Hindawi Publishing Corporation Abstract and Applied Analysis Volume 2011, Article ID 629692, 21 pages doi:10.

1155/2011/629692

Research Article Properness and Topological Degree for Nonlocal Reaction-Diffusion Operators
N. Apreutesei1 and V. Volpert2
1 2

Department of Mathematics, Gh. Asachi Technical University of Iasi, Bd. Carol. I, 700506 Iasi, Romania Institut Camille Jordan, UMR 5208 CNRS, Universit Lyon 1, 69622 Villeurbanne, France e

Correspondence should be addressed to N. Apreutesei, [email protected] Received 5 December 2010; Accepted 25 February 2011 Academic Editor: Nobuyuki Kenmochi Copyright q 2011 N. Apreutesei and V. Volpert. 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. The paper is devoted to integro-dierential operators, which correspond to nonlocal reactiondiusion equations considered on the whole axis. Their Fredholm property and properness will be proved. This will allow one to dene the topological degree.

1. Introduction
Consider the semilinear parabolic equation 2 u x2

u t

F u, J u ,

1.1

where

J u

x y u y, t dy.

1.2

Here : is a bounded function, not necessarily continuous, 0 on . The support of the function is supposed to be bounded, supp N, N . We will also assume that 1. Conditions on the function F will be specied below. y dy Integro-dierential equations of this type arise in population dynamics see 1, 2 and references therein . They are referred to as nonlocal reaction-diusion equations. A travelling

Abstract and Applied Analysis w x ct .

wave solution of 1.1 is a solution of this equation of the particular form u x, t It satises the equation w cw F w, J w 0.

1.3

The constant c is the wave speed. It is unknown and should be found together with the function w x . There are numerous works devoted to the existence 37 , stability and nonlinear dynamics 1, 1, 2, 816 of travelling wave solutions of some particular cases of 1.1 . Properties of travelling waves are determined by the properties of the integrodierential operator in the left-hand side of 1.3 . In this paper we will study the Fredholm property of this operator and its properness. We will use them to dene the topological degree and will discuss some applications. Let E C2 , E0 C , 0 < < 1 the usual Holder spaces endowed with the norms u x u y xy

E0

sup
x,y

sup|u x |,
x

E0

E0

E0

1.4

We are interested in the solutions of equation 1.3 with the limits w at , where the values 0. We are looking for the solutions w of 1.3 under the form w are such that F w , w w for x 1 and x w for x 1. Thus w u , where C , such that x 1.3 becomes u c u F u , J u 0. 1.5

Denote by A the operator in the left-hand side of 1.5 , that is A : E E0 , Au u c u F u , J u . 1.6

Suppose that F is dierentiable with respect to both variables. The linearization of A about a function u1 E is the operator L : E E0 , Lu A u1 u F u1 u F u1 U 1.7

cu

, J u1

, J u1

J u,

where F/u and F/U are the derivatives of F u, U with respect to the rst and to the second variable, respectively. For the linearized operator L, we introduce the limiting operators. Since for w1 u1 , w , it follows that J w1 J u1 w as x there exist the limits limx w1 x and the limiting operators are given by L u u cu a u b J u , 1.8

Abstract and Applied Analysis where

F w , w , u

F w , w . U

1.9

We will now recall the main denitions and results concerning the essential spectrum and Fredholm property for linear operators and the properness of nonlinear operators.

1.1. Essential Spectrum and Fredholm Property

Let us recall that a linear operator M : E1 E2 acting from a Banach space E1 into another Banach space E2 is called a Fredholm operator if its kernel has a nite dimension, its image is closed, and the codimension of the image is also nite. The last two conditions are equivalent 0 to the following solvability condition: the equation Lu f is solvable if and only if i f for a nite number of functionals i from the dual space E2 . Suppose that E1 E2 . By denition, the essential spectrum of the operator L is the set of all complex for which the operator LI, where I is the identity operator, does not satisfy the Fredholm property. The essential spectrum of general elliptic boundary value problems in unbounded domains can be determined in terms of limiting operators 17 . For the integrodierential operators under consideration, since they have constant coecients at innity, the essential spectrum can be found explicitly. It is proved 5, 6 that the operator L I is normally solvable with a nite-dimensional kernel if and only if the equations L u u do not have nonzero bounded solutions. Applying the Fourier transform to the last equations, we obtain ,

ci

b ,

1.10

where is the Fourier transform of the function x . Thus, the operator L is normally solvable with a nite-dimensional kernel if and only if the curves on the complex plane do not pass through the origin. Under some additional conditions, it can be also shown that the codimension of the operator is nite, that is, it satises the Fredholm property, and its index can be found. A nonlinear operator B : E1 E2 is called Fredholm if the linearized operator B satises this property. In what follows we will use the Fredholm property in some weighted spaces see below .

