

Theorem 2.3. For r < s we have

k I − J kL(H s ;H r ) = o(s−r )

Also, for any test function f, we have for all s > 0, J f −→ f ∈ H s . We similarly have the growth estimate when r > s.

Theorem 2.4. Let r ≥ s, then for any test function f

k J f kH r ≤ s−r k f kH s

Let Λ = (1−∂x 2 ) so that for any test function f, we have F(Λs f ) = (1+k2 )s fˆ(k). Then we have the following basic estimates.

Lemma 2.5. Let f be any test function, and σ ∈ R, then kΛσ f kL2 = kf kH σ , (1 − ∂x 2 )−1 f H σ = kf kH σ−2 , k∂x f kH σ ≤

kf kH σ+1 . We define the commutator [Λs , f ] = Λs f − f Λs , in which a test function f is regarded as a multiplication operator.

We will use the following negative Sobolev space estimate.

78
R. Gokilam

Proposition 2.6. If s > 32 , r + 1 ≥ 0 and r ≤ s − 1, then

k[Λr ∂x , f ] gkL2 ≤ cs,r kf kH s kgkH r

Also, we will using the Kato-Ponce commutator estimate.

Proposition 2.7. If s ≥ 0 then

k[Λs , f ] gkL2 ≤ cs (k∂x f kL∞ Λs−1 g L2 + kΛs f kL2 kgkL∞ )

Finally, replacing Λ with the J operator, we have the commutator estimate.

Proposition 2.8. Let J be the mollifier defined above, and f,g be two test functions, then

k[J , f ]gkL2 ≤ C kf kLip kgkH −1 .

Lemma 2.9 (Algebra Property). Let s > 1

2
and f, g ∈ H s , we have

kf gkH s ≤ cs kf kH s kgkH s .

Lemma 2.10 (Sobolev Interpolation Lemma). Let s0 < s < s1 be real numbers, then

s1 −s s−s0
s −s0 s −s0
kf kH s ≤ kf kH1s0 kf kH1s1 .

Lemma 2.11. Let s > 0 and J be defined as in J f (x) = j f (x). Then for any f ∈H s , we have J f → f in H s .

Lemma 2.12. Let w be such that k∂x wkL∞ . Then there is a constant c > 0 such that for any f ∈L2 , we have

k[J , w]∂x f kL2 ≤c kf kL2 k∂x wkL2 .

n
Proposition 2.13. Given σ = p
+ 1 and 1 < s < σ, there exists θ∈(0, 1) such that kf k s,
p ≤c kf kH σ,p and
H θ

kuk p ≤c kukH s−1,p .

L 1−θ

Lemma 2.14. If s > k + n

2
, where k is a nonnegative integer then H s (Rn ) ⊂ C k (Rn ) ∩ L∞ (Rn ), where the inclusion is

continuous. In fact,
X
k∂ α ukL∞ ≤Cs kukH s ,
|α|≤k

where Cs is independent of u.

1


Lemma 2.15. Let σ∈ 2
,1, then

kf gkH σ−1 ≤ kf kH σ−1 kgkH σ .

Lemma 2.16. Given q≥0, let u = u(x)∈H q be any function such that kux kL∞ < ∞. Then the there is a constant cq

depending only on q such that the following inequalities hold

Z 
 Λq uΛ(uux )dx ≤ cq kux k ∞ u2 q
 
  L H
RZ 
 Λq uΛ(u2 )dx ≤ cq kuk ∞ kuk2 q
 
  L H
R

On the other hand, one may estimate the following integral using integration by parts
Z  Z 
 f Λq uΛq ux dx = 1  fx (Λq u)2 dx ≤ 1 kfx k ∞ kuk2 q .
   
L H

R
 2 
R
 2

79
 Existence and Continuous Dependence of the Solutions of the Benjamin-Bona-Mahony-Peregrine-Burger’s Equation on the Circle

3. Local Well-posedness

To prove well-posedness, we employ a Galerkin approximation argument. The strategy will be to mollify the nonlinear

terms in the BBMPB equation to construct a family of ODEs. Then, we will extract a sequence of solutions to the ODEs,
which converges to the solution of the BBMPB equation in an appropriate space. We apply the mollifier Jε to the BBMPB

equation to construct family of ODEs in H s .

