CHAPTER 1

Explicitly Solvable First Order Differential

Equations
In this chapter we will introduce the first order differential equations
and several type of it that can be solved explicitly using integration
methods covered in standard Calculus course.
From Calculus, we know that the derivative dfdt(t) of a function f (t)
is the rate at which the quantity x = f (t) is changing with respect
to the independent variable t. An equation relating an unknown
function and one or more of its derivatives is called an differential
equation. Particulary, an equation of the form x0 = f (t, x) is called a
first order ordinary differential equation(ODE), more precisely a semi
linear first order ordinary differential equation. In this chapter will
will demonstrate how to find explicit solution(s) to a given ODE . In
general one cant find explicit solution to a given ODE , but for special
types of f (t, x), we will have luck to do that.
Some clearance of notations is needed here.
In x0 = f (t, x), t is the independent variable and x(t) is the unknown
function, x0 is dx
, the first derivative of x(t), and f (t, x) is a given
dt
function of two variables t and x. So to find an explicit solution is to
find an formula for x(t) so that when substitute the formula in f (t, x),
we will get an expression in t only that matches x0 (t). For example, for
example, suppose we have f (t, x) = x + t, our ODE reads, x0 = x + t. If
we set x(t) = 2et t 1 then f (t, x) = x + t = 2et t 1 + t = 2eT 1
and x0 (t) = 2et 1 so x0 = f (t, x) and x(t) = 2et t 1 is an solution.
On the other hand if we write y 0 = f (x, y), then in this ODE , x
is the independent variable, y(x) is the unknown function to be found,
dy
and y 0 is dx
, the first derivative of y(x). So for y 0 = y + x, we would
get an solution y(x) = ex x 1.
Sometimes the independent variable does not appear in the equation such as y 0 = f (y) or x0 = f (x) in those case, one could assume the
independent variable is either x (for y 0 = f (y)) or t (for x0 = f (x)) and
write the solution accordingly. Hence, for y 0 = y one solution would be
y(x) = ex and for x0 = x, one would write as solution as x(t) = et .
1

1. EXPLICITLY SOLVABLE FIRST ORDER DIFFERENTIAL EQUATIONS

Note 0.1. It is very important for students to know what variable

represents an unknown function and what variable represents the independent variable that the unknown function is defined with respect
to.
1. Finding antiderivative using Mathcad
You can use Mathcad to find both definite and indefinite integral.
To access those two operators, you can bring up the popup menu, (See

Figure 1. Calculus tool bar

chapter 1 for how to bring up the Calculus popup menu.) Click on the
Rb
R
symbol
to
get
the
operator
for
definite
integral
and
on
the
symbol
a
to get the operator for the indefinite operator. Then just fill in the
place holders accordingly. For definite integral, hit = to get the value
and for indefinite integral, you need either
Select the entire expression and from the Symbolics menu
following the menu item evaluate to the symbolically(Or
[Shift][F9]).
or
Hold down the [Ctrl] key and press the [.] (period) key.
You can also access the integration operators by hot keys [Shift][F7]
for definite integral and [Ctrl][I] for indefinite integral.
The integration operators appear on the workplace as shown in the
following diagram,
Now lets see some examples. The first example is for finding definite
integral.
R3
Example 1.1. Find the definite integral 0 (x5 e3x sin(3x)) dx.
Solution At any blank space in the workplace of Mathcad , type
[Shift][F7], that is hold the [Shift] key and press [F7] key to place the
definite integration operator
Z
d

1. FINDING ANTIDERIVATIVE USING Mathcad

Figure 2. Integration Operators

. Click the upper limit place holder, type 3 and click the lower limit
place holder type 0. Now click the integrant place holder type
x^5 -e^3x

