Volume 2010, Article ID 281238,14pages doi:10.1155/2010/281238
Research Article
Approximate Controllability of
a Reaction-Diffusion System with
a Cross-Diffusion Matrix and Fractional
Derivatives on Bounded Domains
Salah Badraoui
Laboratoire LAIG, Universit´e du 08 Mai 1945, BP. 401, Guelma 24000, Algeria
Correspondence should be addressed to Salah Badraoui,[email protected]
Received 11 July 2009; Accepted 5 January 2010
Academic Editor: Ugur Abdulla
Copyrightq2010 Salah Badraoui. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
We study the following reaction-diffusion system with a cross-diffusion matrix and fractional derivativesuta1Δua2Δv−c1−Δα1u−c2−Δα2v1ωf1x, tinΩ×0, t∗,vtb1Δub2Δv−
d1−Δβ1u−d2−Δβ2v1ωf2x, tinΩ× 0, t∗,uv0 on∂Ω×0, t∗,ux,0 u0x,vx,0
v0xinx∈Ω,whereΩ⊂RNN≥1is a smooth bounded domain,u0, v0∈L2Ω, the diffusion
matrixM a1a2
b1b2
has semisimple and positive eigenvalues 0 < ρ1 ≤ ρ2, 0 < α1, α2, β1, β2< 1,
ω⊂Ωis an open nonempty set, and 1ωis the characteristic function ofω. Specifically, we prove that under some conditions over the coefficientsai, bi, ci, dii1,2, the semigroup generated by the linear operator of the system is exponentially stable, and under other conditions we prove that for allt∗>0 the system is approximately controllable on0, t∗ .
1. Introduction
In this paper we prove controllability for the following reaction-diffusion system with cross diffusion matrix:
uta1Δua2Δv−c1−Δα1u−c2−Δα2v1ωf1x, t inΩ× 0, t∗,
vtb1Δub2Δv−d1−Δβ1u−d2−Δβ2v1ωf2x, t inΩ× 0, t∗,
uv0 on∂Ω× 0, t∗,
ux,0 u0x, vx,0 v0x inx∈Ω,
1.1
We assume the following assumptions.
H1 Ωis a smooth bounded domain inRN N≥1.
H2The diffusion matrixMa1a2 b1 b2
has semisimple and positive eigenvalues 0< ρ1≤
ρ2.
H3cj, dj j 1,2are real constants,αj, βj j1,2are real constants belonging to the interval 0,1.
H4u0, v0∈L2Ω.
H5The distributed controlsf1, f2∈L20, t∗ ;L2Ω.
Specifically, we prove the following statements.
iIf c2d10 and min{c1λ11−α1ρ1, d2λ11−β2ρ1}>0, whereλ1is the first eigenvalue
of−Δwith Dirichlet condition, or if c2/0, d1/0, c1 ≥0, andd2 ≥ 0; then, under
the hypothesesH1–H3, the semigroup generated by the linear operator of the system is exponentially stable.
iiIf c2d10 and under the hypothesesH1–H5, then, for allt∗>0 and all open
nonempty subset ωofΩthe system is approximately controllable on 0, t∗ .
This paper has been motivated by the work done in1 and the work done by H. Larez and H. Leiva in2 . In the work1 , the auther studies the asymptotic behavior of the solution of the system
uta ∂2u
∂x2 β
∂u ∂xb
∂2v
∂x2 ft, u, v, x∈R, t >0,
vtc ∂2u
∂x2 d
∂2v
∂x2 β
∂v
∂xgt, u, v, x∈R, t >0
1.2
supplemeted with the initial conditions
ux,0 u0x, vx,0 v0x, x∈R. 1.3
The author proved that in the Banach spaceX×X where XCubRis the space of bounded
uniformly continuous real valued functions onR, iff andg are locally Lipshitz and under some conditions over the coefficients a, b, c, d, β, and if u0, v0 ∈ C {u ∈ CubR :
limx→∞uxexist},then ut, vt ∈ C for all t < tmax.Moreover, Ut limx→∞ux andVt limx→∞vxsatisfy the system of ordinary differential equations
Ut ft, Ut, Vt,
with the initial data
U0 lim
x→∞u0x, V0 xlim→∞v0x. 1.5
The same result holds forC−{u∈CubR: limx→ −∞uxexist}.
