" On the Calculus of Symbols.-Fifth Memoir .
With Application to Linear Partial Differential Equatiols , and the Calculus of Funetions .
" By W. HI .
L. RUSSELL , Esq. , A.B. Communicated by Professor STOK ES , Sec. R.S. Received April 7 , 1864* .
In applying the calculus of symbols to partial differential equiations , we find an extensive class with coefficients involving the independent variables which may in fact , like differeiltial equiations with constant coefficients , be solved by the rules which apply to ordinary algebraical equations ; for there are certain functions of the symbols of partial differentiation which combine " with certain functions of the independent variables according to the laws of combination of commoni algebraical quantities .
In the first part of this memoir I have investigated the nature of these symbols , and applied them to the solution of partial differential equations .
In the second part I have applied the calculus of symbols to the solution of functional equations .
For this purpose I have given some cases of symbolical division on a modified type , so that the symbols may embrace a greater range .
I have then sliown how certain functional equations may be expressed in a symbolical form , and have solved them by methods analogous to those already explained .
Since ( xd -y )(X2+y2)=O .
we shall have ( d dyJd)(X2 ?
Y2)U(X2 ?
y2)( +-y+)u , or , omitting the subject , also ~~~d d , 2d d ( 2 d __yd ) ( Xto + y2 ) =(X2 y2 ) ( X d__yd -+.+y=+va--y dx also d _yd + S2+2 , , 2+$d _yd ; die -ydx ~d dar dd22 therefore the symbols -y -Yd and x2y2 combine according to the laws of ordinary algebraical symbols , and consequently partial differential equations , which can be put in a form involving these functions exclusively , can be solved like algebraical equations .
We shall give some instances of this .
Consider , first , the equation d2du _d_ d'u du du Pd -$ dx die 2dx -Y S ( x +Y ?
u )=f(x Y ) .
This may be put in the form { dyYd{(2 +Y2 ) } { $jY-yd-+( ( 2+Y2)}u=f( , ) ; orifw={ dy d 2+ 2 ) d we shall have or~~~ if wy X.d d_2H+ Y2 ) -$y dy da Lagrange 's metliod will give in this case the equations dy _dz _ au y f(X , y)+(X2+y2)w Hence we shall have dH _4y2 = el , 2X 2 ) Let Vd)f(x '-')_ ceV _\n c x , -(a,2+ e ' ) sin--I ___ 2W+ey2 ) sin+ A_ ( S+82i-T Vi 2+y2 +y ' ) where cp is an arbitrary function .
'We shall denote this expression by X(x ' y ) , whence we have for the determination of ( u ) , du du 2 ( S+2)U X(X4 y ) dy dx which gilves ( X2+r2 ) Sin-I -+2V2 ' '+ ) sin- , x(ue 4V2+y2 F31(X yS+2 ) + , e 14_2+y2oj ( X2+y2 ) .
where F( &lt; which completes the solution .
Next let us take the equation 2d'u 2 d2u +2d 22 du dx2 $ dx dy dy dy du + ( XI+ y2)Iu f( , .
This equation may be written ( dd u2( 2+y2 ) ( -d + ) U+(a2+ y2)2 =f(X , y ) , { d _ ?
y2 ) } { _y d(x2+y2)}uf( , y ) which may be , treated as before .
In order to find the most general form of equation to which the symbols dd2l x ; -y and 2Xy2 give rise , we must determine the expansion of Ic dd\ dd ( w d-YT d ).As x+Ydand x2,2 likewise combine according to the laws of algebraical symbols , we shall take ( X-6 ?
Y - ) to avoid the negative sign .
Now the expansion of ( d Y ?
) will consist of all the terms of the form *X x-y d,..J , vdY IS in which a+bd Jn .
We shall write &amp; .
for ad , and By for d , where it is to be understood that 4 and By do not apply to the subject .
Moreover we shall use , as in the third memoir , a-I a-2 a-r ?
1 a , fora .
23 Then we shall have , if a+a= r , ( \d ) ( c ; .1 dO 1d die -2cb die ) V TX dfya dx &amp; Sa Jy a-l dxa ( / j o-2 cla + a3 xa Y/ 7ya -3C-X-a +** dga3dXa Again , a+oa+ a bn , / db il ada 62\ wx-- ) i &amp; y Y =ga2rb Xi d_ _ ? ?
4yb + ?
ya ?
+xb B2 d2 In-2 , jJ dlla d+b + b ?
Ja d-a2b-1 die cb/ L+lb2 + 63 ya X3a _. .
, cly a clx a +4-3 +c~.y b~ 32 d n-2 _________.+ax dy-2da+ + ; bz by 232~ ya 2 +And again , if a+M+b ?