∂t u + θJ (J u J ∂x u ) = ∂x (1 − ∂x2 )−1 [−θ(u ∂x2 u ) + α∂x u − γu − β∂x2 u + θ(∂x u )2 ]

u (x, 0) = u0 (x)

Using the fact that

λ−2 = (1 − ∂x2 )−1

The non local form can be written as

∂t u + θJ (J u J ∂x u ) = ∂x λ−2 [−θ(u ∂x2 u ) + α∂x u − γu − β∂x2 u + θ(∂x u )2 ]

Our strategy is now to demonstrate that the Cauchy problem satisfies the hypotheses of the Fundamental ODE theorem.
We will therefore obtain a unique solution u (., t) ∈ H s , |t| < T , for some T > 0.

Energy estimate and lifespan of solution u

For each , there is a solution u to the mollified BBMPB equation. The lifespan of each of these solutions has a lower
bound T . In this subsection, we shall demonstrate that there is actually a lower bound T > 0 that does not depend upon .

To show the existence of T , we shall derive an energy estimate for the u . Applying the operator λs to both sides of i.v.p,

multiplying by λs u , and integrating over the torus yields the H s -energy of u .

Z Z Z
λs ∂t u λs u dx + λs θJ (J u J ∂x u )λs u dx = λs ∂x λ−2 [−θ(u ∂x2 u ) + α∂x u − γu − β∂x2 u + θ(∂x u )2 ]λs u dx

Consider the first term of the left hand side

Z
1 d
ku k2H s
λs ∂t u λs u dx =
2 dt
Z Z
1 d
ku k2H s + λs θJ (J u J ∂x u )λs u dx = λs ∂x (1 − ∂x2 )−1 [−θ(u ∂x2 u ) + α∂x u
2 dt

− γu − β∂x2 u + θ(∂x u )2 ]λs u dx

using the fact that

λ−2 = (1 − ∂x2 )−1

Z Z
1 d
ku k2H s = − λs θJ (J u J ∂x u )λs u dx + λs ∂x λ−2 α∂x u λs u dx
2 dt
Z Z Z
− γ λs ∂x λ−2 u λs u dx − β λs ∂x λ−2 ∂x2 u λs u dxθ λs ∂x λ−2 (∂x u )2 λs u dx

To bound the energy, we will need the following Kato-Ponce commutator estimate. We now rewrite the first term by first
commuting the exterior J and then commuting the operator λs with (J u ) arriving at

Z Z
θ λs J (J u J ∂x u )λs u dx = θ λs [J u ∂x J u ]λs J u dx

80
R. Gokilam

adding and subtracting the term on the right hand side θ (J u )λs ∂x J u λs J u dx
R

Z Z
θ λs J (J u J ∂x u )λs u dx = θ
Z
+θ
Z Z
θ λs J (J u J ∂x u )λs u dx = θ
Z
λs [J u ∂x J u ]λs J u dx − θ (J u )λs ∂x J u λs J u dx
(J u )λs ∂x J u λs J u dx
Z
[λs , J u ]∂x J u λs J u dx + θ (J u )λs ∂x J u λs J u dx

Setting v = J u , we can bound the first term of right hand side by first using the Cauchy-Schwarz inequality and then

applying the lemma (Kato-Ponce) and using the Sobolev theorem, we get

Z
θ [λs , v]∂x vλs vdx ≤ k[λs , v]∂x vkL2 kλs vkL2

cs kλs vkL2 k∂x vkL∞ + k∂x vkL∞ λs−1 ∂x v L2 kvkH s

≤



≤ cs kvkH s k∂x vkL∞ + k∂x vkL∞ k∂x vkH s−1 kvkH s



≤ cs kvkH s k∂x vkL∞ + k∂x vkL∞ kvkH s kvkH s



≤ cs kvkH s kvkH s + kvkH s kvkH s kvkH s

= 2cs kvk3H s

Next consider the second term of eqn , integrating by parts and using the Sobolev theorem, we have