In the work done in2 , the authers studied the system1.1withc2d10, c1d2,
and α1 β2 1/2. They proved that if the diffusion matrix
a b
c d
has semi-simple and positive eigenvalues 0< ρ1≤ρ2,f1, f2∈L20, τ;L2Ω,then ifλ11/2ρ1β >0λ1is the first
eigenvalue of−Δ, the system is approximately controllable on 0, τ for all open nonempty subset ωofΩ.
2. Notations and Preliminaries
In the following we denote by
M2Rthe set of 2×2 matrices with entries fromR,
L2Ωthe set of all measurable functionsu:Ω → Rsuch that
Ω|u|2dx <∞,
H1Ω the set of all the functions u ∈ L2Ω that have generalized derivatives
∂u/∂xj∈L2Ωfor allj1, . . . , N, H1
0Ωthe closure of the setC∞0 Ωin the Hilbert spaceH1Ω,
H2Ω the set of all the functions u ∈ L2Ω that have generalized derivatives
∂u/∂xj, ∂2u/∂xj∂xk∈L2Ωfor allj, k1, . . . , N.
We will use the following results.
Theorem 2.1cf.3 . Let us consider the following classical boundary-eigenvalue problem for the
laplacien:
−Δuλu, onΩ,
u0, on∂Ω, 2.1
whereΩis a nonempty bounded open set inRNandD−Δ H2Ω∩H1 0Ω.
This problem has a countable system of eigenvalues 0< c ≤λ1 < λ2 <· · · < λj < · · · and λj → ∞asj → ∞.
iAll the eigenvalues λj have finite multiplicity mj equal to the dimension of the corresponding eigenspaceSj.
iiLet{ϕjk} mj
k1be a basis of the Sjfor everyj,then the eigenvectors {ϕjk}
mj,∞
k1,j1form
a complete orthonormal system in the space L2Ω.Hence for all u ∈ L2Ωwe
haveu ∞j1mj
k1u, ϕjkϕjk.If we putEju
mj
k1u, ϕjkϕjk then we get u
∞
j1Eju.
ivFor allu∈D−Δwe have −Δu∞j1λjEju.
vThe operatorΔgenerates an analytic semigroup {TΔt}on L2Ωdefined by
TΔtu
∞
j1
e−λjtEju. 2.2
Definition 2.2. Let 0< α <1 a real number, the operator−Δαis defined by
−Δα:D−Δα⊂L2Ω−→L2Ω,
D−Δα
⎧ ⎨
⎩u∈L2Ω|
∞
j1
mj
k1
λαju, ϕjk
2
<∞
⎫ ⎬ ⎭,
−Δαu∞ j1
mj
k1
λαjϕjk, u
ϕjk.
2.3
In particular, we obtain ϕjk ∈ D−Δα and −Δαϕjk λαjϕjk. Since {ϕjk} mj,∞
k1,j1 form a
complete orthonormal system in the spaceL2Ω,then it is dense inL2Ω, and henceD−Δα is dense in L2Ω.
Proposition 2.3cf.4 . LetXbe a Hilbert separable space and{Aj}j≥1and {Pj}j≥1two families
of bounded linear operators inX, with{Pj}j≥1a family of complete orthogonal projections such that AjPjPjAj, j ≥1.
Define the following family of linear operatorsStw∞j1eAjtP
jw, w∈X, t≥0.Then aStis a linear and bounded operator if eAjt ≤gt, j ≥1withgt≥0,continiuous for
t≥0,
bunder the above condition (a),{St}t≥0is a strongly continiuous semigroup in the Hilbert spaceX,whose infinitesimal generatorAis given by
Aw ∞
j1
AjPjw, w∈DA, DA
⎧ ⎨ ⎩w∈X|
∞
j1
AjPjw2
<∞
⎫ ⎬
⎭. 2.4
Theorem 2.4cf.5 . SupposeΩis connected,fis a real function inΩ, andf 0on a nonempty
open subset ofΩ. Thenf≡0inΩ.
3. Abstract Formulation of the Problem
In this section we consider the following notations.
iXL2Ω×L2Ω. Xis a Hilbert space with the inner product
iiWe define
A11u, v a1Δua2Δv−c1−Δα1u−c2−Δα2v,
A12u, v b1Δub2Δv−d1−Δβ1u−d2−Δβ2v.
