Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
A KEONECKER KLTA.. CU) JOHNS HOPKINS UMY LAUREL NDFPPIED PHYSICS LS It A FARRELL ET AL. SEP 85
UNLSSIFIE JN /APL I MM 4-8 --30SiF/ 12/ NL
I.E
--.-- vr ~-.~-- . z - --.- .
'4_
r2.
4,0- 1 2.0
1.11.8
1111.25 [41116
MtCROCOPE 'SOLuTiON TEST CH4PT
f.4 I-. 84 A F ,TAAP S - 963
I
I
I
I
- .... - . .. . - - . -- . , " .
JHU APL
TG 1350
SEPTEMBER 1985
I,.?co
DEMONSTRATION OF AJACOBI POLYNOMIAL REPRESENTATIONOF A KRONECKER DELTA FUNCTION
RICHARD A. FARRELL
ERNEST P. GRAY
ROBERT W. HART
10 22 042
UNCLASSIFIED PLEASE FOLD BACK IF NOT NEEDEDSECURIPY CLASSIFICATION OF THIS PAGE FOR BIBLIOGRAPHIC PURPOSES
REPORT DOCUMENTATION PAGEla REPOP" SECURITY CLASSIFICATION I lb RESTRICTIVE MARKINGS
L nla'.ified
2a SECURITY CLASSIFICATION AUTHORITY 3 DISTRIBUTION'AVAILABILITY OF REPORT
Approved for public release; distribution is unlimited.
2t DECQASSI'-CAT;-N DOWNGRADING SCHEDULE
4-PERFORMINO ORGANIZA'ION NUMBERS S MONITORING ORGAN4IZATION REPORT NUMBERS,.
jilt API I(I 135 iHl' API 163 t350
6a NAME OF PArORMING OPGANIZA71ON Et OFICr SYMBOL 7a NAME OF MONITORING ORGANIZATIONIt AopOCab'e
[ti Ioh1, -li'pki', L 0.Q.'r, .
Na\al Plant Representatis e Office\p;':, . t11 I!'. I . h, , r ,r RI P
r, A_:RI S a . 3 ale 4W P ,7b ADDRESS ICily. Slate. and ZIP Code
I him. Hopkit, Road Johns Hopkins Road
I turcl. \lrland 21l1-- 1 aurel, Maryland 20707
'AYT K*. N."G SONSO;RINi, O"RGANI,,AION T (OIFICE SYMBOL 9 PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERIIt Aophcaitle l
Na a N'c 'bItt ( ,I1and i N00024-85-C-5301
5. A S 7 Slae ad Z, Code 10 SOURCE OF FUNDING NUMBERS
PROGRAM PROJECT TASK WORK UNIT*20,12 ELEMENT NO NO NO 3J ACCESSION NO
T s" ' ,, secure. Classfcat,
Dcrnon, rlil M1 a .J acobl Pol nomal Reprc'entation of a Kroneck er Delta 1'unction (U)
12 PERSONAL AU'14OR,S,
IaTfcIl. Ri hard A, (ira,. [ri.ost P.; Hart, Robert %,
*,a '"A1 2' Pf";n " 3b TIME COVERED 14 DATE OF REPORT (Year Month1 Day 5 , PAGE COUNT
I .lilia \lMtorandum I k iif 1985. September 14
t* ,ATI CFODES 18 SUBJECT TERMS
S,'.- P SUB GROUP i Kronecker delta function
Jacobi polynomials. .. . . .... . fluid shells -
S- 1-- ' -cPssa,, and dertfy by block number,
\ ' n , ! .' Nholc ha' brci gencraltied to obtain an expansion of the Kronecker delta function as an infinite series invols.ing the products
A", . a ,1 A, r, 21 ABSTRACT SECURITY CLASSIFICATION
. *. '-,AM[ AS RPT 0- DTIC USERS I I nclasified
A. ', ", . ', , . t, 22b TELEPHONE (include Area Code) 22c OFFICE SYMBOL
,' (30) 953-50)0(3 NPRO
M .t SECURITY CLASSIFICATION OF THIS PAGE
* -UNCLASSIFIED
* IJHU/APL
TG 1350
SEPTEMBER 1985
Tecbiical Memorandum
DEMONSTRATION OF AJACOBI POLYNOMIAL REPRESENTATIONOF A KRONECKER DELTA FUNCTION
RICHARD A. FARRELL
ERNEST P. GRAY
ROBERT W. HART
ii
- . THE JOHNS HOPKINS UNIVERSITY APPLIED PHYSICS LABORATORYJohns Hopkins Road, Laurel, Maryland 20707
0.,Na.
A,Approved for public release, distribution is unlimited
I "
APPLIED PHYSICS LABORATORY
ABSTrRACT
N\ formula dens ed bN Noble has been generalized to obtain an espansiollof the K roitecker delta f'unction as an inl'inite series insolsi ni the products~ ofmo5 la,:obi polx'noinlials.