1.2. Properness and Topological Degree

An operator B : E1 E2 is called proper on closed bounded sets if the intersection of the inverse image of a compact set K E2 with any closed bounded set in E1 is compact. For the sake of brevity, we will call such operators proper. It is an important property because it implies that the set of solution of the operator equation B u 0 is compact.

Abstract and Applied Analysis

It appears that elliptic or ordinary dierential operators are not generally proper when considered in Holder or Sobolev spaces in unbounded domains. We illustrate this situation with a simple example. Consider the equation w H w 0, x , 1.11

where H w w w 1 . It can be veried that this equation has a positive solution w x , which converges to zero at innity. This convergence is exponential. So the solution belongs to Holder and to Sobolev spaces. Along with the function w x , any shifted function w x h , h is also a solution. Hence there is a family of solutions, and the set of solutions is not compact. Similar examples can be constructed for the integro-dierential equation. k In order to obtain proper operators, we introduce weighted spaces C with a growing at innity polynomial weight function x . The norm in this space is given by the equality u
k C

Ck

1.12

Let us return to the previous example. The family of functions w x h is not uniformly bounded in the weighted space. If we take any bounded closed set in the function space, it can contain the solutions w x h only for a compact set of the values of h. Therefore the set of solutions is compact in any bounded closed set. This example shows the role of weighted spaces for the properness of the operators. Properness of general nonlinear elliptic problems in unbounded domains and in weighted spaces is proved in 18 . In this work, we will prove properness of the integrodierential operators. After that, using the construction of the topological degree for Fredholm and proper operators with the zero index 18 , we will dene the degree for the integro-dierential operators. We will nish this paper with some applications of these methods to travelling waves solutions.

2. Properness in Weighted Spaces

In this section we study the properness of the semilinear operator A. Denition 2.1. If X, Y are Banach spaces, an operator A : X Y is called proper if for any compact set D Y and any bounded closed set B X, the intersection A1 D B is a compact set in X. Remark 2.2. The operator A : E C2 E0 C may not be proper from E to E0 see the comments related to 1.11 from the introduction . We will show in the sequel that A is proper in some weighted spaces associated to E and E0 . Let : be the function given by x 1 x2 , x . Denote E C2 , 0 C endowed with the usual norms E and E0 . We will work in the weighted E 0 u E and Holder spaces E and E , which are E and E0 , respectively, with the norms u u E0 . u 0 We begin with the following estimate for the integral term J u .

Abstract and Applied Analysis Lemma 2.3. Suppose that x supp N, N is bounded,

5 1 x2 and : is a function such that 0 on , y dy 1 and E0 . Then

0

J u for some constant K > 0. Proof. If supp

K u

u E0 ,

2.1

N, N , then supp x
x N

x N, x

N . First we write 2.2

x J u x
xN

x x y y u y dy. y

Since x / y is bounded for |x y| N and K1 u C , for some positive constant K1 . For every x1 , x2 , x1 / x2 , denote H x1 , x2

y dy

1, we have J u

x1 J u x1 x2 J u x2 . |x1 x2 |
C

2.3

If |x1 x2 | 1, we have H x1 , x2 2K1 u H x1 , x2 1 |x1 x2 |

x1 N x1 N

. If |x1 x2 | < 1, then

x1 x1 y x2 x2 y u y dy x2 x2 y u y dy
x2 N x2 N

x1 N x1 N

x2 x2 y u y dy

x1 N x1 N

x1 x2 x1 y y |x1 x2 |

x2 x1 y x2 y y |x1 x2 |

y u y dy 1 |x1 x2 | 1 |x1 x2 |
x2 N x1 N x1 N x2 N

x2 x2 y y u y dy y x2 x2 y y u y dy . y 2.4

Since |x1 x2 | < 1 and |x1 y| N, then |x2 y| N of and of x1 x2 / y |x1 x2 | , implies that H x1 , x2 K2 u
C

1. In this case, the boundedness

K3

E0

2K4 |x1 x2 |1 u

2.5

for some K2 , K3 , K4 > 0. Thus the desired estimate holds and the lemma is proved.

Abstract and Applied Analysis

0 We study the operator A acting from E into E . In order to introduce a topological degree in Section 4 , we prove the properness of A in the more general case when the 0 coecient c and function F depend also on a parameter 0, 1 . Let A : E E , 0, 1 be the operator dened through

A u

c u

F u

, J u

2.6

We note that the linearization L of A about a function u1 E is L u u c u F u1 u , J u1 u F u1 U , J u1 J u, 2.7

while its limiting operators are given by L u u c u F w , w u u F w , w J u . U 2.8

Assume that the following hypotheses are satised H1 For any 0, 1 , the function F u, U and its derivatives with respect to u and U satisfy the Lipschitz condition: there exists K > 0 such that |F u1 , U1 F u2 , U2 | K |u1 u2 | |U1 U2 | , 2.9

for any u1 , U1 , u2 , U2 2 . Similarly for F /u and F /U: F u2 , U2 F u1 , U1 u u F u2 , U2 F u1 , U1 U U K |u1 u2 | K |u1 u2 | |U1 U2 | , 2.10 |U1 U2 | .