-sin(3x)

notice you need it the space bar twice before -sin(3x) . Of

course you need hit space bar once before entering -e^3x to exit the
power mode. Finally in the integration variable place holder type x(see
Figure 2 for the position of each place holder.) Now hit [=] key if in
automatic mode or =[F9] otherwise to find the value 2.5 103 .
a
The second example is for find indefinite integral.
R
Example 1.2. Find the definite integral ( x + 2e3x sin(3x)) dx.
Solution
At any blank space in the workplace of Mathcad ,
type [Ctrl][I], that is hold the [Ctrl] key and press [I] key to place the
indefinite integration operator
Z
d.
Click the integrant place holder and type
\x+2

-e^3x

-sin(3x)

Notice you need to hit the space bar twice before -sin(3x) . Of
course you also need hit space bar twice too before entering -e^3x to
exit the square root operator.
In general, you will need to hit the space bar as many times
an need to escape the power(subscript, division, etc.) mode
before you enter next term.
The backslash \ brings up the square root operator. Finally in
the integration variable place holder type x(see Figure 2 for the position
of each place holder.) Now hold down[Ctrl] to type [.] key and click on

1. EXPLICITLY SOLVABLE FIRST ORDER DIFFERENTIAL EQUATIONS

any region outside the bounding box, if not in the auto execution
mode hit [F9] too, to find the antiderivative
Z

3
2
1
1
x + 2 e3x sin(3x) dx (x + 2) 2 exp3x + cos(3x)
3
3
3
. Here exp represent exponential function. From Calculus, we know
that
Z

3
2
1
1
x + 2 e3x sin(3x) dx = (x + 2) 2 exp3x + cos(3x) + C.
3
3
3
Mathcad omits the C for believing you know it! Remember C is important when we try to find a particular solution for a given, so called,
initial condition.
Example 1.3. Find a function x(t) such that x0 (t) = 2t and x(0) =
3.
Solution

If we take antiderivative on both sides, we have

Z
Z
0
x (t) dt = 2t dt

which leads to
x(t) = t2 + C.

(1)

Evaluate this x(t) at t = 0 we have

x(0) = 02 + C = C
So the requirement x(0) = 3 implies C = 3. Hence x(t) = t2 + 3.
Notice if we drop the C from (1), we would not be able to find x(t)
that satisfies x(0) = 3!
a
The problem of finding a particular solution with initial condition is
called initial value problem, which we will discuss in the later section. a
Practice
1. Find the following definite integral using Mathcad .
R2
(a) 1 x3 x cos(x2 + 1) dx
R4 3
dx
(b) 1 x +4x5
x+1
R 0 3x
(c) 5 e x2x+3 dx
R 10
(d) 1 x2 x sec(x2 + 1) dx
2. Find the following indefinite integral using Mathcad .
p
R
(a) x5 x (x2 + 1) dx
R 3
dx
(b) x x+4x5
2 +1

2. SEPARABLE EQUATIONS

R
x3
(c) (x2 +3)(x+5)
dx
R
2
(d) x tan(x + 1) dx

2. Separable equations
If we let x be the independent variable and y be the dependent
variable, a separable ODE is given by,
y 0 = f (x)g(y)
dy
where both f and g are given one variable functions and y 0 = dx
is the
derivative of y(x) with respect to x and y(x) is the unknown function
we want to find out.
When g(y) 1, that is g(y) is a constant function whose value is
always 1, the separable ODE becomes

y 0 = f (x)
, which is sometimes called integrable ODE . The general solution to
the integrable ODE is the antiderivative of f (x), i.e.
Z
y(x) = f (x) dx
So if we can find the antiderivative of f (x) we can find the explicit
general solution.
Example 2.1. Find solution y 0 = x2 + 3
Solution

The solution is,

Z
1
y(x) = x2 + 3 dx = x3 + 3x + C
3
. Here we used the power rule of antiderivative,
Z
xn+1
xn dx =
+ C.
n+1
To find the solution in Mathcad , you just click any blank area in the
workplace and type [Ctrl][I] to get the indefinite operator
Z
d
and type x^2 +3 in the place holder for integrant and x in the integrating variable place holder. Then press [Ctrl][.] and click on any
region outside the bounding box to get the result. Sometime you
might need to hit [F9] too if the auto execution mode is turn off.
a

