MATH 2243: Linear Algebra & Differential Equations

Instructor: Jodin Morey [email protected]

Website: math.umn.edu/~moreyjc

7.1: First-Order Systems and Applications

We have discussed ways of solving DEQs involving one dependent variable: y f y, t , x f x , x, t , etc.

In this chapter, we will start to learn how to solve systems of DEQs involving several dependent variables. In particular, systems
which have as many equations as dependent variables. For example, this model of two (x, y) vibrating atoms:
m1x k1x k2 y x
m2y k2 y x f t

The matrix manipulation skills we learned will help us to solve these types of problems in coming sections.

To simplify these problems, however, we first introduce a technique to reduce the order of the DEQ. In particular, we have more
tools to solve first-order than higher-order DEQs. And it turns out we can transform higher-order into first-order DEQs.

Transforming Higher-Order DEQs into a System of First-Order DEQs

3
If you’re given: x 3x 2x 5x sin 2t
Then, define some new variables: x 0 : x, x1 : x x0 , x2 : x x1 ,
3
(note below that you don’t need the highest derivative x to be represented by a new variable)
Using this "dictionary," we can define a first-order system to replace our given DEQ.
Particularly, the first two DEQs listed below simply come from the dictionary itself.
The last DEQ is the given DEQ, but with the variable swapped out for the new ones.
x0 x1,
x1 x2,
x2 3x 2 2x 1 5x 0 sin 2t.
(notice that the old variable x is eliminated completely)

Why would we make this transformation? Because first-order DEQs are easier to solve, especially for systems of DEQs. The
price we pay is that now we have to solve three of them! Which is part of why we learned about matrices.

First-Order System Transformed to Second-Order DEQ

Now let’s go the other way around, turning a system of first-order DEQs into one higher-order DEQ.

Why would we make this transformation? Notice the system below is coupled (e.g., there is a y in the x DEQ). For simple low
dimensional coupled systems like this one, instead of using matrix methods (eigenvalues/eigenvectors), it may be simpler to do
the transformation below and solve the resulting DEQ using the simpler methods learned previously in the course.

1
Transforming and solving a 2D System: Example, x 2y, y 2
x.
Attempt to alter the equations toward the aim of substituting one into the other.
 The resulting equation should be either in x or y, depending on your choice above.
For example, above you can differentiate the first DEQ,
and then substitute the second DEQ into it:
1
x 2y 2 2
x x, x x 0.

Next, use your standard techniques to solve for x t .

Finally, use substitution again to plug x t into x 2y and solve for y t .

Existence and Uniqueness of Solutions for Linear Systems

Given: x 1 p 11 t x 1 p 12 t x 2 f1 t ,
x2 p 21 t x 1 p 22 t x 2 f2 t ,
x3 p 31 t x 1 p 32 t x 3 f3 t ,
with initial conditions: x 1 a b1, x2 a b2, x3 a b3.
If p 11 , p 12 , p 21 , p 22 , p 31 , p 32 , f 1 , f 2 , f 3 are continuous on some I containing t a,
then there exists a unique solution to the system x 1 t , x 2 t , x 3 t within the interval I.

Examples of Two-Dimensional Systems

x 2y,
1
y 2
x.

1
Taking the derivative of the first EQ: x 2y 2 2
x x.

So: x x 0 r2 1 0 r i e it cos t i sin t

xt A cos t B sin t.

1 1
From our given equations: y 2
x 2
A sin t B cos t .

planetary orbits, anything cyclic

And similarly for
x y,
y 2x y,
we find solution curves like:

topography
And for
x y,
101 1
y 100
x 5
y,
we find solution curves like:

Hurricanes, fluidic motion, decaying orbit

Why do such similar systems generate such different solution curves?

Stay tuned for an answer in subsequent sections.

Exercises

Problem: #8 Transform the following system into an equivalent system of first-orderDEQs.

x 3x 4x 2y 0, y 2y 3x y cos t.

Dictionary: x 0 : x, x 1 : x x 0 , and y 0 : y, y 1 : y y0.

(note that you don’t need the highest derivatives x or y )

Therefore,
x0 x1, y0 y1, (from the dictionary)
x1 4x 0 3x 1 2y 0 , y1 3x 0 2y 1 y0 cos t. (from the given DEQs)

I wrote the new DEQs this way anticipating the following form:
Let x : x 0 , x 1 , y 0 , y 1 . Then x Ax v OR
 x0 0 1 0 0 x0 0
x1 4 3 2 0 x1 0
.
y0 0 0 0 1 y0 0
y1 3 0 2 1 y1 cos t

Problem: #19 Find the general solution of the following system. Then, find the corresponding particular solution.
x y, y 13x 4y, x0 0, y0 3.

Note that they are coupled. Differentiating the first equation gives us:
x y 13x 4y

13x 4x . (we eliminated y !)

So, x 4x 13x 0

r2 4r 13 0
4 16 52
r 2
2 3i

2 3i t
e e 2t cos 3t i sin 3t

xt e 2t c 1 cos 3t c 2 sin 3t Then...?

x0 0 gives 0 c 1 e 2t , so c 1 0. Therefore, x t c 2 e 2t sin 3t.

y x gives yt 2c 2 e 2t sin 3t 3c 2 e 2t cos 3t . And...?

y0 3, so: 3 3c 2 , and c 2 1.

Therefore, x p t e 2t sin 3t and y p t e 2t 2 sin 3t 3 cos 3t

Problem: #24 Derive the DEQs: m1x1 k1 k2 x1 k2x2 and m2x2 k2x1 k2 k3 x2
for the displacements (from equilibrium) of the two masses shown below:
Recall Newton’s second law (for spring constant k): mx kx.

Looking at the above figure, we see that the first mass is pulled to the left (negatively) by the first spring and to the right
(positively) by the second spring. The second mass is pulled to the left (negatively) by the second spring and to the right
(positively) by the third spring. We also have to take into account the degree each spring is being stretched (to the right). For
instance, the second spring is being stretched to the right a distance of x 2 by mass m 2 , but the stretch is lessened by the distance
x 1 traveled by m 1 . So, the second spring is stretched: x 2 x 1 .

Hence for the first mass Newton’s second law gives:

m1x1 k1 x1 k2 x2 x1 k1 k2 x1 k2x2,

and for the second mass:

m2x2 k2 x2 x1 k3 0 x2 k2x1 k2 k3 x2.

What can you do with this?

atoms vibrating animation

Problem: #27 A particle of mass m moves in the plane with coordinates x t , y t under the influence of a force that is
directed toward the origin and has magnitude x 2 k y 2 , an inverse-square central force field. Show that:
kx ky
mx r3
and my r3
, where r x2 y2 .

k k
Let be the polar angular coordinate of the point x, y and write F x2 y2 r2
.

Also, let F x , F y be the components of force in the x, y directions, respectively.

Recall that cos and sin give the horizontal and vertical components of the vector lying on the unit circle pointing toward x, y .
y
Therefore, they can be represented as xr , r , where r x2 y2 .

Then Newton’s second law gives:

k x kx
mx F cos r2 r r3
,
k y ky
my F sin r2 r r3
.