Differential Equations Notes
Solving systems
In this lesson, we want to extend much of what we know about solutions
of second order equations to solutions of systems of differential
equations.
This will actually be a really helpful framework, because much of what we
know about the solution of a second order equation is consistent with
what we’ll see about the solution of a system of equations.
General solution
For instance, remember previously when we learned to solve second order
homogeneous equations, that the general solution was given by
y = c1y1 + c2 y2, where {y1, y2} formed a fundamental set of solutions.
Following this format produced formulas for the general solution with
distinct real roots, equal real roots, or complex conjugate roots of the
associated characteristic equation.
Roots General solution
Distinct real roots y(x) = c1e r1x + c2e r2 x
Equal real roots y(x) = c1e r1x + c2 xe r1x
Complex conjugate roots y(x) = e αx(c1 cos(βx) + c2 sin(βx))
where r = α ± βi
354
Differential Equations Notes
Similarly, the solution to a system of two homogeneous differential
equations, written in matrix form as x ′⃗ = A x ,⃗ when the matrix A is an 2 × 2
square matrix, will be given by
x ⃗ = c1 x1 ⃗ + c2 x2 ⃗
This time, instead of the solutions {y1, y2} which were functions in x, we’ll
have solutions { x1 ,⃗ x2 }⃗ which are vectors in t. The vectors { x1 ,⃗ x2 }⃗ will be
defined as
x1 ⃗ = k1 ⃗e λ1t
x2 ⃗ = k 2 ⃗ e λ 2 t
if λ1, λ2 are two non-zero real eigenvalues of matrix A. The vectors k1 ⃗ and k2 ⃗
will each have two components, when the matrix A in x ′⃗ = A x ⃗ is a 2 × 2
square matrix.
And when we’re solving a nonhomogeneous system of two equations
x ′⃗ = A x ⃗ + F, where F ≠ O,⃗ the general solution will be the sum of the
complementary solution xc ⃗ = c1 x1 ⃗ + c2 x2 ⃗ (the general solution to the
associated homogeneous equation x ′⃗ = A x ⃗ when the matrix A is an 2 × 2
square matrix) and particular solution xp .⃗
x ⃗ = xc ⃗ + xp ⃗
x ⃗ = c1 x1 ⃗ + c2 x2 ⃗ + xp ⃗
Verifying solutions
355
Differential Equations Notes
In the past, we learned how to show that a solution set like
{y1, y2} = {e −3x, xe −3x} satisfied a second order equation like y′′ + 6y′ + 9y = 0.
We simply substituted each solution, y1 = e −3x and y2 = xe −3x, one at a time
into the differential equation, to show that each one individually satisfied
the equation.
We can do the same thing with the vector solutions to the system of
differential equations. For example, we can show that
[2]
x1 ⃗ =
−1 − 32 t
e
is a solution to the system given in matrix form,
[1 −1]
x ′⃗ = x⃗
−1 4
We just need to substitute the solution vector and its derivative into the
matrix equation. The derivative of the solution vector is
2[2]
3
[−3]
3 −1 − 3 t
x1 ⃗ = −
′ 3
e 2 = 2 e− 2 t
Plugging the solution x1 ⃗ and its derivative x1 ⃗ into the matrix equation gives
′
[ ]
3 1
[−3] [1 −1]
− 32 t −1 −1 − 32 t
2 e = 4 e
2
3 −3t
−1 14
[−3e − 2 t] [ 1 −1] [ 2e − 32 t ]
e 2 −e − 2 t
3
2 =
3
356
Differential Equations Notes
(−1)(−e − 2 t ) + ( 14 )(2e − 2 t )
3 −3t 3 3
[−3e ]
e 2
2 =
− 32 t − 32 t − 32 t
(1)(−e ) + (−1)(2e )
3 −3t
e − 2 t + 12 e − 2 t
3 3
[−3e ] [−e − 32 t − 2e − 32 t]
e 2
2 =
− 32 t
3 −3t 3 −3t
[−3e ] [−3e − 32 t]
e 2 e 2
2 = 2
− 32 t
Because we get equivalent values on both sides of the equation, we’ve
verified that
[2]
x1 ⃗ =
−1 − 32 t
e
is, in fact, a solution to the system of differential equations given by the
matrix equation.
Superposition, linear independence, and the fundamental
set
Furthermore, in the same way that c1y1 + c2 y2 is also a solution to the
second order equation when {y1, y2} form a solution set, c1 x1 ⃗ + c2 x2 ⃗ is also a
solution to the system of differential equations when { x1 ,⃗ x2 }⃗ form a
solution set. This is the superposition principle.
And when it comes to the solution set { x1 ,⃗ x2 },
⃗ we’re primarily interested in
linearly independent solutions, which, just like with solutions to second
357
Differential Equations Notes
order equations, are solutions that aren’t constant multiples of one
another.
If we have just two solutions, they’ll be linearly independent if they aren’t
constant multiples of each other. If we have more than two solutions,
they’ll be linearly independent when none of the solutions can be formed
with a linear combination of the other solutions.
Just like before, we can use the Wronskian to test for linear independence.
⋮ ⋮ ⋮ ⋮
W( x1 ,⃗ x2 ,⃗ . . . , xn ⃗) = x1 ⃗ x2 ⃗ ⋯ xn ⃗
⋮ ⋮ ⋮ ⋮
If the Wronskian of the vector set is non-zero, then the vector set
represents a set of linearly independent solutions.
If the Wronskian is non-zero, that means the solution set is linearly
independent, and equivalently, we can call the solution set a fundamental
set of solutions.
Let’s do an example with a homogeneous system to try to tie all these
facts together.
