• No results found

Large time behavior for a multidimensional chemotaxis model

N/A
N/A
Protected

Academic year: 2020

Share "Large time behavior for a multidimensional chemotaxis model"

Copied!
11
0
0

Loading.... (view fulltext now)

Full text

(1)

R E S E A R C H

Open Access

Large time behavior for a

multidimensional chemotaxis model

Lan Luo

*

*Correspondence:

[email protected] School of Mathematics and Information Science, Guangzhou University, Guangzhou, 510006, P.R. China

Abstract

We investigate the global existence and large time behavior of the solutions to the Cauchy problem of a multidimensional chemotaxis model with initial data close to a constant state. We also obtain the time decay rate ofkth-order derivatives of these solutions.

MSC: 35G25; 35M10

Keywords: chemotaxis model; global existence; large time behavior; time decay rate

1 Introduction

In this paper, we consider the Cauchy problem of the multidimensional chemotaxis model

⎧ ⎨ ⎩

∂tu=u+∇ ·(u∇lnv),

∂tv=uvμv,

(.)

where (t,x)∈R+×Rn,u(t,x) andv(t,x) denote the cell density and the chemical concen-tration, respectively, andμis a constant.

The system (.) was proposed by Othmer and Stevens [] to describe the chemotac-tic movement of parchemotac-ticles where the chemicals are non-diffusible and can modify the lo-cal environment for succeeding passages. For example, myxobacteria produce lime over which their cohorts can move more readily and ants can follow trails left by predecessors []. One direct application of (.) is to model haptotaxis where cells move toward an in-creasing concentration of immobilized signals such as surface or matrix-bound adhesive molecules.

For the sake of simplicity, we setw=μt+lnv. Therefore we get from (.)

⎧ ⎨ ⎩

∂tu=u+∇ ·(uw),

∂tw=u.

(.)

Lettingp¯>  be a constant, we substitute

u=p+p¯, q=∇w, (.)

(2)

into the system (.) to get

⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩

∂tpp¯divqp=div(pq),

∂tq–∇p= ,

∇ ×q= .

(.)

To represent the results in this paper, the following notations are needed. We use (·,·) to denote the standard inner products inL(Rn) and·is the corresponding norm inL(Rn). For anyr∈[,∞],Lr(Rn) is the usual Lebesgue space on Rn. We define the operators withsRby

sf(x) :=

Rn|

ξ|sfˆ(ξ)eπix·ξdξ,

wherefˆ(ξ) =Ff(ξ) is the Fourier transform off(x). We shall useDkwith an integerk to stand for the usual spatial derivatives of order k. Whenk<  ork is not a positive integer,Dkstands fork. We then recall the homogeneous Besov spaces. LetφCc(R)

be such thatφ(ξ) =  when|ξ| ≤ andφ(ξ) =  when|ξ| ≥. Letϕ(ξ) =φ(ξ) –φ(ξ) and

ϕj(ξ) =ϕ(ξj) for anyjZ. Then

jZϕj(ξ) =  ifξ= . We definej:=F–(ϕj)∗f, then, for ∈Rand ≤r,s≤ ∞, we define the homogeneous Besov spacesB˙r,s(Rn) with the norm · B˙r,sdefined by

fB˙r,s:= jZ

sj jfsLr

s

, fB˙r,∞:=sup

jZ

j

jfLr.

We also usek(p, q)(t):=kp(t)+kq(t)and denote byCa generic positive

constant which varies from line to line. For the energy estimates, the instant energy func-tional for a solution (p, q)(t,x) to the system (.), denoted by EN(t) has the equivalent relation with some integerN≥,

EN(t)∼ N

k=

Dk(p, q)(t). (.)

In addition, the corresponding dissipation functional denoted byDN(t) satisfies

DN(t)∼ N

k=

Dk(p, q)(t)+DN+p(t). (.)

Here the notationABis used to denote that there exist two generic positive constants

C<Csuch thatCBACB.

With the above preparation, the main results in this paper can be stated as follows.

Theorem . There exists> such that ifEN()≤with N>n+ and n≥,there is a

unique global solution(p, q) = (p, q)(t,x)to the system(.)with(p, q)(,x) = (p(x), q(x))

that satisfies

d

(3)

This global solution has the following time decay:

(p, q)(t)≤C( +t)– for any∈[,N]. (.)

If we further assume(p, q)B˙–,∞<∞with ∈(,

n

],we have

(p, q)(t)≤C( +t)–(+ ) for any∈[– ,N– ]. (.)