1. EXPLICITLY SOLVABLE FIRST ORDER DIFFERENTIAL EQUATIONS

When g(y) is not a constant function, the general solution to

y 0 = f (x)g(y)
is given by the equation
Z
(2)

dy
=
g(y)

Z
f (x) dx,

which is obtained by dividing both sides of the equation by g(y) and

then taking antiderivative to both sides. R
R
dy
So one has to find two antiderivatives g(y)
and f (x) dx. Sometimes we can solution for y(x) explicit from the (2), sometimes we cant
find explicit formula for y(x), in that case we say y(x) is an implicitly
defined function.
Notice, in the process of finding the general solution to the separable
ODE , we divide both sides of the equation by g(y), that implicitly
requires that g(y) 6= 0. On the other hand if g(c) = 0 then y(x) = c is
an solution to the separable equation
y 0 = f (x)g(y)
This kind solutions are called stationary or steady-state solutions.
Sometimes they are also called singular solution.
Example 2.2. In the study of Epidemics, the spread of disease in
a population can be modelled by I 0 = (b(t)S(t) r(t))I where S(t) is
number of people that is susceptible to the disease but not infected yet.
I(t) is number of people actually infected. b, r are proportional function.
Now suppose S(t) = e3t , b = t2 , and r = 4, find all solutions.
Solution
Then, to

First notice that I(t) 0 is the steady-state solution.

dI
= (b(t)S(t) r(t))I,
dt
divide its both sides by I, multiplying its both sides by dt and taking
antiderivative we have
Z
Z
dI
= b(t)S(t) r(t) dt
I
The left side is easy to find,
Z
dI
(3)
= ln |I| + C
I
Plug in the given functions, we get the right hand side
Z
Z
b(t)S(t) r(t) dt = t2 e3t 4 dt
I0 =

2. SEPARABLE EQUATIONS

Now open Mathcad window, click any blank area, type [Ctrl][I] to get
the antiderivative operator
Z
d
type t^2 *e^{-3t}
- 4 in the integrant place holder (again using
space bar to exit the power mode) and t in the other place holder, then
type [Ctrl][.], click any area outside the box. You will get
Z
1
2
2
(4)
t2 e3t 4 dt = t2 e3t te3t e3t 4t + C
3
9
27
Put (3) and (4) together, we have
ln |I(t)| = 13 t2 e3t 29 te3t

2 3t
e
27

4t + C

Solve for I(t), we finally arrive at

1 2 3t

3t

3t

I(t) = e 3 t e 9 te 27 e 4t+C
1 2 3t 2 3t
2 3t
= Ce 3 t e 9 te 27 e 4t
a
The next example shows that sometimes we can find explicit formula for the solution of an separable ODE .
Example 2.3. Find general solution of y 0 =

ex +2x
sin(y)+3y 2 +3

dy
Solution Write y 0 as dx
, multiply both sides of the equation by
2
(sin(y) + 3y + 3)dy and take antiderivative, we have
Z
Z
2
sin(y) + 3y + 3 dy = ex + 2x dx

Now using Mathcad or compute by hand, we find

Z
sin(y) + 3y 2 + 3 dy = cos(y) + y 3 + 3y + C
and

Z
ex + 2x dx = ex + x2 + C

So we have an implicitly defined solution y(x) by the equation

cos(y) + y 3 + 3y = ex + x2 + C
Here, we cant find an explicit formula for y(t).
Practice

1. EXPLICITLY SOLVABLE FIRST ORDER DIFFERENTIAL EQUATIONS

Project

3. Linear equations and Bernoulli equations

3.1. Linear Equations. An ODE of form
y 0 + a(x)y = b(x)
is called a linear equation. Any equation that can be converted into
this form can also be called a linear equation. Such as x2 y 0 + 4x2 ex y =
sin(x), which can be written as y 0 + ex y = sin(x)
, is a linear equation.
x2
The general solution of the linear equation is
Z
!
R

y(x) = e

a(x) dx

b(x)e

a(x) dx

dx + C

Notice that C is inside the parenthesis.