Example
Verify that the solution vectors satisfy the system of equations. Calculate
the Wronskian of the solution set and use the result to make a statement
about the linear (in)dependence of the solutions, and whether or not they
form a fundamental set. Then write the general solution of the
homogeneous system.
358
Differential Equations Notes
0 6 0
x ′⃗ = 1 0 1 x⃗
1 1 0
6 −3 2
[1] [1]
{ x1 ,⃗ x2 ,⃗ x3 }⃗ = −1 e , 1 e , 1 e 3t
−t −2t
−5
We can verify that each of the vectors { x1 ⃗, x2 ⃗, x3 ⃗} are solutions by
substituting them into the matrix equation. First though, we’ll need the
derivative of each solution vector.
6e −t −6e −t
x1 ⃗ = −e −t x1 ⃗ =
′
e −t
−5e −t 5e −t
−3e −2t 6e −2t
x2 ⃗ = x2 ⃗ = −2e −2t
′
e −2t
e −2t −2e −2t
2e 3t 6e 3t
x3 ⃗ = e 3t x3 ⃗ = 3e 3t
′
e 3t 3e 3t
Substitute x1 ⃗ and its derivative x1 ⃗ into the matrix equation.
′
−6e −t 0 6 0 6e −t
e −t = 1 0 1 −e −t
5e −t 1 1 0 −5e −t
359
Differential Equations Notes
−6e −t (0)(6e −t ) + (6)(−e −t ) + (0)(−5e −t )
e −t = (1)(6e −t ) + (0)(−e −t ) + (1)(−5e −t )
5e −t (1)(6e −t ) + (1)(−e −t ) + (0)(−5e −t )
−6e −t −6e −t
e −t = 6e −t − 5e −t
5e −t 6e −t − e −t
−6e −t −6e −t
e −t = e −t
5e −t 5e −t
Substitute x2 ⃗ and its derivative x2 ⃗ into the matrix equation.
′
6e −2t 0 6 0 −3e −2t
−2e −2t = 1 0 1 e −2t
−2e −2t 1 1 0 e −2t
6e −2t (0)(−3e −2t ) + (6)(e −2t ) + (0)(e −2t )
−2e −2t = (1)(−3e −2t ) + (0)(e −2t ) + (1)(e −2t )
−2e −2t (1)(−3e −2t ) + (1)(e −2t ) + (0)(e −2t )
6e −2t 6e −2t
−2e −2t = −3e −2t + e −2t
−2e −2t −3e −2t + e −2t
6e −2t 6e −2t
−2e −2t = −2e −2t
−2e −2t −2e −2t
360
Differential Equations Notes
Substitute x3 ⃗ and its derivative x3 ⃗ into the matrix equation.
′
6e 3t 0 6 0 2e 3t
3e 3t = 1 0 1 e 3t
3e 3t 1 1 0 e 3t
6e 3t (0)(2e 3t ) + (6)(e 3t ) + (0)(e 3t )
3e 3t = (1)(2e 3t ) + (0)(e 3t ) + (1)(e 3t )
3e 3t (1)(2e 3t ) + (1)(e 3t ) + (0)(e 3t )
6e 3t 6e 3t
3e 3t = 2e 3t + e 3t
3e 3t 2e 3t + e 3t
6e 3t 6e 3t
3e 3t = 3e 3t
3e 3t 3e 3t
Because we found equivalent vectors on both sides of the equation in
each of these three cases, we’ve shown that all three vectors are solutions
to the system of differential equations.
Now let’s take the Wronskian of the solution set.
6e −t −3e −2t 2e 3t
W( x1 ,⃗ x2 ,⃗ x3 ⃗) = −e −t e −2t e 3t
−5e −t e −2t e 3t
e −2t e 3t −e −t e 3t −t
e −2t
W( x1 ,⃗ x2 ,⃗ x3 ⃗) = 6e −t −2t
− (−3e ) −t 3t
+ 2e 3t −e
e −2t
e 3t
−5e e −5e −t e −2t
361
Differential Equations Notes
W( x1 ,⃗ x2 ,⃗ x3 ⃗) = 6e −t[(e −2t )(e 3t ) − (e 3t )(e −2t )]
+3e −2t[(−e −t )(e 3t ) − (e 3t )(−5e −t )] + 2e 3t[(−e −t )(e −2t ) − (e −2t )(−5e −t )]
W( x1 ,⃗ x2 ,⃗ x3 ⃗) = 6e −t(e t − e t ) + 3e −2t(−e 2t + 5e 2t ) + 2e 3t(−e −3t + 5e −3t )
W( x1 ⃗, x2 ⃗, x3 ⃗) = 6e −t(0) + 3e −2t(4e 2t ) + 2e 3t(4e −3t )
W( x1 ,⃗ x2 ,⃗ x3 ⃗) = 12e 2t−2t + 8e 3t−3t
W( x1 ,⃗ x2 ,⃗ x3 ⃗) = 12(1) + 8(1)
W( x1 ,⃗ x2 ,⃗ x3 ⃗) = 20
Because the Wronskian is non-zero, W( x1 ,⃗ x2 ,⃗ x3 ⃗) = 20 ≠ 0, we can confirm
that the vectors in the solution set are linearly independent, which means
that the vector set represents a fundamental set of solutions.
Given that this is a fundamental set of solutions, we can write the general
solution of the system as
x ⃗ = c1 x1 ⃗ + c2 x2 ⃗ + c3 x3 ⃗
6 −3 2
[1] [1]
x ⃗ = c1 −1 e + c2 1 e + c3 1 e 3t
−t −2t
−5
362