H2 c , F u, U and the derivatives of F u, U are Lipschitz continuous in , that is, there exists a constant c > 0 such that |c c 0 | c| 0 |, F0 u, U F u, U u u c| 0 |, |F u, U F0 u, U | c| 0 |, F0 u, U F u, U U U c| 0 |, 2.11

, 0 0, 1 , for all u, U from any bounded set in 2 . H3 Condition NS For any 0, 1 , the limiting equations u c u F w , w u u F w , w J u U 0 2.12

do not have nonzero solutions in E.

Abstract and Applied Analysis Lemma 2.4. Suppose that conditions H1 - H2 hold. If n 0 and un u0 in C Fn un Proof. We have the equality Fn un , J un Fn un Fn u0 F0 u0 , J u0 . , J un F0 u0 , J u0
0

, then 2.13

0.

, J un , J u0

Fn u0 F0 u0

, J u0 , J u0

2.14

Condition H1 leads to the estimate of the rst dierence Fn un , J un Fn u0 , J u0 K |un u0 | |J un u0 | . 2.15

In view of hypothesis un u0 in C , the above inequality allows us to conclude that the weighted norm converges to zero. In order to estimate the second dierence, we begin with the following representation: Fn u0 , J u0
1

Fn , J , J U tJ u0
1

J u0
0

Fn u0

dt

u0
0

Fn tu0

, J u

2.16 dt.

Similarly, F0 u0 , J u0
1

F0 , J , J U tJ u0
1

J u0
0

F0 u0

dt

u0
0

F0 tu0

, J u

2.17 dt.

Therefore, Fn u0 , J u0
1

F0 u0

, J u0

dt
1 0

J u0
0 1

Fn u0

, J U

tJ u0

F0 u0

, J U dt

tJ u0

dt 2.18

u0
0

Fn tu0

, J u

dt A

1 0

F0 tu0

, J u

Fn , J

F0 , J

A denotes the rst line in the right-hand side, B: the second, C: the third . The expressions A and B converge to zero in the weighted norm of E C , due to the Lipschitz condition

Abstract and Applied Analysis

with respect to of the derivatives of F see H2 . The expression C is a function with a nite support. It also converges to zero in the weighted norm as n 0 . This concludes the proof. We can now prove the properness of the -dependent operator A . Denote by Eloc E,loc
0 0 and similarly Eloc and E,loc . 0 Theorem 2.5. If E0 , under assumptions H1 H3 , the operator A u : E 0, 1 E from 2.6 is proper with respect to u, on E 0, 1 . 0 0 Proof. Consider a convergent sequence fn E , say fn f0 in E . Let un , n be a solution fn , such that in E 0, 1 of the equation An un

u C2 u C

I , I I , I

bounded interval , 2.19 bounded interval

un

M,

n 1.

2.20

We prove that one can choose a convergent in E subsequence of the sequence un . Without loss of generality we may assume that n 0 as n . Equation An un fn can be written as un c n u n Fn un , J un fn . 2.21

Multiplying the equation by and denoting vn x that vn 2 c n vn , J un

x un x , g n x
2

x fn x , we derive

c n c n

vn gn .

2.22

Fn un

Indeed, since un un un vn vn / and un un un 2 un 2 vn / vn /2 , by 2.21 one easily obtains 2.22 . The sequence vn un is uniformly bounded in E: vn
E

vn vn /

un

un

M,

n 1.

2.23

Then it is locally convergent on a subsequence. More exactly, for every bounded interval N, N of x, there is a subsequence denoted again vn converging in C2 N, N to a limiting function v0 C2 N, N . By a diagonalization process we can prolong v0 to such that v0 E. Since vn E M, n 1, we can easily see that v0 E M. Let u0 be the limit that corresponds to un . Then un u0 in Eloc and v0 u0 .

Abstract and Applied Analysis

We now want to pass to the limit as n in 2.21 and 2.22 . To this end observe that H2 implies that Fn un , J un Fn un F0 un c1 |n 0 | F0 u0 , J u0 , J un , J u0 F0 u0
0

, J un , J un F0 un

F0 un F0 u0

0 0

2.24 .
0 in E,loc , we

, J un

, J u0

0 0 0 Since F0 is continuous from E E to E see H1 derive that

and J un

J u0

Fn un

, J un

F0 u0

, J u0

0 as n , in E,loc .