 Z  Z 
θ v∂x λs vλs vdx = 1  (λs v)2 ∂x vdx
   
  2  
≤ k∂x vkL∞ kvk2H s

≤ kvkH s kvk2H s

= kvk3H s

Combining, we get

Z
θ λs J (J u J ∂x u )λs u dx ≤ (2cs + 1) kvk3H s

≤ (2cs + 1) kJ u k3H s

≤ (2cs + 1) ku k3H s

Consider the second term of the right hand side is bounded by first applying the Cauchy-Schwarz inequality and then using

the estimate and the algebra property of H s , we get

Z
λs ∂x λ−2 αJ ∂x u λs u dx ≤ λs ∂x λ−2 α∂x u L2 kλs u kL2

≤ ∂x λ−2 α∂x u H s ku kH s

≤ kα∂x u kH s−1 ku kH s

≤ α ku kH s ku kH s

≤ ku kH s ku kH s

= ku k2H s
Z
γ λs ∂x λ−2 u λs u dx ≤ λs ∂x λ−2 u L2 kλs u kL2

81
 Existence and Continuous Dependence of the Solutions of the Benjamin-Bona-Mahony-Peregrine-Burger’s Equation on the Circle

≤ ∂x λ−2 u H s ku kH s

≤ ku kH s−1 ku kH s

≤ ku k2H s
Z
λs ∂x λ−2 ∂x2 u λs u dx ≤ λs ∂x λ−2 ∂x2 u L2 kλs u kL2

β

≤ ∂x λ−2 ∂x2 u H s ku kH s

≤ ∂x2 u H s ku kH s

≤ k∂x u kH s+1 ku kH s

≤ ku kH s+2 ku kH s

≤ ku kH s ku kH s

≤ ku kH s ku kH s

= ku k2H s
Z
λs ∂x λ−2 (∂x u )2 λs u dx ≤ λs ∂x λ−2 (∂x u )2 L2 kλs u kL2

θ

≤ ∂x λ−2 ∂x u2 H s ku kH s

≤ (∂x u )2 H s−1 ku kH s

≤ ku k2H s ku kH s

≤ ku k3H s
1 d
ku k2H s ≤ (2cs + 3) ku k3H s + 3 ku k2H s
2 dt
1 d
ku k2H s ≤ (2cs + 3) ku k3H s + 3 ku k3H s
2 dt
= (2cs + 6) ku k3H s

Solving this inequality, gives

2
ku0 kH s

ku (t)k2H s ≤
1 − (2cs + 6)t ku0 kH s

which yields the minimum lifespan, T and energy estimate

1
T <
2(2cs + 6) ku0 kH s

and

ku (t)kH s ≤ 2 ku0 kH s

for |t| < T .

Refinement 1

Claim : To show that there exists a subsequence {uj } of {u } which converges in L∞ ([−T, T ]; H s ).

The family {u } is bounded in L∞ ([−T, T ]; H s ), since the family {u } is bounded (by the previous energy estimate) in

C([−T, T ]; H s ). Since L∞ ([−T, T ]; H s ) is the dual of L1 ([−T, T ]; H s ), we may apply Alaoglu’s theorem. By Alaoglu’s
theorem there exists a subsequence {uj } of {u } which converges to an element u ∈ L1 ([−T, T ]; H s ) in the weak∗ topology.

Moreover, the limit point u, satisfies the same size estimation bound and minimum lifespan estimate as the u solutions.

Refinement 2

82
R. Gokilam

Claim : To show that there is a further subsequence of our sequence {u } which converges to u in C([−T, T ]; H s−1 ).