T HE JOH~NS HOPKINS UNIVE 14SITYi
APPLIED PHYSICS LABORATORYAUALL MARYLAND
INTRODUCTION
In our ins estigations of the inviscid, incompressible flow of rotating fluidshells confined between concentric, spherical, rigid, co-rotating boundaries.
m~k sC cnCOUntercd the need for a proof of a Kronecker delta function representa-tion in terms of a series of products of two Jacobi polynomials. In particular,%c required a proof that
2"
kkhcrc
F, (A) - E - --- ---- _ __-
An(k+,A+l+ )(k-+p,+ I
X x~j '(A 1'1' (A) ,( 0/ 1 . (2)
For (%Aj )> -- 1, the P,'3 '(A) are the Jacobi polynomials which are orthogo-nal over the interval (-1, + 1) with the integration weight factor w M)(I -- x) (l + x)". They are normalized by the relation Pn " )(I) = i(a, +n + )
*[F'(n i- i)F(c± + )], w ith F(z) the gamma (factorial) function. For (a,,(3) !s - 1,the Jacobi polynomials' P,('' 3)(A) are defined in terms of the polynomials for((x,0) > -I through repeated applications of the contiguous relations2
(2n + o t 1) P,, " (A) = (n + o(+ 3) P,,"' (A) - (n + 3) P,('-', (A) (3a)
K and
(2n - (k+ 3) 1),," ' (A) = (n + u + i) P,4'H + (it + i)P'(A ,) (3b)
with P4'-" (Al I and P-1'((A) =0.
S Certain sums involving products of two Jacobi polynomials have been previ-ously evaluated (see, for example, Refs. 3 - 6). Our method for evaluating thesumn in Eq. 2 employs mathematical procedures analogous to those whichNoble' used to show that 6
*~ ~ ". M1ag[Ws, F. Oberhettinger. and R. P. Soni. Fojrmulas anfd rheoretsOfor the Special Fun(-,tions of Aathernatical Physics. Springer-Verlag. New York (1966).1. S. (iradshtevn and 1. M. Ry/hik, Tables of Integrals, Series, and Products, Alan Jeffrey,led , Academnic Press, Ness York (1965).11. [. %ianocha and B. I . Sharma. "Some Formulae for Jacobi PQlynornials," in Proc. cami,.Phil. So(., pp. 459-462 (1966).1I, 1 - Manocha and B. I . Sharnma. Geinerating Functions of Jacobi Polynomials," in Proc,('utnh. P'hil. Soc., pp. 431-433 (1966).
* II1. Noble, "'Some Dual Series Equations Ins ols'ing Jacobi Poly nornials.'' in Proc. ('arth. Phil.So(.. pp. 363-3*7I (1963).V.. R. Ian~en, A1 Table of Series and Products, F'rent ice- Hall. Ne%\ York I 1975).
I HE JOHNS HOPKINS UNIVERSITY
APPLIED PHYSICS LABORATORYLAUREL, MARYLAND
2''1(1 V) a(l +x) (x--y) +h -- IA(xy)
(a+ b-c) kPF(k+c+l)(2k+c+I) p( X ) P-,,. a (y) (4)-- ) r(k+a+l)r(k+b+l)
where
= I if -1 !; y < X ! 1
0 if I < x _< Y I
PROOF
The proof is done in two stages: first the I = 0 case is treated, and thenthe I e 0 case is analyzed. Both stages use the expansion of a function of twovariables in an infinite series of Jacobi polynomials of one variable and coeffi-cients that depend on the other variable. In particular, the coefficients in theexpansion are obtained by respresenting the Jacobi polynomial as a finite se-ries, evaluating the resulting elementary integrals, and re-summing the resul-tant series.
In proving Eq. 1, it is convenient to use the relationship
-- I +X) . (x) for n > 1 (6)
in order to write Eq. 2 as
*IF,..,._ - p (,-,,:. i (A)(2,u+ l)
(I+A) 2 (2k+,i+I+V)2 A) P(M A',(7)+ _ _ P A( "I / ':'* ( A ) -- k A _+ 1- ' ( A ) , (7 )8 k-I k(k+1)
where the first term is the k=0 term of the sum appearing in Eq. 2.
02
• .