It is common mistake of
R
students forgetting multiplying C by e a(x) dx .
Example 3.1. According to Newtons law of cooling, the time rate
of change of the temperature T (t) of a air conditioned house with external temperature A(t) is proportional to the difference A T . That
is,
dT
= k(A T ),
t
where k is positive constant. Now, suppose A(t) = 80 10 cos(t) find
the general solution.
Solution Since we want to use Mathcad to find the antiderivative
for us, the first thing is to find how to get the Greek letter . To get
, first bring the math tool bar from the View menu

Figure 3. Math Toolbar

3. LINEAR EQUATIONS AND BERNOULLI EQUATIONS

Click on the

Figure 4. Greek Letter Button

which will bring up

Figure 5. Greek Letter Bar

and select the letter.
Now if we rewrite and put in A(t)
dT
= k(A T ),
t
we get
dT
= k(80 10 cos(t)) kT,
t
and is a linear equation
withR a(t) = k and b(t) = k(80 10 cos(t)), it
R
easy to see that a(t) dt = k dt = kt. Apply the solution formula we
have
Z
!

T (t) = ekt

kekt 80 10 cos(t) dt + C .

After bring up the indefinite operator

Z
d
type k*e^k*t (80-10cos(t) as integrant and t as integration variable, press [Ctrl][.], click any place outside the box, we get,

k
kt
kt
kt
kt
e cos(t)+ 2
e sin(t)
ke 8010 cos(t) dt = 80e 10k 2
k + 2
k + 2
So the solution is

k
ekt cos(t)+ 2
ekt sin(t) +C
T (t) = ekt 80ekt 10k 2
k + 2
k + 2

10 1. EXPLICITLY SOLVABLE FIRST ORDER DIFFERENTIAL EQUATIONS

a
Example 3.2. If in Example 3.1, we set k = 0.2 and = 4 , we
would have general solution

3.2
4
0.2t
0.2t
0.2t
0.2t
80e 5
T (t) = e
e cos(t)+
e sin(t) +C
0.64 + 2
0.64 + 2
The following graph displayed curves of several solutions with different
initial values T (0).

Figure 6. Greek Letter Bar

The picture above exhibits the situation when the air condition fails
at t = 0, so after a out 14 hours, the room temperature is the same as
the outside temperature!
Practice
(1) Find general solution to the following problem
(a) y 0 + 4xy = x2
(b) ex y 0 + ex y = sin(x)
(c) xy 0 = 2y + x3 cos(x)
2
(d) y 0 = 2xy + 3x2 ex
(e) (x2 + 4)y 0 + 3xy = x
3 2
(f) (x2 + 1)y 0 + 3x3 y = 6xe 2 x
(2) Application: The equation y 0 = ky models wide range of
natural phenomenaany involving a quantity whose time rate
of change is proportional to its current size.
Continuously Compounded Interest Rate: Let P (t)
denote principle at time t, which will earn interest with rate
= rP. Now
r(t) and compounded continuously, we have dP
dt

3. LINEAR EQUATIONS AND BERNOULLI EQUATIONS

11

suppose r(t) = cos2 (t), find the general solution and graph
the solution for several different value for the constant C in
the same coordinate. Explain the long time behavior of the
solutions.
Drug Elimination: Let A(t) be the amount of certain drug in the bloodstream, measured by the excess over the
natural level of the drug. Then in many situations A(t) will
decline at a rate proportional to the current excess amount.
That is
dA
= A,
dt
where > 0 is called the elimination parameter of the drug.
Suppose for one type of drug, (t) = t21+4 find the general
solution. Discuss its longtime behavior by graphing several
solutions in the same coordinate.