To prove this we will employ Ascoli’s theorem. First to prove equicontinuity, let t1 and t2 ∈ [−T, T ]. By the mean value

theorem

ku (t1 ) − u (t2 )kH s−1 ≤ sup k∂t u kH s−1 |t1 − t2 | (1)
t∈[−T,T ]

Now consider the mollified equation

∂t u + θJ (J u J ∂x u ) = ∂x λ−2 [−θ(u ∂x2 u ) + α∂x u − γu − β∂x2 u + θ∂x u2 ]

Applying norm on both sides and using the triangle inequality and lemma, we have

k∂t u kH s−1 = ∂x λ−2 [−θ(u ∂x2 u ) + α∂x u − γu −β∂x2 u + θ∂x u2 ] H s−1

≤ ∂x λ−2 θu ∂x2 u H s−1 + ∂x λ−2 α∂x u H s−1 + k−γu kH s−1 − β∂x2 u H s−1 + θ∂x u2 ] H s−1

≤ a ku0 k3H s + b ku0 k2H s + ku0 kH s

Substituting in inequality (1), we get

ku (t1 ) − u (t2 )kH s−1 ≤ a ku0 k3H s + b ku0 k2H s + c ku0 kH s |t1 − t2 |

which implies {u (t)} is equicontinuous. Next, we observe that for each t ∈ [0, T ] the set U (t) = {u }∈(0,1] is bounded in

H s . Since T is a compact manifold, the inclusion mapping i : H s −→ H s−1 is a compact operator, and therefore we may

deduce that U (t) is a precompact set in H s−1 . As the two hypotheses of Ascoli’s theorem have been satisfied, we have a

subsequence {uv } that converges in [−T, T ]; H s−1 . By uniqueness of limits, this subsequence must converge to u.

Refinement 3

Claim : To refine the subsequence we show that the limit u is in the space C [−T, T ]; H s−σ for all σ ∈ (0, 1].

As in the previous case, we will prove that the family u satisfies the hypotheses of Ascoli’s theorem, and to do so we will
show that the sequence u is equicontinuous in H s−σ and uniformly bounded. In fact, we will show that the following

modulus of continuity

ku (t1 ) − u (t2 )kH s−σ ≤ ku0 k3H s + ku0 k2H s + ku0 kH s |t1 − t2 |σ


(2)

To prove the above inequality, we begin by estimating

ku (t1 ) − u (t2 )kH s−σ ≤ ku kC σ ([−T,T ];H s−σ ) |t1 − t2 |σ . (3)

By definition of the Holder norm

ku (t1 ) − u (t2 )kH s−σ

ku kC σ ([−T,T ];H s−σ ) = sup ku (t)kH s−σ + sup (4)
t∈[−T,T ] t1 6=t2 |t1 − t2 |σ

The first term of the right hand side of (4) is bounded by 2 ku0 kH s using the Sobolev embedding theorem followed by
estimate.

sup ku (t)kH s−σ ≤ sup ku (t)kH s ≤ 2 ku0 kH s .

t∈[−T,T ] t∈[−T,T ]

83
 Existence and Continuous Dependence of the Solutions of the Benjamin-Bona-Mahony-Peregrine-Burger’s Equation on the Circle

For the second term is more difficult, and we will open the norm to analyze it. We have

!1/2
ku (t1 ) − u (t2 )kH s−σ X
= |t1 − t2 |−σ (1 + k2 )s−σ |ub (k, t1 ) − ub (k, t2 )|2
|t1 − t2 |σ
k

First, as σ ∈ (0, 1), we have

 σ
1 1 1
≤ 1+ ≤1 + .
(1 + k2 )σ |t1 − t2 |2σ (1 + k2 )|t1 − t2 |2 (1 + k2 )|t1 − t2 |2

Using this inequality

(1 + k2 )s (1 + k2 )s
≤ (1 + k2 )s +
(1 + k2 )σ |t1 − t2 |2σ |t1 − t2 |2 (1 + k2 )
!1/2
ku (t1 ) − u (t2 )kH s−σ X X (1 + k2 )s
= (1 + k2 )s |ub (k, t1 ) − ub (k, t2 )|2 + |ub (k, t1 ) − ub (k, t2 )|
|t1 − t2 |σ |t1 − t2 |2 (1 + k2 )
k k
ku (t1 ) − u (t2 )kH s−σ
≤ 2 sup ku kH s + ku kC 1 ([−T,T ];H s−1 ) .
|t1 − t2 |σ t∈[−T,T ]