Remark . Similar time decay rates to (.) and (.) were obtained when the initial data in the additional negative homogeneous Sobolev spaceH˙– (Rn) with ∈(,n) for equa-tion (.) with some perturbaequa-tion term in [] the referee told us. Compared to the results in [], our results can include the endpoint cases while the equation considered in [] is a little more complicated. In some sense our results generalize their partial time decay rates results and we believe that our proof methods can also apply to the equation considered in [] to obtain the time decay rates of the endpoints. In addition, the other small assump-tions of the initial data are imposed in [, ], some time decay rates results can be obtained and our results also improve the time decay rate results in [, ]. As to the chemotaxis sig-nificance of the time decay rates of equation (.), we refer the reader to [–], where the authors give more discussions.

From Theorem ., (.) and (.), we have the following results on the asymptotic be-havior of solutions to the system (.).

Corollary . There exists>  such that ifNk=Dk(u

–p¯,∇lnv) ≤ with N > n

+ and n≥,then the Cauchy problem(.)has a global solution(u,v) = (u,v)(t,x)with

(u,v)(,x) = (u(x),v(x))satisfying for all t> 

N

k=

Dk(up¯,∇lnv)(t)≤C.

This global solution has the following time decay:

(up¯,∇lnv)(t)≤C( +t)– for any∈[,N].

If we further assume(u–p¯,∇lnv)B˙–,∞<∞with ∈(,

n

],we have

(up¯,∇lnv)(t)≤C( +t)–(+ ) for any∈[– ,N– ].

(4)

was in the Chemin-Lerner space framework (seee.g.[, ]). In [], the authors obtained the global weak existence and partial time decay of the solutions to the system (.) near some constant state in the slightly low regular Sobolev space. It was studied in [] and a comprehensive qualitative and numerical analysis were provided. We will not attempt to exhaust references in this paper. We refer reader to [, –, –, ] and the references therein for more discussions in this direction.

In this paper we consider precise large time behavior of the global solutions to the sys-tem (.) near some constant state, motivated by [, , ]. We first construct the uniform estimates, which are used to obtain global existence and time decay rate under the as-sumption of initial energy being small enough. Once a global solution are constructed, by using a time-weighted methods, we can prove the time decay rate of thekth-order deriva-tives of these solutions approaching a constant state at a time decay rateO(tk) without

additional assumption. We remark that the time-weighted methods was also applied to prove the time decay rate for the fluid system [, ]. If we further assume that the initial data is in the negative Besov space (not necessary small) by a similar interpolation to [, ], we can obtain the fast time decay rate.

2 Some basic estimates

In this section we first represent some elementary inequalities and then we use these re-sults to handle the nonlinear term.

Lemma .([]) For any f,gL(Rn)Hk(Rn),we have

Dk(fg)≤CfLDkg+gLDkf. (.)

Lemma .([]) Suppose that m= .We have the interpolation estimate kgLrCmg

–θ

Lg

θ

˙

B,C mg–θ

L g

θ

L, (.)

where <θ< and≤r≤ ∞and

k+n

 –

n

r =m( –θ) + θ. (.)

Lemma . If≤kN– ,it follows that Dkdiv(pq),divDkqCEN(t)Dk+(p, q)

(.)

and

Dkdiv(pq),DkpCE N(t)

N

m=k+

Dm(p, q). (.)

We have

(5)

Proof By (.) and Lemma ., we have, forθ= (n – )/[n],

pL∞≤CDp(–θ)D[

n

]+pθCEN(t). (.)

It follows from this and (.) that

Dkdiv(pq),divDkqC(p, q)LDk+(p, q)≤C

EN(t)Dk+(p, q)

.

It follows from (.) that

Dkdiv(pq),Dkp=Dk(pq),Dk+p

C(p, q)LDk(p, q)Dk+p. (.)

Then we see from (.) that (.) holds. Next we prove (.). Ifk= , (.) follows from (.) and (.). If ≤kn, it follows from (.) that

pLkpCαp k

k+[n]+p

k+k+pkk+pk+, (.)

whereα=n

– ([ n ] +  –

n )

k∈[,

n ).

IfαZ, by (.) and (.), we get (.). Ifα∈/Z, it follows from (.) that

αp[α]p[α]+–α[α]+–[α]. (.)

By this, (.) and (.), we obtain (.). If n

<kN– , we have from (.)

pLkpCp–θk+p

θ

k+p–θαpθ, (.)

whereθ=(kn+) andα= ( –n)(k+ )∈(,k+ ). IfαZ, by (.) and (.), we get (.).