2.25

Passing to the limit as n , uniformly on bounded intervals of x in 2.21 and 2.22 , one obtains that u0 v0 2 c 0 u 0 F0 u0 , J u0
2

f0 , v0 f0 .

2.26

c 0 v0 , J u0

c 0 c 0

2.27

F0 u0

Subtracting 2.27 from 2.22 and denoting Vn 2 v0

vn v0 , one nds
2

Vn

c n V n , J un

c n

Vn 2.28

Fn un

F0 u0

, J u0

c n c 0

v0

fn f0 .

Recall that Vn vn v0 0 as n in Eloc . We show that Vn 0 in C . Suppose that it is not the case. Then, without any loss of generality, we can chose a sequence xn such that |Vn xn | > 0. This means that |vn xn v0 xn | > 0. Let Vn x Then, Vn 0 |Vn xn | > 0. 2.30 Vn x xn vn x xn v0 x xn x xn un x xn u0 x xn . 2.29

10 Writing 2.28 in x 2 xn , one obtains x xn x xn 2

Abstract and Applied Analysis

Vn x

c n V n x x xn x xn , J un v0
2

x xn x xn

c n

x xn x xn , J u0

Vn x x xn

2.31

xn Fn un

F0 u0 x xn

c n c 0

v0

fn f0 x

xn .

We will pass to the limit as n in 2.31 . First we note that by 2.29 and 2.23 , there exists V0 E such that Vn V0 as n in Eloc . Next, it is obvious that 1 x xn 0, x xn 0, x xn x xn 0, x xn n ,

2.32

0 while condition fn f0 in E leads to fn f0 x xn 0. Inequality 2.23 implies a similar estimate for v0 , so v0 x xn and v0 x xn are bounded in E. We also have x xn 0 for x xn > 1 and for x xn < 1 and

xn Fn un x x x

, J un

F0 u0

, J

u0

xn x x x xn xn xn . 2.33

xn Fn un xn F0 un xn F0 u0

, J un , J un , J un

F0 un F0 u0 F0 u0

, J un , J un , J u0

n n n Denote by T1 , T2 , and T3 the three terms in the right-hand side. Hypothesis H2 for F0 infers that

n T1 0,

0 n in Eloc .

2.34

Next, 2.29 leads to F0 un , J un F0 u0 , J un un x xn u0 x xn 1 s u0 , J un , J un x xn , x xn

n T2

Vn x Vn x Vn x

F0 s un u F0 sun u

xn

2.35

1 s u0

Abstract and Applied Analysis

11

for some s 0, 1 . By 2.20 we obtain |un x xn | M/ x xn , |u0 x xn | M/ x xn , hence sun

1 s u0 x

xn w ,

J un

xn

xn y un y dy z, it follows that

xn y y dy.

2.36

By the change of variable xn y

J un

xn

x z un xn

z dz

x z xn

z dz w ,

2.37

uniformly on bounded intervals of x. Hypothesis H1 shows that

n T2

F0 w , w V0 , u

0 as n in Eloc .

2.38

On the other hand,

n T3

J un F0 u0

J u0 , J un J un

F0 u0 1 s J u0 , J u0 J u0 x x 1 xn xn , x2 , x , with the 2.39

In x

F0 u0 U

, sJ un

for some s 0, 1 , where In x aid of 2.29 , we arrive at

J un

J u0

. For x

In x

x z

xn x z Vn z dz xn

x z V0 z dz

J V0 ,

2.40

0 in Eloc . As above, since J un x xn 0, J u0 x on bounded intervals of x, we deduce that

xn 0, J x xn w uniformly

n T3

F0 w , w J V0 , U

0 as n in Eloc .

2.41

Now we may pass to the limit in 2.31 . With the aid of 2.33 2.41 and H2 , one arrives at V0 c 0 V 0 F0 w , w V0 u F0 w , w J V0 U 0, 2.42

which contradicts H3 . Therefore we have proved that Vn 0 in C

12

Abstract and Applied Analysis

Now we have to show that Vn 0 in E C2 . To this end, we write 2.28 in the hn , where S Vn is the linear part from the left-hand side and form S Vn hn fn f0 Fn un c n c 0 v0 , J un . F0 u0 , J u0 2.43

v0

Using Lemma 2.1 from 6 for the linear operator S, we can write Vn
E

S Vn

E0

Vn

2.44

We make use of Lemma 2.4, hypothesis H2 for c n c 0 , and of the convergence fn f0 0 in E , to deduce that S Vn hn 0 in E0 C . Since Vn 0 in C , we conclude that un u0 in E . The theorem is proved.