3.2
iiiLetw u, v,then we can define the linear operator
A:DA⊂X−→X,
DA H2Ω;R∩H1 0Ω;R
2
,
Aw−MΔ−c1B1−Δα1−c2B2−Δα2−d1B3−Δβ1−d2B4−Δβ2
w, 3.3 where M
a1 a2
b1 b2
, B1
1 0 0 0
, B2
0 1 0 0 , B3 0 0 1 0
, B4
0 0 0 1 . 3.4
Therefore, for allw∈DA
A11u, v a1
∞
j1
λjEjua2
∞
j1
λjEjvc1
∞
j1
λα1j Ejuc2
∞
j1
λα2j Ejv,
A12u, v b1
∞
j1
λjEjub2
∞
j1
λjEjvd1
∞
j1
λβ1j Ejud2
∞
j1
λβ2j Ejv.
3.5
If we put
Pj
Ej 0 0 Ej
, j1,2, 3.6
then3.3can be written as
Aw≡
A11u, v
A12u, v
∞ j1
λjMλα1j c1B1λα2j c2B2λβ1j d1B3λβ2j d2B4
Pjw, 3.7
and we have for allw∈X
w ∞
j1
Pjw, w2 ∞
j1
Consequently, system1.1can be written as an abstract differential equation in the Hilbert spaceXin the following form:
•
w−AwBωft, inΩ× 0, t∗,
w0, on 0, t∗×∂Ω,
w0 w0, inx∈Ω,
3.9
wheref ≡colf1, f2∈L20, T ;X andBω
1
ω 0
0 1ω
is a bounded linear operator fromU intoX.
4. Main Results
4.1. Generation of a
C
0-Semigroup
Theorem 4.1. Ifc2 d1 0, then, under hypotheses (H1)–(H3), the linear operator−Adefined by
3.3is the infinitesimal generator of strongly continuous semigroup{St}t≥0given by
Stw ∞
j1
eAjtP
jw, w∈X, 4.1
where
Mj −λjM−λα1j c1B1−λα2j c2B2−λβ1j d1B3−λβ2j d2B4, 4.2
AjMjPj. 4.3
Moreover, if
minc1λ11−α1ρ1, d2λ11−β2ρ1
>0, 4.4
then theC0-semigoup{St}t≥0is exponentially stable, that is, there exist two positives constantsc, δ such that
St ≤ce−δt, for allt≥0. 4.5
Proof. In order to apply theProposition 2.3, we observe that−Acan be written as follows:
−Aw∞ j1
where
Aj−
λjMλα1j c1B1λα2j c2B2λβ1j d1B3λβ2j d2B4
Pj. 4.7
Therefore,AjMjPjandAjPjPjAj.
Now, we have to verify condition a of theProposition 2.3. We shall suppose that 0< ρ1< ρ2.Then, there exists a set{Q1, Q2} ∈M2R 2of complementary projections onR2
such that
eMteρ1tQ
1eρ2tQ2. 4.8
IfG g11g21g12g22is the matrix passage from the canonical basis ofR2 to the basis composed
with the eigenvectors ofM, then
Q1
1 ρ1ρ2
g11g22 −g11g12
g21g22 −g12g21
, Q2
1 ρ1ρ2
−g12g21 g11g12
−g21g22 g11g22
. 4.9
Hence,
e−λjMte−λjρ1tQ
1e−λjρ2tQ2. 4.10
We have also
e−λαj1c1B1t
⎛ ⎝e−λ
α1
j c1t 0
0 1
⎞
⎠, e−λα2
j c2B2t
1 −λα2j c2t
0 1
,
e−λβj1d1B3t
⎛
⎝ 1 0
−λβ1 j d1t 1
⎞
⎠, e−λβ1
j d2B4t
⎛ ⎝1 0
0 e−λβj2d2t
⎞ ⎠.
