• No results found

Characterization of the range of the propagation operator for the spherical reaction-diffusion equation

N/A
N/A
Protected

Academic year: 2021

Share "Characterization of the range of the propagation operator for the spherical reaction-diffusion equation"

Copied!
11
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Characterization of the range of the propagation operator for

the spherical reaction-diffusion equation

Citation for published version (APA):

Liu, G. Z. (1989). Characterization of the range of the propagation operator for the spherical reaction-diffusion equation. (RANA : reports on applied and numerical analysis; Vol. 8911). Technische Universiteit Eindhoven.

Document status and date: Published: 01/01/1989

Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne Take down policy

If you believe that this document breaches copyright please contact us at: openaccess@tue.nl

providing details and we will investigate your claim.

(2)

Eindhoven University of Technology

Department of Mathematics and Computing Science

RANA 89-11 June 1989

CHARACfERIZATION OF THE RANGE OF THE PROPAGATION

OPERATOR FORTHESPHERICAL

REACTION-DIFFUSION EQUATION

by

Liu Gui-Zhong

Reports on Applied and Numerical Analysis

Department of Mathematics and Computing Science Eindhoven University of Technology

P.O. Box 513 5600 MB Eindhoven The Netherlands

(3)

CHARACTERIZATION OF THE RANGE OF THE

PROPAGATION OPERATOR FOR THE SPHERICAL

REACTION-DIFFUSION EQUATION

Liu Gui-Zhong*

Department of Mathematics and Computing Science Eindhoven University of Technology

5600 MB Eindhoven •

The Netherlands

Abstract

The ranges of the propagation operator for the spherical diffusion equation are characterized as weighted Hilbert spaces of hannonic functions.

§

1. Introduction

We are concerned with the spherical reaction-diffusion equation

au

at

=liLB

u.

(1)

Hereu=u(~,t) (~E Sq-l ,t

>

0) is the unknown function andliLB is the Laplace-Beltrami opera-tor on the unit sphereSq-l inRq.

As is well known (cf. e.g., [5], [6] or [4]) we have an identity decomposition for the Hilbert space 00

L2(Sq-l)into spherical hannonics: L2(Sq-l)= El H~ whereH~ is the linear space of spherical

m=O

harmonics of degreem in IRq. For a sequence {A-mI meINo of positive numbers, then, we have a well defined positive self-adjoint operator A onL 2(Sq-l),Le.

(4)

-2-00

00

D(A)={uE L 2(Sq-l) I

L

A~IIPmuII2<00}.

m=Q

The domainD(A),equipped with the inner product 00

(U,V)A

=

L

A~(PmU, Pm v)

m=Q

and corresponding norm 00

lIuIIA

=(L

A~IIPmuII2)'h

m=Q

-is a Hilbert space by itself. If the sequence {Am} -is such thatRmIAm

=

0(1) for anyR

>

0, then each u in D(A) extends uniquely to a harmonic function on IRq (cf. [2] Lemma 2 and [4] Theorem IV. 1.10).

On the other hand, it is easy to see that the space

HAqij,L)

=

{U(X) harmonic on IRq 11Iulll1

=

{J

Iu(x) 121J.(lx I)dx}'h <oo} IR'

is a pre-Hilbert space with the inner product

(U,V)I1=

J

u(x)v(x)IJ.(lxl)dx. IRq

Here IJ.: (0,00)~(0,00)is a Lebesgue measurable weight function.

The following basic result in [3] is on the identification of a spaceHA qij,L) with a spaceD(A)for an appropriate pair of weight function IJ. and sequence {Ale}'

Theorem 1. Assume that a weight function IJ.:(0,00)~(0,00) and a positive sequence {Am}meINosatisfy the conditions below:

1)

J

r2m+q-1lJ.(r)dr

<

00, Vm E IN o

(5)

3

-3) {A~} -

J

r2m+q-lll(r)dr. Here and afterwards, for two sequences of positive numbers

o

{am} and {bm }we write {am} - {bm}iff 0

<

liminf(amb;l)::; limsup(amb;l)

<

00.

Then the spaceHAqCll) is isomorphic to the space D(A)as nonned spaces. The isomorphism is

exactly the restriction-extension mapping. Furthennore, if, instead of condition 3) above,

A~

=

J

r2m+q-lIl(r)dr,then the isomorphism is actually an isometry.

o

It is well known that (cf. [5], [6] or [4])

.1.LBu=-m(m+q-2)u, 'f;fUE HZt.

o

(2)

Thus the operator.1.LB , as defined on the algebraic direct sum of HZt(mEIN0), is essentially self-adjoint inL2(Sq-I).Its self-adjoint closure, still denoted.1.LB ,is given by

l!1LBU=-

L

m(m +q -2)PmU

m=O

UE D(l!1LB )= {u E L2(Sq-l) ,

:E

[m(m+q-2)f IIPm uII2<oo}.

m=O

The solutionU

=

u(~,t)of(1)for the initial valueu(O)