I3=n , we shall have+ &amp; c ( $ d ) ( dXb($ d-2)o(Ydx ) ya+bxa+13 ?
~iy ~ a+b +a+ dXa+b +b--Xa+b-I + C2dYxn-2y( n2 +6a3y2 ?
bx1 ax d3 z-3n + 2dl +X2 dyoc+1-3dx.a + b3 + as1bPy3yy dfli a b3 aYY yy dn + +a7Y6Z2aB23ya joc+2-d-i+ba+6f2l ?
Cib1x13yb3x stz , / s2 R+13-I r~a+ 3 +3y c-3 dTe a+bd An agin if+J a+ d'e a , +b a +36l 6y + nn weyshl have gw+ ddn-d2b d d_ _+d =ya+ bG Xa u(t+13 d3yyb +by a 1 a7b1 W/ tb dn +I 3-2d a+b-i die ( x2 die +ya -y 'd xb + ?
+b 2 +b , y( , 32 d. dya+ d , V+1-3-d+ +I3~1a~~1335~ , a d'z-2 I_b_x________2.IcPIx+~23x+3sb2Sa1 12 dya+-1 dv+-3d a+6 d|+if ?
U-2+ +61132 d Xb dn-3 + Xxp2Yab dya+P-2 ClXa+b _XM22Ya+ dya ?
+P-2d8(Xa+6bI + aL02 32Y ?
yb a.ya _ -3 ?
a+b We are consequently able to see that the general term of iS ~ Cqi6pi**I3qO1p** TIE WG TIE %j dx~o ~d s ; r ** cq , bp , q* *13 , ] ocP YY*x 4a+p+.-2dXyb +3+.a a+b+..-s where P+q +** =r , ******* * We may consider B=1 , which gives the following conditions : a'b-ab ' , a=-b .
Also A=CE=- , F=- .
abaa And .
the symbols may be written b(ax+ by + c ) d -axa+ by + cld and a2 &amp; 2+ 2ab xy + b2y2+ 2ac'x + 2cby + H. It follows hence that in order to find the form of the differential equations to which these symbols give rise , we must know the expansion of ( x d+Y di ) , where X and Y are functions of x and y. The expanded form will be a series of terms like ( die ( d ) bid ) ( X )a We must consequently find an expression for ( X in powers of d. It must be remembered that X is a function of x and y , in which ( y ) during the present process is considered as constant , and therefore X may be looked upon as a function of ( x ) only .
Now we shall find after a few differentiations , that ( x+7 ) =x5d- .
dx5 + ( XaX4 + X23X3 + X33X2 + X4aX ) d+ ( X3X3Xl + XcX23X2 + X2aX3X2 + XX3X + X23X2X + X3OX3X ) dW + ( XMXX8X2 + X3X6X23X +XX23X3X ) + X3X3X3X ; X d dxd Now let ( X d- ) =Xn de n d-n n dXns + ... + Then X(")-1nX ?
O'SX'SX ... where there are r 8 's , and + 3+ y + ... .
e=n .
Hence we shall have X d &gt; / ) TX ) Chi ) { _ x ?
xY(p ) dJ-l +Y(2 ) Jl3-2 7}{ Xb g ?
- ?
X+( ) Cl+ cX2)lb2 ?
&lt; Xl aJa I+Xa g*a-2 Whence it is obviouls that the gbeneral expression for the expansion of ( d d\fl ~ db dbI b(X-+Y- ) will depencl upon prinlciples noJt materially differing from dz dyl those already considered .
The syTmbols we have already considered are only of the first order of differentiation ; we shall show that there exist symbols of the secon}d order which combine with certain algebraical qnantities as if they were themselves algebraic .
Let us take the symbols a-+ 2b6 Jr cJ++ dxco dsc die dy2 dL v die and Ax2+ 2Bxy ?
y2+ 2A ' +2B'y+H .
Proceedint as before .
, we arrie at the following conditionos : A6+Bc0,..+..@..(1 ) Bb +Cc-=0 ... .(2 ) A'6+B'c=O0 ... ..(3 ) Aa ' +B6'=0 , ... ... ..(4 ) Baf + C6'= &gt; ... ... *.(a ) Ac + B6 =0..*(6 ) Ba +C d.(7 ) A'a +Bt'6= , ... .**..(8 ) 2tA+r46B +2cC +4ca'A'+46B'3+e0 ... ..(9 ) Whence we have , putting B= 1 , A=- &gt; C=-c ; d. , v die ~~ 6 ' and with the following conditions , 4a/ A/ ?r46'13 ' ?
e- ; the condition a beae may be otherwise writtene are oy in confsequence of the equation ace s2 .
It will be observed that several of the nine above equations are not independent of the rest ; so that the r sult is much simplified .