3.2. Bernoulli Equations. A Bernoulli equation is an ODE of

the form
(5)

dy
+ a(x)y = b(x)y n .
dx

You can see that when n = 1, we get linear equation. So linear equation
is a special case of Bernoulli equation. On the other hand if we let
v = y 1n , then
dv
dy
= (1 n)y n
dx
dx
n
Dividing both side of (5) by y , we have
y n

dy
+ a(x)y 1n = b(x)
dx

So
(6)

1 dv
+ a(x)v = b(x).
1 n dx

That is using the transformation v = y 1n , the Bernoulli equation for

y becomes a linear equation for v.
Therefore to find the general solution for Bernoulli equation (5) we
first find the general solution for the linear equation (6) of v, then using
1
y = v 1n to find y.

12 1. EXPLICITLY SOLVABLE FIRST ORDER DIFFERENTIAL EQUATIONS

Example 3.3. In modelling a population with its births(per unit

time) proportional to current population level and the deaths(per unit
time) is proportional to the square of the current population, we have
dx
= ax bx2
dt
, which is called logistic differential equation. It is a Bernoulli equation
with n = 2 (Notice it is also a separable equation if a, b are constants.)

Solution Let v = x12 = x1 , then xv = 1 apply product rule of

differentiation, we have x0 v + v 0 x = 0, plug in x0 = ax bx2 , we have
(ax bx2 )v + v 0 x = axv bx2 v + v 0 x = a bx + v 0 x = 0 since xv = 1
So
(7)

a bx + v 0 x = 0
1
x

Divide (7) by x and notice

is v, we have

v 0 + av b = 0,
or
v 0 + av = b.
Apply the solution formula for linear equation we have
Z
!
v = e

a dt

be

So
1
x= =
v
R

e
be

a dt

dt + C

a dt

a dt

dt + C
a

Example 3.4. In the logistic model,

dx
= ax bx2
dt
, suppose a > 0, b > 0 are constants, we will discuss behavior of the
solutions.

3. LINEAR EQUATIONS AND BERNOULLI EQUATIONS

Solution

13

From Example 3.3,

x=

e
R

be

a dt

a dt

dt + C

So
x=

eat
R

eat

!=

beat dt + C

b at
e
a

!,

+C

is the general solution.

When C = 0 we have x = ab , a stead-state solution. Also, x = 0 is
another steady-state solution, which cant be represented in the general
solution.
We can rewrite x(t) as
x(t) =

a
,
b + aCeat

since limt eat = 0 So we have limt = ab . That is, as time goes

on the population will gradually settles itself at the level of ab . The
following picture shows that the solution curves approaches the line
y = ab = 10, with a = 0.2 and b = 0.002.

Figure 7. Population a = 0.2 and b = 0.02

14 1. EXPLICITLY SOLVABLE FIRST ORDER DIFFERENTIAL EQUATIONS

Here is how to get the graph in Mathcad :

0.2
- we first defined a function x(t, C):= 0.02+Ce
0.2t with two arguments t and C by typing x(t,C):0.2/0.02+C*e^-0.2t,
- then we defined a range variable t:=0,0.1. . .20 by typing
t:0,0.1;20,
- finally, type @ at a blank area to get the xy-plot; in the function
place holder, we put
x(t,0),x(t,-0.01),x(t,-0.008),x(t,0.004),x(t,0.01)
and type t in the variable place holder, and click at outside of the
box to get the graph.

4. Initial value problem and existence theorem

The problem of finding a particular solution x(t) of an ODE x0 =
f (t, x) that satisfying condition x(t0 ) = x0 is called an initial value
problem. The problem can be written
0
x = f (t, x)
x(t0 ) = x0
Since solving an ODE requires a lots of hard work, it is wise to know if
a given equation has any solution before march on the journey of find
a solution. For the linear first order ODE x0 + a(t)x = b(t) we following
existence-uniqueness theorem.
Theorem 4.1. If the function a(t) and b(t) are continuous on the
open interval I containing the point t0 , then the initial value problem
x0 + a(t)x = b(t),
x(t0 ) = x0 has a unique solution x(t) on I, given
by
Z
!
R
R
(8)
x(t) = e a(t) dt
b(t)e a(t) dt dt + C
with an appropriate value of C.
So solving an initial value problem for linear first order ODE is easy.
One just apply the formula (8) and using the equation x(t0 ) = x0 to
find C.
Example 4.1. Solve the initial value problem x0 +2tx = 4t3 , x(0) =
1.
Solution