4.11
From4.10-4.11into4.7we obtain
eAjte−λjρ1tQ
1e−λjρ2tQ2
KjtPj, 4.12
where
Kjt
⎛ ⎜ ⎝e
−λα1
j c1tλα2β1
j c2d1t2e−λ α1
j c1t −λα2
j c2te−λ α1
j c1λ β2
j d2t
−λβ1
j d1t e−λ
β2
j d2t
⎞ ⎟
Asc2d10 we get
Kjt
⎛ ⎝e−λ
α1
j c1t 0
0 e−λβj2d2t
⎞
⎠. 4.14
Asλj → ∞asj → ∞,then this implies the existence of a positive numbercand a real number δ such that eAjt ≤ ceδt, for every j ≥ 1.Therefore−Ais a strongly continious
semigroup{St}t≥0given by4.1. We can even estimate the constantscandδas follows. i If min{c1 λ11−α1ρ1, d2 λ11−β2ρ1} ≤ 0. As limj→ ∞{−λα1j c1 λ
1−α1
j ρ1} limj→ ∞{−λβ2j c1λj1−β2ρ1}−∞, then there exist constants
δ1max
−λα1 j
c1λ1j−α1ρ1
|λα1j c1λ1j−α1ρ1
≤0, j ≥1,
δ2max
−λβ2 j
d2λ1j−β2ρ1
|λβ2j d2λ1j−β2ρ1
≤0, j ≥1,
4.15
hence, if we put
δmax{δ1, δ2} ≥0, 4.16
c0
1 ρ1ρ2 max
g11g22,g11g12,g21g22,g12g21, 4.17
we easily obtain
eAjt≤4c
0e−δt, j≥1. 4.18
iiIf min{c1λ11−α1ρ1, d2λ11−β2ρ1}>0. If we put
δminλα11c1λ11−α1ρ1
, λβ21 d2λ11−β2ρ1
>0, 4.19
then we find that
eAjt≤4c
0e−δt, j≥1. 4.20
Therefore, the linear operator−Agenerates a strongly continuous semigroup{St}t≥0 onX given by expression4.1.
Finally, if min{c1 λ11−α1ρ1, d2 λ11−β2ρ1} > 0,we have already proved 4.20. Using
4.20into4.1we get that theC0-semigoup{St}t≥0is exponentially stable. The expression
Theorem 4.2. If
c2/0, d1/0, c1≥0, d2≥0, 4.21
then, under the hypotheses (H1)–(H3), the linear operator−A defined by3.3 is the infinitesimal generator of strongly continuous semigroup exponentially stable{St}t≥0defined by4.1. Specially, there exist two positives constantsc, δsuch that
St ≤ce−δt, ∀t≥0. 4.22
To prove this result, we need the following lemma.
Lemma 4.3. For every two real positives constantscandλ, one has for every0< δ < λ/c
cte−λt≤ 1 eλ/c−δe
−δct, ∀t≥0, 4.23
and for every0< δ < λ/√c
ct2e−λt ≤ 4
e2λ/√c−δe
−δ√ct, ∀t≥0. 4.24
Proof ofLemma 4.3.It is easy to verify that for every ε >0 :te−εt ≤1/eε, for allt≥0. Let 0< δ < λ/candελ/c−δ >0, then we get
te−λ/ct≤ 1 eλ/c−δe
−δt, ∀t≥0. 4.25
Hence, we get4.23.
Also, it is easy to verify that for every ε > 0 : t2e−εt ≤ 4/e2ε2, for all t ≥ 0. Let
0< δ < λ/√candελ/√c−δ >0, then we get
t2e−λ√ct≤ 4
e2λ/√c−δ2e
−δt, ∀t≥0.
4.26
Hence, from 4.26we get ct2e−λt √ct2e−λ/√c√ct ≤ 4/e2λ/√c−δ2e−δ√ctfor allt≥ 0 and 0< δ < λ/√c, which gives4.24.
With the same manner we can prove that for every 0< δ < λc−1/nand every n∈N∗ we have
tne−λc−1/nt≤ n n
eεne
and consequently, for every two real positives constantscandλand every n∈N∗we have
ctne−λt ≤ n n
eεne −δc−1/nt
, for allt≥0 and every 0< δ < λ. 4.28
Now, we are ready to proveTheorem 4.2.
Proof ofTheorem 4.2. By applyingProposition 2.3we start from formula4.12and we put
Kjt
K11,jt K12,jt
K21,jt K22,,jt
, 4.29
where
K11,jt e−λ
α1
j c1tλα2β1
j c2d1t
2e−λα1
j c1t, K
12,jt −λα2j c2te−λ α1
j c1λ β2
j d2t,
K21,jt −λβ1j d1t, K22,,jt e−λ
β2
j d2t, ∀j≥1.