Ifα∈/Z, we use (.), (.) and (.) to obtain (.).

3 Energy analysis

In this section we first deduce the uniform estimates of the system (.) and make use of the time-weighted methods to the time decay rate (.). Then we use the properties of the Besov space and the interpolation to get the fast time decay rate.

Theorem . Letting(p, q) = (p, q)(t,x)be the solution to the system(.)satisfyingEN(t) <

,we have the estimate

d

dtEN(t) +CDN(t)≤. (.)

HereEN(t)andDN(t)are defined as(.)and(.).

Proof Taking the derivativeDkof (.) with ≤kN, we can obtain 

d dtD

(6)

For the system (.), we have for ≤kN–  

d dtdivD

kq d

dt

Dkp,divDkq+p¯divDkq–∇Dkp

= –Dkdiv(pq),divDkq. (.)

For anyκ>  small enough, we have from (.) and (.)

d dt

N

k=

Dkp+p¯Dkq+κ N–

k=

divDkq– Dkp,divDkq

+ N–

k=

( –κ)∇Dkp+ κp¯divDkq+ DN+p

=  N

k=

Dkdiv(pq),Dkp– κ

N–

k=

Dkdiv(pq),divDkq. (.)

Notice that the equality∇xq=divq+∇ ×qholds on p. of []. By the fact that∇ ×q= , we havedivDkq=Dk+q. By this, for anyκ>  small enough, one

has

EN(t) = N

k=

Dkp+p¯Dkq+κ N–

k=

divDkq– Dkp,divDkq

N

k=

Dk(p, q),

DN(t) = N–

k=

( –κ)∇Dkp+ κp¯divDkq+ DN+p

N

k=

Dk(p, q)+DN+p.

By Lemma . we have

N

k=

Dkdiv(pq),Dkp +

N–

k=

Dkdiv(pq),divDkqC

EN(t)DN(t).

SinceEN(t) <, by these, we see from (.) that (.) holds.

Remark . By Theorem . we have deriveda prioriestimates. Based on this uniform estimate the global existence is proved by the standard continuation argument together with the local in time existence theory; for example, see [].

(7)

Lemma . Assume that (p, q) = (p, q)(t,x)be the solution to the system(.)satisfying EN(t) <.For any fixed m∈ {, , . . . ,N– },we have the following estimate:

d dtE

m(t) +CDm(t). (.)

HereEm(t)andDm(t)are defined by

Em(t) = N

k=m

Dkp+p¯Dkq+κ

N–

k=m

divDkq– Dkp,divDkq, (.)

Dm(t) = N–

k=m

( –κ)∇Dkp+ κp¯divDkq+DN+p, (.)

whereκ> is small enough.

Proof For anyκ>  small enough, we have from (.) and (.)

d dt

N

k=m

Dkp+p¯Dkq+κ N–

k=m

divDkq– Dkp,divDkq

+ N–

k=m

( –κ)∇Dkp+ κp¯divDkq+ DN+p

=  N

k=m

Dkdiv(pq),Dkp– κ

N–

k=m

Dkdiv(pq),divDkq. (.)

SincedivDkq=Dk+qandκ>  is small enough, one has from (.) and (.) that

Em(t) N

k=m

Dk(p, q), and Dm(t)∼ N

k=m+

Dk(p, q)+DN+p. (.)

By Lemma . we have

N

k=m

Dkdiv(pq),Dkp +

N–

k=m

Dkdiv(pq),divDkqC

EN(t)Dm(t).

SinceEN(t) <, by these, we see from (.) that (.) holds.

In the following we will use the time-weighted methods to show the time decay of the solutions to the system (.).

Theorem . Assume that(p, q) = (p, q)(t,x)be the solution to the system(.)satisfying EN(t) <.We have the time decay estimate,for any m∈[,N],

(8)

Proof For anym∈ {, . . . ,N– }, we see from (.) that

d dt

( +t)mEm(t)m( +t)m–Em(t) + ( +t)mDm(t). (.) Letting the constantsCm>  andCmCm+be large enough, we define the functional

E(t) = N–

m=

Cm( +t)mEm(t). (.)

By (.), (.) and (.) withm= , we see that

E(t) +C

N–

m=

t

( +s)mDm(s)dsE(). (.)

SinceEN(t) <, for anym∈ {, , . . . ,N– }, we have from (.) and (.)

Dm(p, q)≤CEm(t)≤C( +t)–m. (.)