First we apply the formula (8)

Z
R

x(t) = e

a(t) dt

b(t)e

a(t) dt

!
dt + C ,

4. INITIAL VALUE PROBLEM AND EXISTENCE THEOREM

15

R
R
with a(t) = 2t and b(t) = 4t3 . So a(t) dt = 2t dt = t2 , and
Z
!
x(t) = et

4t3 et dt + C .

Using method of substitution or Mathcad , we find

t2
2 t2
t2
x(t) = e
2t e 2e + C
Let t = 0 we have x(0) = C 2. From initial condition x(0) = 1 we see
that C = 3. So the solution is

t2
2 t2
t2
x(t) = e
2t e 2e + 3
a
However, for general first order ODE , the problem is not so nice,
the following theorem is a little bit hard to state and to understand.
Theorem 4.2. Let f (t, x) be defined on interval I. If the partial
(t,x)
derivative of f (t, x) with respect to y (denoted as fy
, reads partial f(t,x) over partial y) and f (t, x) are continuous at an open disk
centered at (t0 , x0 ), then the initial value problem
0
x = f (t, x)
x(t0 ) = x0
has an unique solution on some open interval containing the point t0 .
Remark 4.1. We will not discuss this existence and uniqueness
theorem for general ODE . But we would like to make some comments.
(t,x)
- To find ft
, you just need to treat x variable as a constant,
(t,x)
for example, let f (t, x) = t2 x3 + 3t 4x, we have ft
=
3
2tx +3, here 4x is a constant with respect to t, so the derivative
is 0, and x3 is also a constant, so t2 x3 when taking derivative
against t we have 2tx3 . Therefore finding partial derivative is
as easy as to find ordinary derivative.
- This theorem only guarantees a solution defined for t in an
open interval, which might be a bounded interval. For example
1
is the general solution, if
x0 = x2 , we see that x(t) = t+c
c = 2 then the solution is only defined on interval (0, 2) as
1
x(t) = t2
will become undefined at t = 2.
- Some time we could have more than one solution that satisfies
2
the same initial condition, for example, for x0 = x 3 and x(0) =
0 we have two different solution, one is x(t) = 0 another one

16 1. EXPLICITLY SOLVABLE FIRST ORDER DIFFERENTIAL EQUATIONS

1

(t,x)
1
is x(t) = t 3 . The reason is that for f (t, x) = x 3 , fx
=
3x
which is undefined at (0, 0). So the conditions of the theorem
are violated.

5. Phase and vector field diagrams

Sometimes it is more important to know the behavior of solutions
than to know precisely detail of point-by-point of some particular solutions. Here we will discuss two diagrams to get an feel of behavior
of solutions for a ODE without actually find any solution. The first,
phase diagram, is used for the autonomous equations and is easy to
draw. The second, vector field diagram, can be used for any equation but harder to draw. These two diagrams make use of the facts:
(1) The derivative f 0 (x) represent the slope of tangent line to the
graph y = f (x) at the point (x, f (x)).
(2) The sign of f 0 (x) tell whether the graph is going up or down.
If f 0 (x) > 0 the graph is increasing.
If f 0 (x) < 0 the graph is going down.
(3) The second derivative f 00 (x) gives information about concavity:
If f 00 (x) > 0 the graph is concave upward.
If f 00 (x) < 0 the graph is concave downward.
The concepts of concavity (up or down) and monotonicity (going up
or down) is illustrated in the following diagram.