Using the solution size estimate and the estimate found in the previous refinement, we obtain

ku kC σ ([−T,T ];H s−σ ) ≤ (4 + c) ku0 kH s + a ku0 k3H s + b ku0 k2H s

Substituting into inequality (3) we establish a uniform modulus of continuity, and we conclude that the family {u } is

equicontinuous in the variable t. The precompactness condition is established in exactly the same fashion as the previous

case as the inclusion mapping of H s into H s−σ is a compact operator. As the two hypotheses of Ascoli have been satisfied,

we may extract a subsequence that converges to u in C [0, T ]; H s−σ . Similarly, we can refine the sequence {u } several

times, by finding a sub-sequence of solutions which converges to a solution to BBMPB equation. Hence the proof of existence

of a solution to the BBMPB equation.

Continuity of the data-to-solution map

Here we show that the dependence of the solution of the BBMPB equation on initial data is continuous.

Theorem 3.1 (Continuous dependence). The data-to-solution map u0 7−→ u(t) for the cauchy problem of the BBMPB

equation is continuous from H s −→ C(I; H s ).

Proof. Fix u0 ∈ H s and let {u0,n } ⊂ H s be a sequence with lim u0,n = u0 . If u is the solution to the BBMPB equation
n−→∞

with initial data u0 and if un is the solution to the BBMPB equation with initial data u0,n , we will demonstrate that

lim un = u in C(I; H s ). Equivalently, let η > 0. We need to show that there exists an N > 0 such that
n−→∞

n > N ⇒ ku − un kC(I;H s ) < η

As we will be using energy estimates in the H s norm, to get around the difficulty of estimating the terms, we will use the

J convolution operator to smoooth out the initial data. Let ε ∈ (0, 1]. We take uε be the solution to the Cauchy problem

for BBMPB equation with initial data Jε u0 = jε ∗ u0 and uεn be the solution with initial data Jε u0,n . Applying the triangle

inequality, we arrive at

ku − un kC(I;H s ) ≤ ku − uε kC(I;H s ) + kuε − uεn kC(I;H s ) + kuεn − un kC(I;H s ) .

84
R. Gokilam

η
We will prove that each of these terms can be bounded by 3
for suitable choices of ε and N. We note that the ε we have

introduced will be independent of N and will only depend on η; whereas, the choice of N will depend on both η and ε.
Estimating kuε − uεn kC(I;H s ) : Setting v = uε − uεn

∂t uε = (−θuε ∂x uε ) + λ−2 ∂x −θuε ∂x uε + α∂x uε − γuε − β∂x2 uε + θ(∂x uε )2

∂t uεn = (−θuεn ∂x uεn ) + λ−2 ∂x −θuεn ∂x uεn + α∂x uεn − γuεn − β∂x2 uεn + θ(∂x uεn )2

Subtracting

∂t (uε − uεn ) = (−θuε ∂x uε + θuεn ∂x uεn ) + λ−2 ∂x (−θuε ∂x uε + θuεn ∂x uεn ) + λ−2 ∂x α∂x (uε − uεn )

−γ(uε − uεn ) − β∂x2 (uε − uεn ) + θ (∂x uε )2 − (∂x uεn )2

Let

F (uε ) = (−θuε ∂x uε ) + λ−2 ∂x (−θuε ∂x uε ) + λ−2 ∂x θ(∂x uε )2

Let

F (uεn ) = (−θuεn ∂x uεn ) + λ−2 ∂x (−θuεn ∂x uεn ) + λ−2 ∂x θ(∂x uεn )2

∂t (v) = (−θuε ∂x uε + θuεn ∂x uεn ) + λ−2 ∂x (−θuε ∂x uε + θuεn ∂x uεn ) + λ−2 ∂x α∂x (v)

− λ−2 ∂x γ(v) − λ−2 ∂x β∂x2 (v) + λ−2 ∂x θ (∂x uε )2 − (∂x uεn )2

∂t (v) = (F (uε ) − F (uεn )) + λ−2 ∂x α∂x (v) − γ(v) − β∂x2 (v)



We calculate the H s energy of v.