3. Fredholm Property in Weighted Spaces

Consider the operator L : E L u u c u
2 C 0 E C ,

F u1 u

, J u1

F u1 U

, J u1

J u,

3.1

and its limiting operators L u u c u F w , w u u F w , w J u . U 3.2

Recall here condition NS for L , that is, hypothesis H3 : for each 0, 1 , the limiting equations L u 0 do not have nonzero solutions. We prove now the Fredholm property of L as an operator acting between the above weighted Holder spaces.
2 0 Theorem 3.1. If condition NS is satised, then the operator L : E C E between weighted spaces) is normally solvable with a nite-dimensional kernel. C (acting

Proof. Like in Theorem 2.2 from 6 , we can prove that L from E to E0 is normally solvable with a nite-dimensional kernel. To verify the property in the weighted spaces, we use Lemma 2.24 in 18 : if L : E E0 is normally solvable with a nite-dimensional kernel 0 L u L u is compact, then L : E E is and the operator K : E E0 , Ku normally solvable with a nite-dimensional kernel. Let {ui } be a sequence such that ||ui ||E M. We prove the existence of a subsequence C . Consider the sequence vi ui . Since vi of {Kui } which converges in E0 ui E M, one can nd a subsequence, denoted again {vi }, which converges locally in C2 to a function v0 , which can be prolonged to by a diagonalization process. We have v0 E, v0 E M and vi v0 in Eloc in C2 I , for every bounded interval I .

Abstract and Applied Analysis Let u0 be such that v0 Kui Ku0 Observe that zi K u0 . Then K zi ,
E0

13

E0

zi

vi v0

ui u0 .

3.3

2M and zi 0 in Eloc . Now we can write L zi L zi , J u1 2

2

zi

c J zi .

zi 2

z i

3.4

F u1 U But J zi

zi J

J zi

x y zi y zi x

x 1 dy y x 1 d. x

3.5

Since x / x 1 h x , where h x 0 as x , zi x is uniformly bounded and zi x 0 as i locally with respect to x, it follows that J zi / J zi 0 as i , uniformly with respect to x on . Similarly, zi , zi are uniformly bounded, zi 0, zi 0 as i locally and / 0, / 0 as x , so the rst two terms from 3.4 tend to zero uniformly with respect to x , as i . This implies that K zi / 0 as i in C . Therefore, with the aid of the local convergence zi 0 in C2 , we conclude that Kui Ku0 as i in E0 C . The theorem is proved.
0 We prove now the Fredholm property for L : E E , under an additional hypothesis. To this end, let I be the identity operator on E .

Condition NS[]
For each 0, 1 , the limiting equations L u u 0 associated to the operator L I do not have nonzero solutions in E , for any 0. We recall an auxiliary result from 6 which will be employed below. Lemma 3.2. Consider the operators L0 , L1 : C2 C dened by L0 u L u u, L1 u 1 s L0 sL1 , s 0, 1 . Then u u 0 and the homotopy Ls : C2 C , Ls 0 do not have nonzero there exists 0 large enough such that the limiting equations Ls u solutions for any s 0, 1 .
0 Theorem 3.3. If Condition NS is satised, then L , regarded as an operator from E to E , has the Fredholm property and its index is zero. 2 0 Proof. We put L0 u L u u, L1 u u u and Ls : E C E s 0 1 L 1 s L sL , s 0, 1 . Condition NS for L implies Condition NS for L0 C , L I.

14

Abstract and Applied Analysis

0 Then, Theorem 3.1 ensures that L0 L I, regarded from E to E , is normally solvable 0 with a nite-dimensional kernel. For operator L1 , we have ker L1 {0}, Im L1 E , hence L1 1 is a Fredholm operator and its index is ind L 0. By Lemma 3.2 applied for Ls , there exists 0 large enough such that Condition NS holds for all Ls , s 0, 1 . In view of Theorem 3.1, it follows that the operators Ls are normally solvable with a nite-dimensional kernel. In other words, the homotopy Ls gives a continuous deformation from the operator L0 to the operator L1 , in the class of the normally solvable operators with nite-dimensional kernels. Such deformation preserves the Fredholm property and the index. Since the index of L1 is zero, we derive that the index of all Ls is zero. In particular, for s 0 and 0, one arrives at the conclusion that L has the Fredholm property and its index is zero. This completes the proof.

4. Topological Degree
In this section we apply the topological degree construction for Fredholm and proper operators with the zero index constructed in 18 to the integro-dierential operators. Denitions. Recall in the beginning the denition of the topological degree. Consider two Banach spaces E1 , E2 , a class of operators acting from E1 to E2 and a class of homotopies H {A u : E1 0, 1 E2 , such that A u , 0, 1 }. 4.1

Let D E1 be an open bounded set and A such that A u / 0, u D, where D is the boundary of D. Suppose that for such a pair D, A , there exists an integer A, D with the following properties. i Homotopy invariance. If A u H and A u / 0, for u D, 0, 1 , then A0 , D A1 , D . 4.2

ii Additivity. If A , D is the closure of D and D1 , D2 D are open sets, such that and A u / 0, for all u D \ D1 D2 , then D1 D2 A, D A, D1 A, D2 . 4.3

iii Normalization. There exists a bounded linear operator J : E1 E2 with a bounded inverse dened on all E2 such that, for every bounded set D E1 with 0 D, J, D 1. 4.4

The integer A, D is called a topological degree.