Figure 8. Concavity and Monotonicity

5.1. The phase diagram. First-order equations whose right-hand
side does not depend on the independent variable are called autonomous.
= f (x)
That is, autonomous equations are those of the form x0 = dx
dt
dy
0
(The independent variable is t)or y = dx = f (y) (The independent
variable is x). To draw a phase diagram of x0 = f (x), you would

5. PHASE AND VECTOR FIELD DIAGRAMS

17

Step one: Set f (x) = 0 solve the equation. The solution is called the
critical number of the equation x0 = f (x).
Step two: Plot the solution one a horizontal (or vertical)number line.
The solutions will divide the line into intervals (segments),
pick any number from each interval (segment) and determine
the sign of the value of f (x) at the picked value.
Step three: Using the sign in Step two to draw arrow on each segment, a
right (up) arrow for positive sign and a left (down) arrow for
negative sign.
Example 5.1. Draw phase diagram for x0 = (x 1)(x + 2)(x 3).
Solution
Step one: Here f (x) = (x 1)(x + 2)(x 3) so f (x) = 0 gives
(x 1)(x + 2)(x 3) = 0,
and it has three solutions, x = 2, 1, 3
Step two: Plot the three solution on the number line,

Figure 9. Number Line

we see that we have four intervals: (-, -2), (-2, 1), (1, 3), and (3, ).
The following table organizes our computation,
Interval
(-, -2)
(-2, 1)
(1, 3)
(3, )
x-value picked
-3
0
2
4
Value of f(x)= f(-3)=(-3-1)
(x-1)(x+2)(x-3) (-3+2)(-3-3)
=-24
Sign of f (x)
-

f(0)=(0-1)(0+2) f(2)=(2-1)(2+2) f(4)=(4-1)(4+2)

(0-3)=6
(2-3)=-4
(4-3)=24

+
+
Step three: Draw the arrows on the number line. A right arrow for
positive sign and a left arrow for negative sign.
a

Figure 10. Phase Diagram

If c is a critical number of x0 = f (x), i.e. f (c) = 0,
c is a sink if arrows on both side of it point toward it.
c is a source if arrows on both side of it point away from it.
c is a saddle if one arrow points away from and another points
toward it.

18 1. EXPLICITLY SOLVABLE FIRST ORDER DIFFERENTIAL EQUATIONS

In Example ??, c = 1 is a sink, c = 2, 3 are source.

From phase diagram and the sign of derivative of x00 = dfdt(x) =
df (x) dx
(x)
= f (x) dfdx
, we can get an very good picture about the behavior
dx dt
of the solutions.
Example 5.2. For x0 = f (x) = (x 1)(x + 2)(x + 3), we have
(x)
f (x) dfdx
= (x 1)(x + 2)(x + 3)(3x2 4x 5), which has the following
sign distribution,

2
(a) d dtx(t)
> 0 on intervals, (2, 23 13 19), (1, 23 + 13 19), and
2
(3, ).

2
< 0 on intervals, (, 2), ( 32 31 19, 1), and ( 23 +
(b) d dtx(t)
2

1
19, 3).
3
With this information and the phase diagram we have the following
picture of typical solutions for x0 = (x 1)(x + 2)(x 3),

Figure 11. Typical solutions for x0 = (x 1)(x + 2)(x + 3)

5.2. Vector field diagram. We know that the derivative x0 (t0 )
of x(t) at t0 is the slope of the tangent line to the graph of x(t) at
the point (t0 , x(t0 )), and the equation of the tangent line is y x(t0 ) =
x0 (t0 )(tt0 ). The idea of vector field for an ODE x0 = f (t, x) is to draw
a segment of tangent line to many chosen points (t0 , x0 ) with slope
x0 (t0 ) = f (t0 , x0 ). When we draw enough such line segments, patterns
shall emerge and we shall have a clear idea(with some training) of the
behavior of solutions to the equation. The following diagram shows the
vector field of x0 = sin(t). Clearly we see the wave pattern of typical
solution x(t) = cos(t) + C.
Here two diagrams are shown, the right one is provided for novel
eyes, which has solutions curves superimposed on the vector field. For
complicated vector field only trained eyes will see the subtle patterns