Z Z Z
λs ∂t vλs vdx = λs (F (uε ) − F (uεn )) λs vdx + λs λ−2 ∂x α∂x (v) − γ(v) − β∂x2 (v) λs vdx


Applying Cauchy-Schwarz inequality

1 d
kv(t)k2H s ≤ kλs (F (uε ) − F (uεn ))kL2 kλs vkL2 + λs λ−2 ∂x α∂x (v) − γ(v) − β∂x2 (v) L2 kλs vkL2

2 dt
≤ kF (uε ) − F (uεn )k s kvk s + λ−2 ∂x α∂x (v) − γ(v) − β∂x2 (v) s kvk s

H H H H

Consider the first term of inequality

kF (uε ) − F (uεn )kH s kvkH s ≤ kvk2H s

Consider the second term of inequality and using the triangle inequality

−2 
λ ∂x α∂x (v) − γ(v) − β∂x2 (v) s kvk s ≤ λ−2 ∂x α∂x (v) s + γ λ−2 ∂x v s + β λ−2 ∂x ∂x2 (v) s

H H H H H

≤ α kvkH s + γ kvkH s−1 + kvkH s+1

M
≤ kvk2H s
ε

Where M is positive constant. Combining the above estimates, we obtain the differential inequality

 
1 d M
kvk2H s ≤ 1+ kvk2H s
2 dt ε

85
 Existence and Continuous Dependence of the Solutions of the Benjamin-Bona-Mahony-Peregrine-Burger’s Equation on the Circle

Let

cs M
= (1 + )
ε ε
1 d cs
kvk2H s ≤ kvk2H s (5)
2 dt ε

for some constant cs , Solving (5) gives for all t ∈ I

cs T
kv(t)kH s ≤ e ε kv(0)kH s
cs T
≤ e T ku0 − u0,n kH s (6)

We observe that (6) does not place any constraints on ε; however, handling the first and third terms of (6) will require ε to

be small. After ε is chosen, we take N sufficiently large so that

η cs T
ku0 − u0,n kH s < e ε
3

there by yielding

η
kuε − uεn kC(I;H s ) < .
3

Estimation of ku − uε kC(I;H s ) and kuε − uεn kC(I;H s ) : Let

∂t u = −(θu∂x u) + λ−2 ∂x [−θu∂x2 u + α∂x u − γu − β∂x2 u + θ(∂x u)2 ]

∂t uε = (−θuε ∂x uε ) + λ−2 ∂x −θuε ∂x uε + α∂x uε − γuε − β∂x2 uε + θ(∂x uε )2

As the differences uε − u and uεn − un satisfy the same inequalities, we will use the unified notation v = uε − u and

v = uεn − un and omit all n subscripts in formulae until we reach the point where different analysis for each case is needed.
In constructing this Cauchy problem for v, we note that as we are taking energy estimates in H s , we will want to avoid

having any u coefficients for the Burgers-type part of the equation as this may give rise to an expression of the form kukH s+1

which is undefined. There is no such problem for the nonlocal part of the equation so we will have F (uε − F (u)) as this can

be estimated.

∂t (u − uε ) = λ−2 ∂x [α∂x (u − uε ) − γ(u − uε ) − β∂x2 (u − uε )] − [F (u) − F (uε )]

∂t v = λ−2 ∂x [α∂x v − γv − β∂x2 v] − [F (u) − F (uε )]

v(x, 0) = Jε u0 − u0

Now let us obtain the H s energy of v,

Z Z Z
λs ∂t vλs vdx = λs λ−2 ∂x [α∂x v − γv − β∂x2 v]λs vdx − λs (F (u) − F (uε )) λs vdx