THE JOHNS HOPKINS UNIVERSITY
APPLIED PHYSICS LABORATORYLAUEL, MARYL ANO
CASE OF 1 0For I = 0, we expand
(I+x)(I+Y) -(
= . C"(")(y) P, ( ), (8a)k=O
where
h xif y > xy if y < x. (8b)
Multiplying Eq. 8a by (1 +x)(i -x)') P,,' ' (x) and integrating from xI -to I yields
(m + 1)2p 5 /2C,"(
0 (y) 2 512
(m+u+3/2)(2m+p+5/2) -1+y)
x dx~l X) 'A+ : P (11+:' ( ) d ( - ) (p + 3/2)
+[ d(l AMp4~ W() [ t~ -rI-) ~ dl+ v(I -- v| P" 0
+~ ~ ~ ~ -dxx(x)" P,,),+ ' 2 ).)( -)(x) :dI
20(l+y) (+ V- -4 dx(I-x) - g + 'l"(x)
The integrals are readily evaluated by expanding the Jacobi polynomial as'
= -y (m,ao, 3;r)( ~x)r (10a)r =-
3
................................. (-)(') -)+:p...... .
THE JOHNS HOPKINS UNIVERSITY
APPLIED PHYSICS LABORATORYLAUREL. MARYLAND
with
r y(m,a,f3;r) r(af+m+lnr(af+(+m+r+I)(-1/)r/
integrating each term; and re-summing the series. In particular,
(m + 1)2"+5/2 -C,/()=(yV(+ '
(mn+ A+3/2)(2m+,u+5/2)
M -2 r4 -7 /2 2 1 +S/
2+I r+
.y~~p '',lr) (r+Mu+ 3 / 2 ) (r+ 1)
x-Y L+ I IM~ yr~~T
(r+ 1 (k+3/2))r=
x [ 2 r+PA+7/2 2-SVA(+ -Y) r+]
(r+1)(r+,u+3/2)
=-2A+/ (+y) (m + I+3/2)2 P M(++I h(l (y]
___________ _ (11)
where the fact that
* y(m + l,pA+ /,- l;r +1)=
-(m + j± + 3/2)2,y (mM + 11,l;r)/[2(r+)A+ 3/2)(r + 1))
is used to arrive at the next to the last line, and Eq. 6 is used in the final step.Inserting this result in Eq. 8a yields
4
........................................
THE JOHNS HOPKINS UNIVERSITY
* APPLIED PHYSICS LABORATORYLAUREL. MARYLAND
2P ~~ (-w) -A")-H(0 ) (xy)=
(+y)(1+X) (+ 1)
(2k ++ 5 /2) P"P+ (V 4A ()(I2a)k=0 (k +1)2
or
2 'A +
\2A l 1-w/ (2 1A+l) 8
(2k +,u+ '2 ."y) (I2b)
k2 Pk,( Y) P_
with w defined by Eq. 8b. Since Po(")~ (A) = 1, comparing Eq. 7 with Io to Eq. 12b with x=y=A yields
_ ( )'+ ) (13)(2)A+1) ]-
which demonstrates that Eq. I holds for 1 0.
CAEFor 1 0, we consider
0for I : y 2) x-2t
H"' (x~v)r~p-+5/2) (I -+x y) - (A +3/2) (x-_y)]
fo f an ~x >y 1 (14)
Application of Leibnitz's rule for differentiating a product yields
I-I Y 1+1
W) (x,y) Es -(x-y)(15
5
THE JOHNS HOPKINS UNIVERSITY
APPLIED PHYSICS LABORATORYLAUREL. MARYLAND
with +1 = 3/2)P(,u+ t+3/2)r (1)) -y)r(A-I+5/2)r(t+2)r(t+ l)rI-t)
and U(x,y) I if 1 > x > y -1, and U(xy) =0 otherwise. The func-tion HUl) (xy) is expanded in terms of Jacobi polynomials
H~(I (xy) = CA!~ (Y) Pk;' '~ x (16)k=O
with the coefficients determined from
(m + 1) 9 1 5/
= H(")(xy) (x)I (lxPI+ )()dx
1-1 m= S, r,y(m, 14-1+11,1;r) dx( i -xyu-+ 12+r (Xy)1+l 1 (17)i=0 =O
The integral is evaluated by changing the variable from x to z M (x-y)l/(I -y),which leads to
m
J11)2 ~,MI ,1;r)(1 _Y),,/+ V,+r+ 2 r(l + r+ 3/2)r=O
, r(,u-+r+t+7/2) (8
Substituting for S, the inner summation appearing in Eq. 18 reduces to
(,+/)-I+512 _ rs+31 2+t )]
r~,)= (Mu-+/2)=1);-'+1 (-2r)
(1) 1(19)
For r + r2, we note that
6
K THE JOHNS HOPKINS UNIVERSITYAPPLIED PHYSICS LABORATORY
LAUREL. MARYLAND
u+/2 a-2- + V Z I- I
(Ma/)(l,+ aI (l) 0 for r 0.ar, (±-I+ 5/2) z'"11(20)
The derivative is zero at Z I because (I - 2 - r) < I1- 1. The fact that thederivative reproduces the series in Eq. 19 can be readily verified by writingthe factor (I - )-1as a polynomial in z, differentiating each term, and thensetting z= 1. For I < r +2, we note that
ds s. I ~S1 . + IS
(l)'-'r(Mu+ 5/2)r(r+ 1)
The multiple integral in Eq. 21 is readily evaluated by recalling that integra-tion by parts leads to
ds1 ... (1 -se (sjf(s 1 )dsl. (22)0 ~(j-l1)!