5. PHASE AND VECTOR FIELD DIAGRAMS

19

Figure 12. Vector field for f (t, x) = sin(t)

of the dynamics of solutions as shown in the next not so complicated
diagram,

Figure 13. Vector field for f (t, x) = t x2

To graph the vector field diagram for y 0 = f (x, y) by hand over
rectangular region [a, b] [c, d], you would do the following:
(1) Divide interval [a, b] into N equal length subintervals and [c,
d] into M equal length subintervals, for [a, b] the length is
and the subintervals are [a, a+h], [a+h, a+2h], ;
h = ba
N
for [c, d] the subinterval length is s = dc
and the subintervals
M
are [c, c+s], [c+s, c+2s], .
(2) If we use xi = a + ih, i = 0, 1, , N to denote the endings
of subintervals of [a, b] and yj = c + js, j = 0, 1, 2, to
denote the endings of subintervals of [c, d], then (xi , yj ) form
the grid points, i = 0, 1, , N ; j = 0, 1, , M. For each
point (xi , yj ) compute the slope mij = f (xi , yj ) and, from the

20 1. EXPLICITLY SOLVABLE FIRST ORDER DIFFERENTIAL EQUATIONS

equation of tangent line y yj = mij (x xi ) we get another

point ((xi + h2 , yj + mij h2 ).
(3) Draw line segment from (xi , yj ) to ((xi + h, yj + mij h) for i =
0, 1, , N ; j = 0, 1, , M.
Example 5.3. Suppose we want to graph vector field for y 0 = x
over [0,1][0,1] and divide [0, 1] into two subinterval, h = s = 0.5.
Then we will have nine grid points (0, 0), (0, 0.5), (0, 1), (0.5, 0),
(0.5, 0.5), (0.5, 1), (1, 0), (1, 0.5), and (1, 1). For the grid point
(0.5, 0.5) we have slop m = 0.5 as the function f (x, y) = x. So we can
get another point (0.5 + h2 , 0.5 + m h2 ) = (0.75, 0.625), connect (0.5, 0.5)
to (0.75, 0.625) you get one line segment, continuous this we get the
following diagram.

Figure 14. Vector field for f (x, y) = x

So it is obviously tedious to draw the vector field by hand.
To graph vector field in Mathcad is not an easy task. This is especially
true when you want to superimpose solution curve over the graph. You
need to write some code to instruct how you want Mathcad to graph the
vector field for what function. you can get the code from the website
www.unf.edu/mzhan.
Project
At beginning you should enter: Project title, your name, ss#, and
due date in the following format
Project One: Define and Graph Functions
John Doe
SS# 000-00-0000
Due: Mon. Nov. 23rd, 2003
You should format the text region so that the color of text is different
than math expression. You can choose color for text from Format
>Style select normal and click modify, then change the settings for

5. PHASE AND VECTOR FIELD DIAGRAMS

21

font. You can do this for headings etc. Graph Vector field
Goal: Familiar your self with an many different kind of vector fields,
defined by x0 = f (t, x), as possible and identify many general features
of solutions, especially the following features,
Does the equation has constant solutions? A constant solution is shown in vector field as horizontal line segments since
constant solution has slope equals zero.
Does solutions converge to an particular solution in the lang
run?
Do solutions grow to infinite an a finite time? Her vertical bars
indicate a solution might become unbounded in short time.
Do solutions display any periodicity? We know sin(t) has period of 2 and solution curve repeats itself. Here large interval
for t might be needed.
You can use the following functions
f (t, x) = sin(x)
f (t, x) = cos(t)
f (t, x) = x(x 1)
f (t, x) = et x + sin(x)
f (t, x) = x + xsin(t)
You should use the provide Mathcad file at the website
www.unf.edu/mzhan/vectorfield.mcd
as a starting point and superimpose solutions to the vector fields so
support your observation.