4.30
To estimatee−λjρ1tK
11,jtwe have in taking into accountc1 ≥0
e−λjρ1λαj1c1t≤e−λ1ρ1t, ∀t≥0, 4.31
and applying theLemma 4.3cλjα2β1|c2d1|we get
λα2j β1c2d1t2e−λjρ1λ α1 j c1t
≤ 4
e2λ1−α2β1/2
j /
|c2d1|
ρ1
λjα1−α2β1/2/|c2d1|
c1−γ1
e−γ1λαj2β1/2 √
|c2d1|t,
4.32
for allt ≥ 0 and 0 < γ1 < λj1−α2β1/2/
|c2d1|ρ1 λjα1−α2β1/2/
|c2d1|c1. But we have
λj1−α2β1/2/|c2d1|ρ1λjα1−α2β1/2/
|c2d1|c1≥λ11−α2β1/2/
|c2d1|ρ1, for allj≥1.Then
we get for every 0< γ1<λ11−α2β1/2/
|c2d1|ρ1that
λα2j β1c2d1t2e−λjρ1λ α1 j c1t
≤ 4
e2λ1−α2β1/2
1 /
|c2d1|
ρ1−δ1
e−γ1λ1α2β1/2 √
|c2d1|t, ∀t≥0. 4.33
From4.31-4.33we get
e−λjρ1tK
where
σ114
⎛
⎝λ11−α2β1/2
|c2d1|
ρ1−δ1
⎞ ⎠
−1
, δ1min λ1ρ1, γ1λ1α2β1/2
!
|c2d1|
"
, 4.35
and 0< γ1 <λ11−α2β1/2/
|c2d1|ρ1.
ApplyingLemma 4.3and taking into account4.21we get with the same manner that for every 0< δ2<λ11−α2/|c2|ρ1
e−λjρ1tK
12,jt≤σ2e−δ2λ α2
1 |c2|t, ∀t≥0, j≥1, 4.36
where
σ2
1 eλ11−α2/|c2|
ρ1−δ2
, 4.37
and or every 0< δ3<λ11−β1/|d1|ρ1
e−λjρ1tK
21,jt≤σ3e−δ3λ β1
1 |d1|t, ∀t≥0, j≥1, 4.38
where
σ3
1 eλ11−β1/|d1|
ρ1−δ3
, 4.39
e−λjρ1tK
22,jt≤e−λ1ρ1t, ∀t≥0, j ≥1. 4.40
From4.34-4.40into4.12we get
eAjt≤4c
0σe−δt, ∀t≥0, j ≥1, 4.41
where c0is defined by4.17and
σ1σ1σ2σ3, 0< δ <min
δ1, δ2λα21 |c2|, δ3λβ11|d1|, λ1ρ1
. 4.42
Using4.41into4.1we get that theC0-semigoup{St}t≥0generated by−Ais exponentially
4.2. Approximate Controllability
Befor giving the definition of the approximate controllabiliy for the sytem3.9, we have the following known result: for allw0 ∈Xandf ∈ L2 0, T;U,the initial value problem3.9
admits a unique mild solution given by
wt Stw0
#t
0
St−τBωfτdτ, t∈0, T . 4.43
This solution is denoted bywt;f.
Definition 4.4. System3.9is said to beapproximately controllableat timet∗whenever the set Ft∗{wt∗;f| ∀f ∈L2 0, t∗;U}is densely embedded inX; that is,
∀w0, w1 ∈X, ∀ε >0; ∃f∈L2 0, t∗;U:w
t∗;f−w1< ε. 4.44
The following criteria for approximate controllability can be found in6 .
Criteria 1. System3.9is approximately controllable on0, t∗ if and only if
B∗S∗tw0, ∀t∈0, t∗ ⇒w0. 4.45
Now, we are ready to formulate the third main result of this work.
Theorem 4.5. If the following condition
c2d10 4.46
is satisfied; then, under hypotheses (H1)–(H5), for all t∗>0and all open subsetω⊂Ω,system3.9
is approximately controllable on0, t∗ .
Proof. The proof of this theorem relies on the Criteria1and the following lemma.
