c Dr.

Igor Zelenko, Fall 2017 1

9: Systems of FIRST Order Equations and their relation to higher order equations
(section 7.1 )

1. A first order system of ordinary differential equations (ODEs):

x01 = F1 (t, x1 , x2 , . . . , xn )
x02 = F2 (t, x1 , x2 , . . . , xn )
.. (1)
.
x0n = Fn (t, x1 , x2 , . . . , xn )

2. A set of differentiable functions x1 (t), x2 (t), . . . , xn (t) satisfying the system (1) is called a
solution of the system (1).

3. System of ODE using a vector notation:

   
x1 F1 (t, x1 , x2 , . . . , xn )
 x2   F2 (t, x1 , x2 , . . . , xn )
   

X=  ..  , F = 
  ..  (2)
 .  .

 
xn Fn (t, x1 , x2 , . . . , xn )

Then the system (1) can be written as

X0 = F(t, X). (3)

Symbolically this is exactly the same expression as for a single first order equation.

DEFINITION 1. The system (1) is called autonomous if the right-hand side of it is inde-
pendent of t, i.e. is of the form F(X),

X0 = F(x). (4)
and non-autonomous otherwise.

REMARK 2. Autonomous systems are special first order systems. However, any non-
autonomous system on n unknown functions (x1 , . . . , xn ) of t can be seen as an autonomous
system in n + 1-unknown functions (x0 , x1 , . . . , xn+1 ) such that x00 (t) = 1. Namely, we
consider the following system of n + 1 equations of n + 1 unknown functions:

The very first equation of this system implies that if x0 (0) = 0 then x0 (t) = t and then
the column vector function X as in (2) is the solution of the original system (1). Therefore
conceptually we can restrict ourselves to autonomous systems only.

4. Vector fields and autonomous first order systems:


c Dr. Igor Zelenko, Fall 2017 2

5. A vector field F on Rn : at each point X of Rn a vector F(X) starting at this point X is

given.

6. An integral curve (integral trajectory X(t) of a vector field F is a curve in Rn such that the
velocity X0 (t) to this curve at every its point X(t) (or, equivalently, at every time moment
t ) coincides with the vector fields F at this point, i.e. with the vector F (X(t)).
In other words,

X0 (t) = F (X(t))

i.e. X(t) is an integral curve of the field F if and only if it is a solution of the autonomous
equation (4).

REMARK 3. One can define the analog if direction field for nonautonomous system (1):
it is a direction field in Rn+1 such that the line segment is generated by the vector field of
the corresponding autonomous system in Rn+1 as described in Remark 2.

7. To any autonomous system of n equations with n unknown function one can assign a vector
field in Rn and vice versa, to any vector field F in Rn corresponds an autonomous system
(4). Then Rn is called the phase space of the system (4). More generally, the vector field
may be defined not on the whole Rn but in some region R of Rn or , for example, on a
surface or higher dimensional analog of a surface S (like a sphere, a torus etc, depending on
a model) in Rn (in the latter case the vector field must be tangent to such surface). In this
case the sets R and S are also called phase spaces.
The phase portrait of a system is a representative sketch of integral curves of the system on
the phase space.

8. For n = 1 we have just one equation and the phase space in the autonomous case is just a
line R (what we called the phase line in the previous section). The nonzero vectors tangent
to R have only two directions, positive and negative, so those directions were the only what
matters when we analyzed the behaviour of the solutions. For n > 1 there are much richer
variety of how the phase portrait may look like compared to the case when n = 1.
Here are several examples of sketch of the vector fields corresponding to a given system
of two equations using pplane. The corresponding phase portraits can be sketched by
drawing several representative phase lines. I mark each example with certain name but at
this moment you do not have to put any attention on those names

EXAMPLE 4 (saddle point). (

x01 = x1 + x2
x02 = x1 − x2

c Dr. Igor Zelenko, Fall 2017 3

x´ = x+y
y´ = x-y
10

y 0

-2

-4

-6

-8

-10
-10 -8 -6 -4 -2 0 2 4 6 8 10

[spiral source]
EXAMPLE 5. (
x01 = x1 + x2
x02 = −x1 + x2

x´ = x+y
y´ = -x+y
10

y 0

-2

-4

-6

-8

-10
-10 -8 -6 -4 -2 0 2 4 6 8 10

x

c Dr. Igor Zelenko, Fall 2017 4

EXAMPLE 6 (center). (
x01 = 3x1 + 4x2
x02 = −4x1 − 3x2

x´ = 3x+4y
y´ = -4x-3y
10

y 0

-2

-4

-6

-8

-10
-10 -8 -6 -4 -2 0 2 4 6 8 10

EXAMPLE 7 (competing species).

