Determinants and Their Applications in Mathematical Physics

(Chris Devlin) #1
A.10 Formulas Related to the Function (x+


1+x^2 )

2 n
333

Proof. Replacexby−x
− 1
in (A.10.1), multiply byx
2 n
, and putx
2
=z.


The result is


(−1+


1+z)

2 n
=

[

z

n
+

n

i=1

λniz

n−i

]

−(1 +z)

1
2

n

i=1

μniz

n−i

=

n+1

j=1

λn,n−j+1z

j− 1
−(1 +z)

1
2

n

j=1

μn,n−j+1z

j− 1
.

Rearrange, multiply by (1 +z)


− 1 / 2
and apply (A.10.4):



i=0

νiz

i

n+1

j=1

λn,n−j+1z

j− 1
=

n

j=1

μn,n−j+1z

j− 1
+(1+z)

− 1 / 2
(−1+


1+z)

2 n
.

In some detail,


(1 +ν 1 z+ν 2 z


2
+···)(λnn+λn,n− 1 z+···+λn 1 z

n− 1
+λn 0 z

n
)

=(μnn+μn,n− 1 z+···+μn 1 z

n− 1
)+

(

z

2

) 2 n

(1 +z)

− 1 / 2

(

1 −

1

4

z+···

) 2 n

.

Note that there are no terms containingz
n
,z
n+1
,...,z
2 n− 1
on the right-


hand side and that the coefficient ofz
2 n
is 2
− 2 n


. Hence, equating coefficients


ofz
n−1+i
,1≤i≤n+1,


n+1

j=1

λn,j− 1 νi+j− 2 =

{

0 , 1 ≤i≤n

2
− 2 n
,i=n+1.

The theorem appears when n is replaced by (n−1) and is ap-


plied in Section 4.11.3 in connection with a determinant with binomial


elements. 


It is convenient to redefine the functionsλnrandμnrfor an application

in Section 4.13.1, which, in turn, is applied in Section 6.10.5 on the Einstein


and Ernst equations.


Ifnis a positive integer,

(x+


1+x
2
)

2 n
=gn+hn


1+x
2
, (A.10.6)

where


gn=

n

r=0

λnr(2x)

2 r
, an even function,

g 0 =1;

hn(x)=

n

r=1

μnr(2x)

2 r− 1
, an odd function,

h 0 =0. (A.10.7)
Free download pdf