2.3 Linear Algebra 73
2.3.4 Systems of Linear Equations...............................
A system ofmlinear equations withnunknowns can be written as
Ax=b,
whereAis anm×nmatrix called the coefficient matrix, andbis anm-dimensional
vector. Ifm=n, the system has a unique solution if and only if the coefficient matrixA
is invertible. IfAis not invertible, the system can have either infinitely many solutions
or none at all. If additionallyb=0, then the system does have infinitely many solutions
and the codimension of the space of solutions is equal to the rank ofA.
We illustrate this section with two problems that apparently have nothing to do with
the topic. The first was published inMathematics Gazette, Bucharest,byL.Pîr ̧san.
Example.Consider the matrices
A=
(
ab
cd
)
,B=
(
αβ
γδ
)
,C=
⎛
⎜
⎜
⎝
aα bβ aγ bγ
aβ bβ aδ bδ
cα dα cγ dγ
cβ dβ cδ d δ
⎞
⎟
⎟
⎠,
wherea, b, c, d, α, β, γ, δare real numbers. Prove that ifAandBare invertible, thenC
is invertible as well.
Solution.Let us consider the matrix equationAXB=D, where
X=
(
xz
yt
)
and D=
(
mn
pq
)
.
Solving it forXgivesX=A−^1 DB−^1 , and soXis uniquely determined byA,B, and
D. Multiplying out the matrices in this equation,
(
ab
cd
)(
xz
yt
)(
αβ
γδ
)
=
(
mn
pq
)
,
we obtain
(
aαx+bαy+aγ z+bγ t aβx+bβy+aδz+bδt
cαx+dαy+cγ z+dγt cβx+dβy+cδz+dδt
)
=
(
mn
pq
)
.
This is a system in the unknownsx, y, z, t:
aαx+bαy+aγ z+bγ t=m,
aβx+bβy+aδz+bδt=n,