(
x0 = x(7 − x − 2y)
y 0 = y(5 − y − x)

c Dr. Igor Zelenko, Fall 2017 5

x´ = x(7-x-2y)
y´ = y(5-y-x)
6

5.5

4.5

3.5

y 3

2.5

1.5

0.5

0
0 1 2 3 4 5 6 7 8

We will learn how to solve explicitly systems in Examples 4-6, which are linear homogeneous
systems with constant coefficients and how to analyze the nonlinear system in Example 7
based on the theory of linear systems (without the knowledge of this theory the software
will not be really useful).

9. Existence and Uniqueness Theorem for IVP defined by a system: Consider the IVP:
x01 = F1 (t, x1 , x2 , . . . , xn )
x02 = F2 (t, x1 , x2 , . . . , xn )
..
.
x0n = Fn (t, x1 , x2 , . . . , xn )
(5)
x1 (t0 ) = x01
x2 (t0 ) = x02
..
.
xn (t0 ) = x0n
is literally the same as Theorem 3 in section 7 of the notes devoted to single equation: If each
∂F1 ∂F2 ∂Fn
of the functions F1 , F2 , . . . , Fn and the partial derivatives , ,..., (1 ≤ k ≤ n)
∂xk ∂xk ∂xk
are continuous in a region

R = {α < t < β, α1 < x1 < β1 , α2 < x1 < β2 , . . . , αn < xn < βn }

and the point (t0 , x01 , . . . , x0n ) belongs to R, then there is an interval (t0 − h, t0 + h) in which
there exists a unique solution of the IVP (5).

c Dr. Igor Zelenko, Fall 2017 6

How to transform a scalar ODE or order n to a system

of n first order equations
10. Any scalar ODE equation of order n,

y (n) = f (t, y, y 0 , y 00 , . . . , y (n−1) )

can be transformed to a system of n DE of the first order by introducing derivatives up to order

n − 1 as new variables.

11. To transform the following n-th order IVP,

y (n) = f (t, y, y 0 , y 00 , . . . , y (n−1) ), (6)

y(t0 ) = α0 , y 0 (t0 ) = α1 , . . . , y (n−1) (t0 ) = αn−1

into the system of first order equations we set

x1 (t) = y(t)
x2 (t) = y 0 (t)
..
.
xn (t) = y (n−1) (t)

to get
x01 = x2
x02 = x3
.. (7)
.
x0n = f (t, x1 , x2 , . . . , xn )
subject to
x1 (t0 ) = α0 , x2 (t0 ) = α1 , . . . , xn (t0 ) = αn−1 .

12. Consider the following ODE of unforced undamped vibration:

y 00 + y = 0. (8)

Transform (8) into a system of first order ODE. Is the obtained system autonomous?

13. Transform the equation

y (3) + (sin t)y 00 + et ((y 0 )2 + y 2 ) = 0

to the system of differential equations.


c Dr. Igor Zelenko, Fall 2017 7

14. An obvious but very important remark is:

REMARK 8. Afunction y(t) is a solution of the equation (6) if an only if

 
y(t)
y 0 (t)
 
 
X(t) =  .. 
.
 
 
y (n−1) (t)

is a solution of the corresponding first order system (7).

This simple passage from single differential equation (6) of higher order to first-order system
(7) allows us to consider the theory of second order and even higher order equations discussed
in chapter 3 and 4 of the textbook as a particular case of the theory of systems of first order
equation discussed in chapter 7. So, instead of first covering chapter 3 and then repeating
the same in more general setting of chapter 7 we will start now with chapter 7 and treat the
material of chapter 3 and (and even of chapter 4) simultaneously.