Applying the Cauchy-Schwarz inequality

1 d
kv(t)k2H s ≤ λs λ−2 ∂x [α∂x v − γv − β∂x2 v] L2 kλs vkL2 + kλs (F (u) − F (uε ))kL2 kλs vkL2

2 dt
86
R. Gokilam

1 d
kv(t)k2H s ≤ λ−2 ∂x [α∂x v − γv − β∂x2 v] H s kvkH s + k(F (u) − F (uε ))kH s kvkH s

2 dt

Consider the second term of the inequality

kF (u) − F (uε )kH s kvkH s ≤ kvk2H s

Consider the first term of the inequality and using the triangle inequality

−2
λ ∂x [α∂x v − γv − β∂x2 v] kvkH s ≤ λ−2 ∂x α∂x v H s + λ−2 ∂x γv H s + λ−2 ∂x β∂x2 v H s

Hs

≤ α kvkH s + γ kvkH s−1 + β kvkH s+1

Choosing M a positive constant

−2
λ ∂x [α∂x v − γv − β∂x2 v] M
kvk2H s

Hs
kvkH s ≤
ε

Combining the above estimates, we obtain the following differential inequality

 
1 d M
kvk2H s ≤ 1+ kvk2H s
2 dt ε

As in the previous estimation we will get

η
ku − uε kC(I;H s ) <
3

Similarly we can obtain

η
kuε − uεn kC(I;H s ) <
3

By combining all the inequalities ,we get

η η η
ku − un kC(I;H s ) < + +
3 3 3

Hence

ku − un kC(I;H s ) < η

Hence the data-to-solution map is continuous.

References

[1] R. Camassa and D. Holm, An integrable shallow water equation with peaked solitons, Physics Review Letter, 71(1993),
1661-1664.

[2] A. Degasperis, D.D. Holm and A.N.W. Hone, A new integrable equation with peakon solutions, Theoretical and Mathe-

matical Physics, 133(2002), 1463-1474.

87
 Existence and Continuous Dependence of the Solutions of the Benjamin-Bona-Mahony-Peregrine-Burger’s Equation on the Circle

[3] G. Fornberg and G.B. Whitham, A numerical and theoretical study of certain nonlinear wave phenomena, Philosophical

Transactions of the Royal Society London, 289(1978), 373-404.

[4] A. Himonas and C. Holliman, The Cauchy problem for the Novikov equation, Discrete Continuous Dynamical Systems,

31(2011), 469-488.

[5] A. Himonas and C. Holliman, On well-posedness of the Degasperis-Procesi equation, Nonlinearity, 25(2012), 449-479

[6] A. Himonas and C. Holliman, The Cauchy problem for a generalized Camassa-Holm equation, Advanced Differential
Equations, 19(2014), 161-200.

[7] T. Kato and G. Ponce, Commutator estimates and the Euler and Navier-stokes equations, Communications on Pure

and Applied Mathematics, 41(1988), 891-907.

[8] Y. Li and P. Olver, Well-posedness and blow-up solutions for an integrable nonlinearly dispersive model wave equation,
Journal of Differential Equations, 162(2000), 27-63.

[9] M. Li and Z. Yin, Blow-up phenomena and local well-posedness for a generalized Camassa-Holm equation with cubic

nonlinearity, Nonlinear Analysis: Theory, Methods and Applications, 151(2017), 208-226.

[10] M. Taylor, Pseudodifferential Operators and Nonlinear PDE, Birkhauser, Boston, (1991).
[11] M. Taylor, Commutator estimates, Proceedings of the American Mathematical Society, 131(2003), 1501-1507.

[12] W. Yan, Y. Li and Y. Zhang, The Cauchy problem for the Novikov equation, Journal of Nonlinear Differential Equations

and Applications, 20(2013), 1157-1169.

[13] A. Yin, Well-posedness and blow-up phenomena for a class of nonlinear third-order partial differential equations, Houston
Journal of Mathematics, 31(2005), 961-972.

[14] J. Yin, L. Tian and X. Fan, Classification of traveling waves in the Fornberg-Whitham equation, Journal o Mathematical

Analysis and Applications, 368(2010), 133-143.

88