Abstract and Applied Analysis

15

4.1. Degree for Fredholm and Proper Operators

We now recall a general result concerning the existence of a topological degree which was proved in 18, 19 . Let E1 and E2 be Banach spaces, E1 E2 algebraically and topologically and let G E1 be an open bounded set. Denote by I : E1 E2 the imbedding operator, Iu u, and by a class of bounded linear operators L : E1 E2 satisfying the following conditions: a the operator L I : E1 E2 is Fredholm for all 0, b for every operator L , there is 0 bounded inverse for all > 0 . Denote by F the class F B C1 G, E2 , B proper, B x , x G , 4.5 0 L such that L I has a uniformly

where B x is the Fr chet derivative of the operator B. e Finally, one introduces the class H of homotopies given by H B x, C1 G 0, 1 , E2 , B proper, B , F, 0, 1 . 4.6

Here the properness of B is understood in both variables x G and 0, 1 . Theorem 4.1 see 18 . For every B H and every open set D, with D G, there exists a topological degree B, D . Remark 4.2. Condition b can be weakened. Let E1 and E2 be two Banach spaces such that Ei Ei , i 1, 2 where the inclusion is understood in the algebraic and topological sense. In the case of the Holder space Ck , this can be the space Ck with an integer nonnegative k. k,p We can also consider some integral spaces W 17 . Instead of b above we can impose the following condition 20 : b for every operator L : E1 E2 , there is 0 bounded inverse for all > 0 . 0 L such that L I has a uniformly

4.2. Degree for the Integrodifferential Operators

0 Now, let E1 E and E2 E be the weighted spaces introduced in the previous section, with x 1 x2 , x . We will apply Theorem 4.1 for the integro-dierential operator A of the w , for x 1, x w for x 1 and form 1.6 , where function C , x

H4 F u, U and its derivatives with respect to u and U are Lipschitz continuous in u, U ;

16 H5 the limiting equations u cu F w , w u u

Abstract and Applied Analysis

F w , w J u u U

4.7

do not have nonzero solutions in E, 0. Under these hypotheses, Theorem 2.5 assures that operator A is proper. Moreover, its Fr chet derivative is A e L from 1.7 and it is a Fredholm operator with the index zero Theorem 3.3 . Consider F the class of operators A dened through 1.6 , such that H4 - H5 are 0 satised. Consider also the class H of homotopies A : E E , 0, 1 , of the form 2.6 , satisfying H1 - H2 and H6 for every 0, 1 , the equations u c u F w , w u u F w , w J u u U 0 4.8

do not have nonzero solutions in E, 0. By Theorem 2.5 and Theorem 3.3, we infer that operators A u are Fr chet dierentiable, proper with respect to u, e L verify condition a above. Condition b and their Fr chet derivatives A e follows from the lemma in the appendix. Hence H has the form 4.6 . Applying Theorem 4.1 for the class of operators F and the class of homotopies H, we are led to the following result. Theorem 4.3. Suppose that functions F and c satisfy conditions H1 - H2 and H4 H6 . Then a topological degree exists for the class F of operators and the class H of homotopies.

5. Applications to Travelling Waves

In this section we will discuss some applications of the Fredholm property, properness and topological degree to study travelling wave solutions of 1.1 . Let us begin with the classication of the nonlinearities. Denote F0 w F w, w . 5.1

We obtain this function from F w, J w if we formally replace the kernel x of the integral by the -function. The corresponding reaction-diusion equation u t 2 u x2 F0 u 5.2

is called bistable if F0 w < 0, monostable if one of these derivatives is positive and another one negative and, nally, unstable if F0 w > 0. As it is well-known, it can have travelling wave solutions, that is solutions, which satisfy the problem w cw F0 w 0, w w . 5.3

Abstract and Applied Analysis Let w0 x be a solution of 5.3 with some c solution, L0 u u c0 u

17 c0 . The operator L0 linearized about this

F0 w 0 u

5.4

has the essential spectrum given by two parabolas: 0 2 c0 i F0 w , . 5.5

Therefore the operator L0 satises the Fredholm property if and only if F0 w / 0. If this condition is satised, then the index of the operator is well dened. In the bistable case it equals 0, in the monostable case 1, in the unstable case 0 9 . In the case of the integro-dierential operator Lu u c0 u Fu w, J w u FU w, J w J u , 5.6

the essential spectrum is given by the curves 2 c0 i Fu w , w FU w , w , , 5.7

where is the Fourier transform of the function x . If we replace J u by u, that is, x by the -function, then the spectrum of the integro-dierential operator coincides with the spectrum of the reaction-diusion operator. We note that Fu w , w

FU w , w 0

F0 w , 5.8 x dx 1.

