2 1. Determinants, First Minors, and Cofactors
It follows from (i) and (ii) that
x 1 x 2 ···xn=
∑n
k 1 =
···
∑n
kn=
a 1 k 1 a 2 k 2 ···anknek 1 ek 2 ···ekn. (1.2.2)
When two or more of thek’s are equal,ek 1 ek 2 ···ekn= 0. When thek’s are
distinct, the productek 1 ek 2 ···ekncan be transformed into±e 1 e 2 ···enby
interchanging the dummy variableskrin a suitable manner. The sign of
each term is unique and is given by the formula
x 1 x 2 ···xn=
(n! terms)
∑
σna 1 k 1 a 2 k 2 ···ankn
e
1 e 2 ···en, (1.2.3)
where
σn= sgn
{
1234 ··· (n−1) n
k 1 k 2 k 3 k 4 ··· kn− 1 kn
}
(1.2.4)
and where the sum extends over alln! permutations of the numberskr,
1 ≤r ≤n. Notes on permutation symbols and their signs are given in
Appendix A.2.
The coefficient of e 1 e 2 ···en in (1.2.3) contains alln
2
elements aij,
1 ≤i, j ≤n, which can be displayed in a square array. The coefficient
is called a determinant of ordern.
Definition.
An=
∣ ∣ ∣ ∣ ∣ ∣ ∣
a 11 a 12 ··· a 1 n
a 21 a 22 ··· a 2 n
...................
an 1 an 2 ··· ann
∣ ∣ ∣ ∣ ∣ ∣ ∣ n
=
(n! terms)
∑
σna 1 k 1 a 2 k 2 ···ankn. (1.2.5)
The array can be abbreviated to |aij|n. The corresponding matrix is
denoted by [aij]n. Equation (1.2.3) now becomes
x 1 x 2 ···xn=|aij|ne 1 e 2 ···en. (1.2.6)
Exercise.If
(
12 ··· n
j 1 j 2 ··· jn
)
is a fixed permutation, show that
An=|aij|n=
n! terms
∑
k 1 ,...,kn
sgn
(
j 1 j 2 ··· jn
k 1 k 2 ··· kn
)
aj 1 k 1 aj 2 k 2 ···ajnkn
=
n! terms
∑
k 1 ,...,kn
sgn
(
j 1 j 2 ··· jn
k 1 k 2 ··· kn
)
ak 1 j 1 ak 2 j 2 ···aknjn.