=

vEL2(Sq-l)is given by

U

=

ett.u

v

=

:E

e-

tm(m+q-2) Pm

v.

m=O

(3)

(4)

The range of the propagation operator at time t, R(/t.U

),Le., the possible states of the system at

timetstarting from the initial states inL 2(Sq-l),is given by

D(e -tAu)={u E L 2(Sq-l) I

~

e2tm(m+q-2) IIPm U112

<

oo}. m=O

(5)

If instead of Equation (1) we consider the more general fractional spherical reaction-diffusion equation

(1)'

(6)

-4-and

-(-liLB?/2U=-

1:

{m(m+q-2)} v12Pm u

m=O

00

U E D(-(-liLB?/2)

=

{U E L 2(Sq-l) I

L

{m(m+q-2)r IIPm uII2 <oo}

m=O

v/2 00 12

u

=

e

-I(-6u ) V

=

L

e-

l(m(m+q-2)jv PmV

m=O

v/2 00 12

D(e '(-6u ) )= {u E L2(Sq-1) I

L

e21 (m(m+q-2)jv IIPm ull 2 <oo}.

m=O

(3)'

(4)'

(5)'

The purpose of this paper is to characterizeD(e '(-6u)v/2) applying Theorem 1 above, more

pre-cisely. to identify it with a weighted Hilbert space of harmonic functions on JRq. Note that,

1(-6 )v/2

D(e LB ) equipped with the norm

IIUII

v,t

=

(1:

e

21lm(m+q-2)}v/2 IIPmU112)112 m=O

is just the space

X~

forX

=

L 2(Sq-1)andB

=

e

(_6u)v/2 ,inthe notation of[4]ChapterI.

(6)

§2. StatementofResults

Theorem 2. Given I

<

vS; 2 and t >O.The space

X~

=

D(e '(-6u)'·f2) is isomorphic to the space HA q(ll)as normed spaces under the restriction-extension mapping. Here 11:(0,00)~(0,00)is the weight function given by

2-v 11(')

=

,-2 Ilog , I2(v-1)

e

2(v-1) • 1 v-;::It-;:} v Ilog , I v-1 , ,

>

O.

(7)

Obviously conditions 1) and 2) in Theorem 1 hold true. The assertions in the present theorem are proved provided condition 3) in Theorem 1 is verified, which is done in the next lemma.

Lemma 3. For 1

<

v S; 2 and t

>

0 fixed and 11 described as in Theorem 2 above we have

j

,2m+q-l 11(') dr _ e2t[m(m+q-2)}v/2 .

o

(8)

(7)

5

-Inorder to prove this lemma we need the following delicate result of Brands ([ 1]) on the asymp-totic behaviours of integrals.

Theorem 4. Assume that a functionM /R+ ~ /R+satisfy the following conditions:

(i) ME C([O,oo,R), M E C2([a,00,R) for some a~ 0, M"(x)

>

0 (x~a) and M'(x) ~00 (x-+00).

(ii) There exists a positive function

a

on[b,oo)for someb~ 0 such that a(x) (M"(x»'h ~00 (x~oo);

a(x)=:;; x (x~b) ;

V't>03A >0V'x~A V'y~0 [Iy-x I=:;; a(x) ~ I C"(y)-C"(x) I =:;; tC"(x)].

Then the integral

+00

I(M,t)=

J

e,x-M(lxI)dx

- 0 0

has the following asymptotic behaviour ast ~00:

I(M,t)

=[

21t ]

~

eW(I)(1 +0(1»

m'(m~(t))

=[21t(m~)'(t)]~ eM%(I) (1+0(1»

I

wheremet)

=

M'(t),m~(t) is the inverse function ofmet)andMX(t)

=

J

m~('t)d"C.

o

Corollary 5. ForM : /R+~ /R+satisfying the conditions in the theorem above and

(9)

o

M(x)=~xa+l

+clogx forxsufficiently large(A >0, a>O, CE /R) (10) a+l

we have

(11)

The proofs are given in the next section. Here we remark that, in the Hilbert spaceHAq(11)for the

(8)

6

-§3. Proofs

Proof of Corollary 5. We need to know the asymptotic behaviourofm+-(t)ast ~00.We have

Since

y

=

m(x)

=

M'(x)

=

A xa

+

£ (x large).

x

m'(x) =M"(x)

=

Aaxa-1-

--%-

>

0 forx large enough

x

(12)

the functionm(x) is strictly increasing in a neighbourhood of00. Therefore the inverse function

m+-(y) exists in a neighbourhood of00. SUbstituting

+

for

y

and 1

1

forx in (12) we

-

-a -a+1 a a+1 ro

have

F(a,ro)= roa - A - c aroa+1 = O. (13)

ClearlyF(O,AVa)

=

0 and F(a,ro) is analytic in (a,ro)in the neighbourhood of(O,AVa). Accord-ing to Weierstrass' theorem we have the followAccord-ing expansion in a neighbourhood of

a

=

0:

Wz

A-lIaro=l+w1a+-~+ ... 2

So is true the expansion

1

-.,...---=l-W10+ ...