b The fact that the multiple integral reproduces the series in Eq. 19 can bereadily verified by writing (I - sj) '- as a polynomial in sj and integratingeach term separately. Inserting Eqs. 20 and 21 into Eq. 18 yields
()r ,+5/2) r=1
I(m,1) = r(1-1+r+312)r! for M - -r (MA+52+r) (r+ 1-1)!
0 otherwise. (23)
* .Form 2tI 1, using the definition of y(m,,u-4 I ,l2 ;r) and introducingthe indices r' -r--(I - 1) and m' =m-(I - 1) leads to
7
THE JOHNS HOPKINS UNIVERSITY
* APPLIED PHYSICS LABORATORYLAUREL. MARYLAND
III') - __________
- min + A + /) r(4+ 5/2 -1
[(y- 1 )/2y r(in' +.u+l+ +r')
where the definition of -y(m-l+ 1, At+1+ V2, - l;r) was employed to obtainthe last line. Inserting Eqs. 23 and 24 into Eqs. 17 and 16 yields
- GA+ 3 12 'r(/L+5/2) (2m+ji-1+5/2)
~j - r(I+5,2-/) m=I-1 (m+l)(m+±+312)
X P,( J('/2 1-(y) P,,(A-1+'4'I)(X) (25)
Changing the summation index to m' = m - (I - 1), writing the m' =0 termseparately, and using Eq. 6 in both the mn' = 0 and the mn' d0 terms lead to
H(1) (x.y) ( P~+/)[(+.I() + (~)Iy)(l+x)r(,u+5/2 -1) L(2p+ 1) 8
(2XfM y (26)M, ' (in' + 1) rnIpM,+'.)xj
Setting x=y= A in Eq. 26 and comparing with Eq. 7, one finds
H (A1A 2 -)r ( +5/2) )F..f 0, 1 60 , (27)
where the vanishing of W)' (AA) follows directly from the form of Eq. 15.This completes the proof of Eq. 1.
8
THE JOHNS HOPKINS UNIVERSITY
APPLIED PHYSICS LABORATORYLAUREL.MARYLAND
dl REFERENCES
'W. Miagnus, F. Oberhettinger, and R. P. Soni, Formulas and Theoremnsforthe Special Functions of Mathematical Ph sics, Springer-Verlag, New York(1966).
21. S. Gradshteyn and 1. M. Ryzhik, Tables of Integrals, Series, and Products,Alan Jeffrey, ed., Academic Press, New York (1965).'H. L. Manocha and B. L. Sharma, "Some Formulae for Jacobi Polvnomi-als," in Proc. Camb. Phil. Soc., pp. 459-462 (1966).
4 H. L. Manocha and B. L. Sharma, "Generating Functions of Jacobi Poly-nomials," in Proc. Camb. Phil. Soc., pp. 43 1-433 (1966).
5 5B. Noble, "Some Dual Series Equations Involving Jacobi Polynomials, "inProc. Camb. Phil. Soc., pp. 363-371 (1963).
6 E. R. Hansen, A Table of Series and Products, Prentice-Hall, New York(1975).
9
TH1E iOHNS H~OPKINS UNIVERSITY
APPLIED PHYSICS LABORATORYL AL I MAR,rIANC
INITIAL DISTRIBUTION EXTERNAL TO THE APPLIED PHYSICS LABORATORY*tic sLrk repocd III I(' 0Sii'.I'doneuniirdc! N"'\ ( ,.ilrar~ N(x)i'i4 S ( 10I Ih '-1 l111d 11 1.4~ \", I'0111 id h' I I I
I oiniiind
ORGANIZATION LOCATION ATTENTION Copief
IM P ARI N T 0I ) I M \,\\I
W)1 PR AIl )NI ]I N11 41
()\I IRM I ORd XI
ENR'd "1 k' va~ 0 Pailouill", AX I kIL 1 e
4 1 \[\I RNI III
H ' lr ' I l. wol '' ~ r, 141111o. \11)k NOi
ti~~~I Il ii ii' Ii'ni i
R I I on'!r
* Requests for copies of this report from 000 activities and contractors should be directed to OTIC,Cameron Station, Alexandria. Virginia 22314 using OTIC Form 1 and, if necessary. DTIC Form 55.
'initial distribution of this document within the Applied Physics Laboratory has been made in accordance with aslist on tite in the APL Technical Publicatirns Group
FILMED
12-85
* DTIC