Determinants and Their Applications in Mathematical Physics

(Chris Devlin) #1

38 3. Intermediate Determinant Theory


Hence, removing the factorAnfrom each row,


|cij|n=A

n
n





δijxi+

Hij

An





n

which yields the stated result.


This theorem is applied in Section 6.7.4 on the K dV equation. 

3.6 The Jacobi Identity and Variants


3.6.1 The Jacobi Identity — 1


Given an arbitrary determinantA=|aij|n, the rejecter minorMp 1 p 2 ...pr;q 1 q 2 ...qr


of order (n−r) and the retainer minorNp 1 p 2 ...pr;q 1 q 2 ...qr of orderr are


defined in Section 3.2.1.


Define the retainer minorJof orderras follows:

J=Jp
1 p 2 ...pr;q 1 q 2 ...qr
= adjNp
1 p 2 ...pr;q 1 q 2 ...qr

=

∣ ∣ ∣ ∣ ∣ ∣ ∣

Ap 1 q 1 Ap 2 q 1 ··· Aprq 1

Ap 1 q 2 Ap 2 q 2 ··· Aprq 2

.........................

Ap
1 qr
Ap
2 qr
··· Ap
rqr

∣ ∣ ∣ ∣ ∣ ∣ ∣ r

. (3.6.1)

Jis a minor of adjA. For example,


J 23 , 24 = adjN 23 , 24

= adj





a 22 a 24

a 32 a 34





=





A 22 A 32

A 24 A 34





.

The Jacobi identity on the minors of adjAis given by the following theorem:


Theorem.


Jp
1 p 2 ...pr;q 1 q 2 ...qr

=A

r− 1
Mp
1 p 2 ...pr;q 1 q 2 ...qr
, 1 ≤r≤n− 1.

Referring to the section on the cofactors of a zero determinant in Section

2.3.7, it is seen that ifA=0,r>1, thenJ= 0. The right-hand side of the


above identity is also zero. Hence, in this particular case, the theorem is


valid but trivial. Whenr= 1, the theorem degenerates into the definition


ofAp 1 q 1 and is again trivial. It therefore remains to prove the theorem


whenA=0,r>1.


The proof proceeds in two stages. In the first stage, the theorem is proved

in the particular case in which


ps=qs=s, 1 ≤s≤r.
Free download pdf