The matrix exponential is a powerful means for representing the
solution to nn linear, constant
coefficient, differential equations. The initial value problem
for such a system may be written
x
′t=Axt
x
t
A
x
t
x0=x0
x
0
x0
where AA is the n-by-n matrix of
coefficients. By analogy to the 1-by-1 case we might expect
xt=ⅇAtu
x
t
A
t
u
(1)
to hold. Our expectations are granted if we properly define
ⅇAt
A
t
. Do you see why simply exponentiating each element of
At will no suffice?
There are at least 4 distinct (but of course equivalent)
approaches to properly defining
ⅇAt
A
t
. The first two are natural analogs of the single
variable case while the latter two make use of heavier matrix
algebra machinery.
Please visit each of these modules to see the definition and a
number of examples.
For a concrete application of these methods to a real dynamical
system, please visit the
Mass-Spring-Damper module.
Regardless of the approach, the matrix
exponential may be shown to obey the 3 lovely properties
-
ddtⅇAt=AⅇAt=ⅇAtA
t
A
t
A
A
t
A
t
A
-
ⅇAt1+t2=ⅇAt1ⅇAt2
A
t1
t2
A
t1
A
t2
-
ⅇAt
A
t
is nonsingular and
ⅇAt-1=ⅇ-At
A
t
A
t
Let us confirm each of these on the suite of examples used in
the submodules.
If
A=1002
A
1
0
0
2
then
ⅇAt=ⅇt00ⅇ2t
A
t
t
0
0
2
t
-
ddtⅇAt=ⅇt002ⅇ2t=1002ⅇt00ⅇ2t
t
A
t
t
0
0
2
2
t
1
0
0
2
t
0
0
2
t
-
ⅇt1+t200ⅇ2t1+2t2=ⅇt1ⅇt200ⅇ2t1ⅇ2t2=ⅇt100ⅇ2t1ⅇt200ⅇ2t2
t1
t2
0
0
2
t1
2
t2
t1
t2
0
0
2
t1
2
t2
t1
0
0
2
t1
t2
0
0
2
t2
-
ⅇAt-1=ⅇ-t00ⅇ-2t=ⅇ-At
A
t
t
0
0
2
t
A
t
If
A=01-10
A
0
1
-1
0
then
ⅇAt=costsint-sintcost
A
t
t
t
t
t
-
ddtⅇAt=-sintcost-cost-sint
t
A
t
t
t
t
t
and
AⅇAt=-sintcost-cost-sint
A
A
t
t
t
t
t
-
You will recognize this statement as a basic trig
identity
cost1+t2sint1+t2-sint1+t2cost1+t2=cost1sint1-sint1cost1cost2sint2-sint2cost2
t1
t2
t1
t2
t1
t2
t1
t2
t1
t1
t1
t1
t2
t2
t2
t2
-
Using the cofactor expansion we find
ⅇAt-1=cost-sintsintcost=cos-t-sin-tsin-tcos-t=ⅇ-At
A
t
t
t
t
t
t
t
t
t
A
t
If
A=0100
A
0
1
0
0
then
ⅇAt=1t01
A
t
1
t
0
1
-
ddtⅇAt=0100=AⅇAt
t
A
t
0
1
0
0
A
A
t
-
1t1+t201=1t1011t201
1
t1
t2
0
1
1
t1
0
1
1
t2
0
1
-
1t01-1=1-t01=ⅇ-At
1
t
0
1
1
t
0
1
A
t