A-Varo

(14)

(15)

Inserting (14) in (13) and equating the coefficient ofaon both sides we obtainW 1

=

c AVa •This

a

together with (15) implies that 1 X

=--1--o

a+1 ro A-Va

=

- - 1 - (1-W1a+ ... ) a a+1 (16)

where c' is a constant. The series converges unifonnly in a neighbourhood ofy

=

00, so does its

(9)

7 -t Jm<-('t)d't eM'(t) _ ea ~A-lIatI+lIa C - e I+a - - log t.

a

Relation (11) then follows from relations (9), (17) and (I8).

Proof ofLemma3. For1

<

v::;; 2we have

v v~

e

2t(m(m+q-2)}v/2 _

e

21m [1+2" m I

ForM(x)in Corollary5we have

l(M,2m+q-2)

.!.(_I_I)-..£.. ~rlla(2m)I+lIa[1+(l+lIa)~l

_ (2m +q_2)2 a a eI+a 2m (a~ 1).

Consequently, if we could take a,eandAsuch that

l I e -(--1)--=0 2 0 : 0 : 1 1+-=v 0: I+l... _0:_A-lla2 a =2t 1+0: then l(M, 2m +q-2) _

e

2t (m(m+q-2)}v/2.

It turns out that the system of equations (62) indeed has a set of solutions:

(17) (18)

o

(19) (20) (21)

(10)

8

-1 a =

-v-I

v-2 c = -2(v-l) 2 A

=

---:-1-(tV)v-l

On the other hand

00

J

r2m +q-1 Jj.(r)dr

=

I

e(2m+q-Z)xll(eX)e2l:dx.

o

- 0 0

(22)

(23)

Relation (8) then follows from the relations (21), (22), (23), (7) and the fact that for any continu-ous function Il on(0,00)and 0

<

1 there exists a constant CIl such that

1+01

J

r2m +q-l Jj.(r)dr I

Hi

S CIl 1 [(l+O-I)2m+Q -(l-o)2m+q]

2m+q

=

o(e2Jm(m+q-Z)}v/2) (m

~oo).

Thus Lemma 3 is proved, so is Theorem 2.

Finally we remark that Theorem 1 is still valid for v

>

2incaseq

=

2.

(11)

-9-Acknowledgement:

The author would like to express his gratitude to Prof. J. de Graaf for his inspiring advices.

References

[1] Brands, I.I.A.M., Asymptotics of a certain integral, to appear in Applied Mathematics

Letters.

[2] Graaf, J. de, Two spaces of generalized functions based on harmonic polynomials, in C.

Berzinski, et al. (ed.), Polynomes Orthogonaux et Applications, Proc. Bar-Ie-Due 1984, Lecture Notes in Math. 1171, Springer, Berlin etc., 1984.

[3] Liu, G., Hilbert spaces of harmonic functions in which differentiation operators are continu-ous or compact, preprint 1989, Eindhoven University of Technology, Eindhoven.

[4] Liu,

G.,

Evolution Equations and Scales of Banach Spaces, Ph.D. thesis, June 1989,

Ein-dhoven University of Technology, EinEin-dhoven.

[5] Miiller, C., Spherical Hannonics, Lecture Notes in Math. 17, Springer, Berlin etc., 1966. [6] Stein, E.M., and Weiss G., Introduction to Fourier Analysis on Euclidean Spaces, Princeton

Univ. Press, Princeton, New Jersey, 1971.

Key Words: Spherical diffusion equation; Hilbert spaces of hannonic functions; Spherical har-monics; Asymptotic expansions;

Referenties

GERELATEERDE DOCUMENTEN

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of

Aansluitend op het onderzoek in fase 1 van de verkaveling werd in fase 3 een verkennend onderzoek met proefsleuven uitgevoerd; dit onderzoek bevestigde de aanwezigheid van

presenteerde reeds in 1953 een dispersieformule voor lucht op basis van metingen gedaan door Barrell en Sears in 1939 voor het NFL. Metingen uitgevoerd na 1953 wezen voort- durend

Figure 4.24: Effect of fertilizer type, NPK ratio and treatment application level on soil Na at Rogland in

Indien ook een contrastmiddel in een (arm)ader is toegediend, is het aan te raden om na het onderzoek extra veel te drinken.. Hierdoor wordt het contrastmiddel makkelijker en

Om de stroomgebiedbeheersplannen bilateraal tussen Nederland en Duitsland op elkaar af te stemmen (een KRW-verplichting), werd rond 2002 een ‘Steuerungsgruppe’ opgericht. Deze

The wave number corresponding to a value on the dashed line of Figure 3(a) and beyond which the pattern is unstable, corresponds to the critical current: beyond the critical

The diffusive motion of a colloidal particle trapped inside a small cavity filled with fluid is reduced by hydrodynamic interactions with the confining walls. In this work, we