I now proceed to apply the calculus of symbols to the solution of functional equations .
Let Then the following formula are known : dd ( e{~(Z ) dv)Ax2fjX)-1(X , v+ 2)1 } These formulae .
may be thus expressed in the notation of the calculus of ' d symbols : if p=e ( ) e , I=x , 0a functional symbol acting onf(r ) in such a manner as to convertf(ir ) into fX-'(x ?
X I ) ; then pf(7r ) =`Ofr .
p , a genieral law of symbolical combination due to Professor Boole .
We will now consider two cases of internal and external division in which the symbols combine according to this law .
The results , as will appear afterward , will be found useful in the solution of functional equations .
And , first , for internal division .
We shall determine the condition that p#1(w ) +4k(r ) may divide pno , ( 7r ) + pnI q)-1(7r ) + .
The process will be , mutatis mutandis , the same as in my former memoir .
The symbolical quotient is n-lo2 Co +*n ; A7r 7 lr lb 7r r and the required condition is found by equating the symbolical final remainder to zero , and we have ~7r 'di 070~ I 007 Aot + +o0 _ +074 0 2L _ &amp; CG .
r owo ; ?X .
2T +1 -1 &amp; 1 kT __ Or T , 7r c. ? ?
4/ 7r0 ~17r ~b7 0 affecting every part of the term which succeeds it .
I shall now give the correspolnding condition for an external factor .
The symbolical quotient is pn2 n-(n-1 ) 4_ p2 { 2 ) nI2 o(21 0l4ir}2+ 3517r v,7r +17 'PI* The required condition is found by equating the final remainder to zero ; we have.0-1 IP 2T{ ~ 07 - &amp; C. + 0 , -t A ?
!r-1 '{070-1 +or 0-l .
Ao70 ' , &amp; 7r7r V^ 1+ 4ir01 4'oT Or +1# 0-1 in each term affecting everything which comes after it .
I conclude with some examples of ftunctional equations .
Let the functional equation be f(x ) af ---= ( x ) this may be written d AX)(Maed f((x)=F(X ) , or 1-1~ ddd ={l +edx-2+a2(edx2)2+ .
e.}F(x ) ___2XF ( x ) a ' ]+ 2zF +aF V 4+'* Vj1+2X2 Il+4X2 To make this solution complete , we must add a complementary function , and we have C C1 1~j+ ( / a ( -V/ a ) As an example of this , put F( ) w , and the series becomes aY G3 , i Cs32 +a ~~+ F-_ aF F+ V1+2'2 V1+42 V1+2'i2 -U2 i= dV de-V2(j +ae2X2V2 +a2 6-4$2V2+ OD =r Jo ae2 , V As a second example , we will take the equation f ( 4 2)3f G+ 2f(x ) =F ( x ) .
This equation may be written 2 . ?
d ( e X -3e Z +2)f(x)==F(x)l d. d ( e x-11)(e dx -2 )f F(x)Now let d ( d " 2)fxX(x ) , and the functional equation resolves itself into the two , f(jA-2j(x)=Xx( ) , and X knw fo)=rF(ms wh'ich are known forms .
As a last example , we will take the equation 13x-2 x2+3x-1 2x-1 fl F(x ) or putting _(X_1)2 de dx =p , X= , d d. x we can write the equation ( since p=e { p2_p = +r ( Xr+ 1 ) }uF(x ) Applying the method of divisors , we see that if the symbolical portion of the first member admit of an internal factor , it must be either p-7v or p(7+ 1 ) .
Now in this case Pf( ) =f ( 27r-)p .
Hence 27r-1 Of(T)=f ( __ .
Wherefore the equation por-+07r94 wO + Lo7rO &amp;c. 4y.7r 4i1w 07 .
0 becomes , if we take the factor p-xr , and put Or 1O= -o =-7r 4d17=z1 , 7'0-3r ( 7+ 1)( )-07r+07r=0 ; 7r-2 an identical equation if we put for the symbol 0 its equivalent as given above .
Hence p-7 is an internal factor of the symbolical portion of the first member .
Effecting the internal division , we have ( p-(7+ 1))(p--7r)f(x)F= ) .
Let ( p7rf(X ) = X(X ) , and the equation resolves itself into the two , ( p-(7+ 1))x( ) = F(x ) and ( p 7)J(X ) X %(X ) or ( 2-(x+ 1)X(x ) =FQ(x ) and f( 2x1 ) Al)=A(x ) ; 2 K forms which I have considered in my memoir on the Calculus of Functions published in the Philosophical Transactions for 1862 , in which the general solution of the equationis o(x)-x(x)o +{ 6 } -F(x ) , where 0 is the unknown function , has been obtained .