By (.) and (.) one has

 

d dtD

Np

+p¯DNq+CDNp

(p, q)

LD

N(p, q)+εDNp.

Takingε>  small enough, we have

 

d dtD

Np

+p¯DNq+CDNp≤(p, q)

LD

N(p, q)

. (.)

It follows from (.) that

d dt

( +t)NDNp+p¯DNq–CN( +t)N–DN(p, q)

C( +t)N(p, q)L∞DN(p, q). (.)

By (.) and (.), one has(p, q)

L∞≤C( +t)–. It follows from (.) that ( +t)NDN(p, q)≤C

t

( +s)N–DN(p, q)(s)ds+CDN(p, q).

SinceEN(t) <, we have from (.) and the definition ofDN–(s) that

C t

( +s)N–DN(p, q)(s)ds+CDN(p, q) 

C.

By the above estimates one has

DN(p, q)≤C( +t)–N.

(9)

To show the fast time decay under the additional assumption, for this, we need to show the propagation of the solutions inB˙–,.

Lemma . For any ∈(,n],for the system(.),we have d

dt

p˙

B–,∞+p¯q

˙

B–,∞

+Cp˙

B–,∞≤Cp

L n

x

q. (.)

Proof Taking the operationjon the sides of (.), we have

 

d dtjp

p¯(

jdivq,jp) +∇jp=

divj(pq),jp

and

 

d dtjq

– (

jp,jq) = .

By these we have

 

d dt

jp+p¯jq

+∇jp=

divj(pq),jp

. (.)

We have

divj(pq),jp=j(pq),∇jp≤ jCpqLrjp

≤ jCεp

L n

x

q+εjp.

Here we have used the Bernstein type inequality in [] for =n r

n ∈(,

n

] andr∈[, ).

By using the definition ofB˙–,and takingε>  small enough, it follows from (.) that

(.) holds.

To finish the proof of the fast time decay we will make use of an interpolation method similar to [, ].

Proof of(.)in Theorem. By (.) we have, for any ∈(,n],

(p, q)(t)B˙

,∞≤C(p, q)

˙

B,+C

t

p

L n

x

qds. (.)

For ∈(,n–

 ], we have ≤[ n – ] <

n  and

p

L n

x

q≤CD[n– ]p(–θ)D[n– ]+pθq≤CEN(s)DN(s). (.)

Hereθ=n– – [n – ]∈(, ) for n– ∈/Zby Lemma . andθ =  for n – ∈Z. If

EN()≤, it follows from (.) that for anyt> 

EN(t) +C

t

(10)

The above three inequalities imply that if(p, q)B˙–,C

(p, q)(t)B˙

,∞≤C(p, q)

˙

B–,∞+

C.

For fixedm∈ {, , , . . . ,N– }we have from (.)

Dm(p, q)CDm+(p, q)

+m

+m+(p, q)  +m+

˙

B,CD

m+(p, q) ++mm+.

It follows from this, (.), (.) and (.) that

Em(t)+ +m

C N

k=m

Dk(p, q)

+ +m

C N–

k=m

Dk(p, q)

+ +m

+CDN(p, q)(+

 +m)

C

N

k=m

Dk(p, q)+CDN(p, q)≤CDm(t).

We have from (.) and the above inequality, form∈ {, , , . . . ,N– },

d dtE

m(t) +CEm(t)+ +m. (.)

By using (.), we solve this inequality to obtain

N

k=m

Dk(p, q)≤C( +t)–(m+ ). (.)

By using Lemma ., we can obtain (.) for ∈(,n– ].

Next we consider the case ∈(n– ,n) and(p, q)B˙–, <∞. In this case, we have

(p, q)˙

B–,∞ <∞for any

(, ) by interpolation:

(p, q)˙

B,C(p, q)

/

˙

B,(p, q) ( – )/

˙

B,

C(p, q)

/

˙

B,(p, q) ( – )/

<∞. (.)

Then (.) holds for any ∈(,n– ] and anym∈ {, , , . . . ,N– }. We set =n– with

n≥. By (.) and (.) we have

(p, q)(s)L n

x

(p, q)(s)

CD[n– ](p, q)(–θ)D[n– ]+(p, q)θ(p, q)(s)

C( +s)–(–θ)([n– ]+ )( +s)θ([n– ]++ )( +s)

(11)

Since ∈(n– ,n) and n≥, one has n – +n–  > . It follows from (.) that

(p, q)(t)B˙–,<∞for anyt≥. By this we may repeat the arguments leading to (.)

and prove that (.) holds for the case ∈(n– ,n).