Re

x cos x dx <

5.1. Fredholm Property Bistable Case

Let F0 w we recall that 0 for all since Fu w , w Fu w , w FU w , w < 0 < 0 and FU w , w FU w , w 5.9 > 0. Then Re < 0

1 . Suppose that Fu w , w

Re FU w , w Fu w , w

F0 w .

5.10

Hence the essential spectrum is completely in the left-half plane. This allows us to prove properness of the corresponding operators in weighted spaces and to dene the topological degree.

18

Abstract and Applied Analysis

ci 0

ci 0

Figure 1: Schematic representation of a possible location of the essential spectrum of the operator L.

Consider now the case where Fu w , w > 0 and FU w , w < 0. The principal dierence with the previous case is that the essential spectrum of the integro-dierential operator may not be completely in the left-half plane Figure 1 though this is the case for the reaction-diusion operator. Depending on the parameters, the essential spectrum can cross the imaginary axis for some pure imaginary values. However the linear operator remains Fredholm since the essential spectrum does not cross the origin; the nonlinear operator remains proper in the corresponding weighted spaces. Thus, the bistable case for the reaction-diusion equation gives rise to two dierent cases for the integro-dierential equation. We will call both of them bistable but will distinguish them when necessary.

Monostable Case
Suppose that F0 w > 0 and F0 w < 0. Then 0 is in the left-half plane for all ; 0 is partially in the right-half plane, 0 0 > 0. The essential spectrum of the integro-dierential operator L given by the curves has a similar structure. It does not cross the origin, so that the operator satises the Fredholm property. The curve is partially in the right-half 0 0 > 0. The curve can be completely in the left-half plane or partially plane, 0 in the right-half plane Figure 1 . Similar to the bistable case, there are two subcases in the monostable case.

Index
In order to nd the index of the operator L, we consider the operator L which depends on the N , N . parameter characterizing the width of the support of the function , supp We recall that x dx 1. Let L1 L, that is the value 1 corresponds to the function in the operator L.

Abstract and Applied Analysis

19

Since the essential spectrum of the operator L can be determined explicitly, then we can arm that it converges to the essential spectrum of the operator L0 as 0. Moreover, L converges to L0 in the operator norm. The essential spectrum of the operator L does not cross the origin. Therefore it is normally solvable with a nite-dimensional kernel. Hence the index of the operator L equals the index of the operator L0 12 . It is 0 in the bistable case and 1 in the monostable case cf. 9 .

5.2. Topological Degree and Existence of Solutions

In the bistable case we can dene the topological degree for the integro-dierential operator and use the Leray-Schauder method to prove existence of solutions. In order to use this method we need to obtain a priori estimates of solutions. In 10 , a priori estimates are obtained in the case where F u, J u J u u 1 u u. 5.11

Thus, we can now conclude about the existence of waves for this particular form of the nonlinearity. More general functions will be considered in the subsequent works.

5.3. Local Bifurcations and Branches of Solutions

Other conventional applications of the degree are related to local bifurcations and global branches of solutions see, e.g., 14 . We can now use the corresponding results for the integro-dierential operator in the bistable case. Let us emphasize that these results apply in particular for the case where the essential spectrum of the linearized operator crosses the imaginary axis see above . Therefore the wave persists in this case unless a priori estimates are lost.

Appendix

Sectorial property of an operator implies certain location of its essential spectrum and an estimate of the resolvent. For general elliptic problems in unbounded domains it is proved in 20 . A simple particular case of second-order operators on the axis is considered in 21 . In the lemma below we prove an estimate of the resolvent using this last result. Lemma A.1. Let M0 : C2 R C R , M0 u u b x u x c x u d x J u , A.1

where the coecients of this operator are suciently smooth bounded functions. Then the operator M u M0 u u, considered as acting in the same spaces, has a bounded inverse with the norm independent of for 0 > 0, where 0 is suciently large.

20 Proof. Consider the equation M u We need to obtain the estimate u

C2

Abstract and Applied Analysis

f.

A.2

K f

A.3

of this equation where K is independent of for all suciently large. Here and below we denote by K the constants independent of u, f, and . We rst prove the estimate

u Since the operator Mu is sectorial 21 , then Mu u

A.4

b x u x

c x u

A.5

f d x J u

A.6

Estimate A.4 follows from the last one for suciently large. We can write A.2 in the form M0 u u f u u. A.7

We can choose > 0 such that the operator in the left-hand side is invertible. Hence u
C2

A.8

This estimate and A.4 give A.3 . The lemma is proved. This lemma remains valid for the operators acting in the weighted spaces.