Lemma 4.6. Let{α1j}j≥1, {β1j}j≥1 and {α2j}j≥1,{β2j}j≥1 be sequences of real numbers such that α11 > α12 > α13 >· · ·,α21 > α22 > α23 >· · · andα1j > α2j, for allj ≥0, then for any t∗ ∈R∗one
has
∞
j1
eα1jtβ
1jeα2jtβ2j
0, ∀t∈0, t∗ ⇒β1jβ2j 0, ∀j≥1. 4.47
Proof ofLemma 4.6.By analyticity we get∞j1eα1jtβ
1jeα2jtβ2j 0, ∀t≥0 and from this we getβ11
∞
j2eα1j−α11tβ1j
∞
j1eα2j−α11tβ2j0, ∀t≥0. Under the assumptions of the lemma we get∞j2eα1j−α11tβ
1j∞j1eα2j−α11tβ2j → 0 as t → ∞and so β11 0.Ifα12 > α21, we
divide∞j2eα1jtβ 1j
∞
j1eα2jtβ2j0 byeα12tand we passt → ∞we getβ120. Ifα21> α12,we
divide∞j2eα1jtβ
we divide∞j2eα1jtβ
1j∞j1eα2jtβ2j 0 byeα12tand we passt → ∞and getβ12β21 0.But
in this we case we can integrate under the symbol of sommation over the intervall0, t and we getβ12eα21tβ21eα12t0. Henceβ12β21 0. Continuing this way we see thatβ1j β2j0, for all j≥1.
We are now ready to proveTheorem 4.5. For this purpose, we observe that
B∗ωBω, S∗tw ∞
j1
eM∗jtP∗
jw, w∈X, t≥0, 4.48
where{St}t≥0is the C0-semigroup generated by−A.
Without lose of generality, we suppose that 0< ρ1< ρ2.Hence
Bω∗S∗tw ∞
j1
B∗ωeM∗jtP∗
jw ∞
j1
Bω∗eM∗jtP∗
jw ∞
j1 2
s1
B∗ωK∗jte−λjρstP∗
sj
w, 4.49
wherePsjQsPjPjQs, s1,2.
Now, suppose forw∈XthatBω∗S∗tw0, for allt∈0, t∗ .Then
B∗ωS∗tw0⇐⇒ ∞
j1 2
s1
Bω∗K∗jte−λjρstP∗
sj
wx 0, ∀x∈Ω. 4.50
If4.46is satisfied, then4.50take the form
∞
j1 2
s1
⎛ ⎝e−
λjρsλαj1c1t 0
0 e−λjρsλβj2d2t
⎞
⎠Bω∗Psj∗wx 0, ∀x∈Ω. 4.51
Then, from lemma4.6we obtain that for s1,2 and allx∈ω
B∗ωQ∗sPj∗wx Q∗s
⎛ ⎜ ⎜ ⎜ ⎜ ⎜ ⎝ mj
k1
u, ϕjk
1ωϕjkx
mj
k1
v, ϕjk
1ωϕjkx
⎞ ⎟ ⎟ ⎟ ⎟ ⎟ ⎠ 0 0
, j≥1. 4.52
SinceQ1Q2IR2,we get that allx∈ω
⎛ ⎜ ⎜ ⎜ ⎜ ⎜ ⎝ mj
k1
u, ϕjk
1ωϕjkx
mj
k1
v, ϕjk
1ωϕjkx
⎞ ⎟ ⎟ ⎟ ⎟ ⎟ ⎠ 0 0
On the other hand, fromTheorem 2.4we know that ϕjkare analytic functions, which implies the analticity ofEjumkj1u, ϕjkϕjkandEjvmkj1v, ϕjkϕjk.Then we can conclude that fors1,2 and allx∈Ω
⎛ ⎜ ⎜ ⎜ ⎜ ⎜ ⎝
mj
k1
u, ϕjk
ϕjkx
mj
k1
v, ϕjk
ϕjkx
⎞ ⎟ ⎟ ⎟ ⎟ ⎟ ⎠
0 0
, j≥1. 4.54
Hence Pjw 0, for all j ≥ 1, which implies that w 0. This completes the proof of
Theorem 4.5.
References
1 S. Badraoui, “Asymptotic behavior of solutions to a 2×2 reaction-diffusion system with a cross diffusion matrix on unbounded domains,”Electronic Journal of Differential Equations, vol. 2006, no. 61, pp. 1–13, 2006.
2 H. Larez and H. Leiva, “Interior controllability of a 2×2 a reaction-diffusion system with cross diffusion matrix,” to appear inBoundary Value Problems.
3 E. Zeidler,Applied Functional Analysis, vol. 109 ofApplied Mathematical Sciences, Springer, New York, NY, USA, 1995.
4 H. Leiva, “A lemma on C0-semigroups and applications,”Quaestiones Mathematicae, vol. 26, no. 3, pp.
247–265, 2003.
5 S. Axler, P. Bourdon, and W. Ramey,Harmonic Function Theory, vol. 137 ofGraduate Texts in Mathematics, Springer, New York, NY, USA, 1992.