Finally if =n and(p, q)B˙,<∞, it follows from (.) that(p, q)B˙

,∞<∞for any ∈(,n). Assume that =nδ

 withn≥ andδ>  small enough such that

(p, q)(s) L n

x

(p, q)(s)=(p, q)(s)≤C( +s)–(nδ)≤C( +s)–/.

By this we see from (.) that(p, q)(t)B˙–,<∞. By similar arguments to (.), we can

obtain (.) for the case =n. By using Lemma ., we can obtain (.) for any ∈(,n].

This completes the proof of Theorem ..

Competing interests

The author declares that she has no competing interests.

Acknowledgements

We would like to thank the referees for their helpful comments and pointing us out [3, 15]. The research of the author was supported by the NNSFC Grant 11371151.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Received: 16 January 2017 Accepted: 14 March 2017

References

1. Othmer, H, Stevens, A: Aggregation, blowup and collapse: the ABC’s of taxis in reinforced random walks. SIAM J. Appl. Math.57, 1044-1081 (1997)

2. Eisenbach, M: Chemotaxis. Imperial College Press, London (2004)

3. Wang, Z-A, Xiang, Z-Y, Yu, P: Asymptotic dynamics on a singular chemotaxis system modeling onset of tumor angiogenesis. J. Differ. Equ.260(3), 2225-2258 (2016)

4. Deng, C, Li, T: Well-posedness of a 3D parabolic-hyperbolic Keller-Segel system in the Sobolev space framework. J. Differ. Equ.257(5), 1311-1332 (2014)

5. Li, D, Li, T, Zhao, K: On a hyperbolic-parabolic system modeling chemotaxis. Math. Models Methods Appl. Sci.21(8), 1631-1650 (2011)

6. Li, T, Pan, R-H, Zhao, K: Global dynamics of a hyperbolic-parabolic model arising from chemotaxis. SIAM J. Appl. Math. 72(1), 417-443 (2012)

7. Zhang, M, Zhu, C-J: Global existence of solutions to a hyperbolic-parabolic system. Proc. Am. Math. Soc.135(4), 1017-1027 (2007)

8. Li, T, Wang, Z-A: Nonlinear stability of traveling waves to a hyperbolic-parabolic system modeling chemotaxis. SIAM J. Appl. Math.70(5), 1522-1541 (2009)

9. Li, T, Wang, Z-A: Asymptotic nonlinear stability of traveling waves to conservation laws arising from chemotaxis. J. Differ. Equ.250(3), 1310-1333 (2009)

10. Corrias, L, Perthame, B, Zaag, H: A chemotaxis model motivated by angiogenesis. C. R. Math. Acad. Sci. Paris336(2), 141-146 (2003)

11. Corrias, L, Perthame, B, Zaag, H: Global solutions of some chemotaxis and angiogenesis systems in high space dimensions. Milan J. Math.72, 1-28 (2004)

12. Hao, C-C: Global well-posedness for a multidimensional chemotaxis model in critical Besov spaces. Z. Angew. Math. Phys.63(5), 825-834 (2012)

13. Danchin, R: Global existence in critical spaces for compressible Navier-Stokes equations. Invent. Math.141(3), 579-614 (2000)

14. Danchin, R: Fourier Analysis Methods for PDE’s (2005); http://perso-math.univ-mlv.fr/users/danchin.raphael/ courschine.pdf

15. Ahmad, S, Le, D: Global and blow up solutions to cross diffusion systems. Adv. Nonlinear Anal.4(3), 209-219 (2015) 16. Li, H-L, Matsumura, A, Zhang, G-J: Optimal decay rate of the compressible Navier-Stokes-Poisson system inR3. Arch.

Ration. Mech. Anal.196(2), 681-713 (2010)

17. Matsumura, A: An energy method for the equations of motion of compressible viscous and heat-conductive fluids. MRC Technical Summary Report, Univ. of Wisconsin-Madison,2194(1981)

18. Sohinger, V, Strain, R-M: The Boltzmann equation, Besov spaces, and optimal time decay rates inRn. Adv. Math.261, 274-332 (2014)

19. Yu, H-J: Large time behavior of solutions for hyperbolic balance laws. J. Differ. Equ.261(9), 4789-4824 (2016) 20. Majda, A: Compressible Fluid Flow and Systems of Conservation Laws in Several Space Variables. Applied

Mathematical Sciences, vol. 53. Springer, New York (1984)

References

Related documents