Acknowledgments
This work was partially supported by the LEA Math Mode between CNRS France and Romanian Academy through the joint project Existence of travelling waves for nonlocal reaction-diusion equations.

Abstract and Applied Analysis

21

References
1 S. Genieys, V. Volpert, and P. Auger, Pattern and waves for a model in population dynamics with nonlocal consumption of resources, Mathematical Modelling of Natural Phenomena, vol. 1, no. 1, pp. 6582, 2006. 2 V. Volpert and S. Petrovskii, Reaction-diusion waves in biology, Physics of Life Reviews, vol. 6, pp. 267310, 2009. 3 S. Ai, Traveling wave fronts for generalized Fisher equations with spatio-temporal delays, Journal of Dierential Equations, vol. 232, no. 1, pp. 104133, 2007. 4 N. Apreutesei, N. Bessonov, V. Volpert, and V. Vougalter, Spatial structures and generalized travelling waves for an integro-dierential equation, Discrete and Continuous Dynamical Systems. Series B, vol. 13, no. 3, pp. 537557, 2010. 5 N. Apreutesei, A. Ducrot, and V. Volpert, Competition of species with intra-specic competition, Mathematical Modelling of Natural Phenomena, vol. 3, no. 4, pp. 127, 2008. 6 N. Apreutesei, A. Ducrot, and V. Volpert, Travelling waves for integro-dierential equations in population dynamics, Discrete and Continuous Dynamical Systems. Series B, vol. 11, no. 3, pp. 541 561, 2009. 7 H. Berestycki, G. Nadin, B. Perthame, and L. Ryzhik, The non-local Fisher-KPP equation: travelling waves and steady states, Nonlinearity, vol. 22, no. 12, pp. 28132844, 2009. 8 N. F. Britton, Spatial structures and periodic travelling waves in an integro-dierential reactiondiusion population model, SIAM Journal on Applied Mathematics, vol. 50, no. 6, pp. 16631688, 1990. 9 J. F. Collet and V. A. Volpert, Computation of the index of linear elliptic operators in unbounded cylinders, Journal of Functional Analysis, vol. 164, no. 1, pp. 3459, 1999. 10 I. Demin and V. Volpert, Existence of waves for a nonlocal reaction-diusion equation, Mathematical Modelling of Natural Phenomena, vol. 5, no. 5, pp. 80101, 2010. 11 A. Ducrot, Travelling wave solutions for a scalar age-structured equation, Discrete and Continuous Dynamical Systems. Series B, vol. 7, no. 2, pp. 251273, 2007. 12 I. C. Gohberg and M. G. Kren, The basic propositions on defect numbers, root numbers and indices of linear operators, American Mathematical Society Translations, vol. 13, pp. 185264, 1960. 13 S. A. Gourley, Travelling front solutions of a nonlocal Fisher equation, Journal of Mathematical Biology, vol. 41, no. 3, pp. 272284, 2000. 14 M. A. Krasnoselskii and P. P. Zabreko, Geometrical Methods of Nonlinear Analysis, vol. 263, Springer, Berlin, Germany, 1984. 15 R. Lefever and O. Lejeune, On the origin of tiger bush, Bulletin of Mathematical Biology, vol. 59, no. 2, pp. 263294, 1997. 16 B. Perthame and S. G nieys, Concentration in the nonlocal Fisher equation: the Hamilton-Jacobi e limit, Mathematical Modelling of Natural Phenomena, vol. 2, no. 4, pp. 135151, 2007. 17 A. Volpert and V. Volpert, The Fredholm property of elliptic operators in unbounded domains, Transactions of the Moscow Mathematical Society, vol. 67, pp. 127197, 2006. 18 V. Volpert and A. Volpert, Properness and topological degree for general elliptic operators, Abstract and Applied Analysis, vol. 2003, no. 3, pp. 129181, 2003. 19 V. A. Volpert, A. I. Volpert, and J. F. Collet, Topological degree for elliptic operators in unbounded cylinders, Advances in Dierential Equations, vol. 4, no. 6, pp. 777812, 1999. 20 V. Volpert, Elliptic partial dierential equations, in Fredholm Theory of Elliptic Problems in Unbounded Domains, vol. 1, Birkh auser, Boston, Mass, USA, 2011. a 21 A. I. Volpert, V. Volpert, and V. A. Volpert, Traveling Wave Solutions of Parabolic Systems, vol. 140 of Translations of Mathematical Monographs, American Mathematical Society, Providence, RI, USA, 1994.