Dr. Z.

’s Calc4 Lecture 1 Handout: Introducing Differential Equations

By Doron Zeilberger

Section 1: Direction Fields

A general first order differential equation looks like

dy
= f (y, t) ,
dt
where f (t, y) is (usually) a function of both t (usually time), and y (the value of the desired function
at time t, or graphically, the elevation).

It often has an initial condition: y(0) = y0 (or, more generally, y(t0 ) = y0 ), for some number y0 .
we often write y 0 instead of dy
dt .

Here are some examples

y0 = t + y , y(0) = 5 ;
y 0 = ty , y(0) = 5 ;
y 0 = cos(ty) , y(0) = π .

Such a differential equation is a puzzle. For example the differential equation y 0 = y + t means

I am a certain function of t, let’s call me y(t). My derivative (rate of change) at any time t in the
future is exactly equal to t plus my value at that very same time (y(t)), who could I possibly be?
How do I look like?

If you also have the initial condition y(0) = 5, for example, then there is an additional clue.

“In addition, at t = 0, my value is exactly 5”. Who am I?

For quite a few (but by no means for all!) differential equations, people (and Maple!) know how
to find an exact formula for y(t). For example, for the diff.eq. y 0 = y + t, y(0) = 5, you type

dsolve({diff(y(t), t) = y(t) + t, y(0) = 5}, y(t));

and you get immediately

y(t)=-1-t+6 exp(t) ,

and if you want a plot, all you have to do is type:

plot(op(2,%),t);

But for many diff.eqs. there is no way to find a formula for the solution. For these one can do it
graphically, by drawing a direction field.

1
To solve a diff.eq. of the form y 0 = f (t, y) draw a tiny arrow of slope f (t, y) starting at the point
(t, y) for as many points (t.y) as you can. Then starting at (0, y0 ) “follow the arrows” and you
would get the graph of the solution.

If you are not given an initial condition, then there are infinitely many solutions! If you are lucky,
then amongst them would be a very simple solution, a horizontal line, aka as a constant function,
y(t) = c for some constant c. Since for such constant functions y 0 = 0 always, to find out whether
you are lucky, you have to solve the (non-differential!) equation

f (c, t) = 0 ,

and see whether you can solve for c that does not depend on t. This usually happens for the special
case that f (t, y) does not depend on t, i.e. the slope of the function y(t) only depends on the
elevation, but not on the time! Such equations are called autonomous, and their format it

y 0 = F (y)

for some function F (y) of only y, and then the constant solutions are those c for which F (c) = 0.

Such solutions, of the form y(t) = c are called equilibrium solutions, since, if in luck, in “the long
run” all solutions converge to them as time, t, gets larger and larger. Otherwise solutions diverge
from it.

Problem 1.1.1: Draw a direction field for the given differential equations. Based on the direction
field, determine the behavior of y as t → ∞

a: y 0 = 6 − 2y

b: y 0 = 6 + 2y

Solution of 1.1.1.a: Solving 6 − 2y = 0, we get the constant function y = 3. If y > 3 then the
slope is negative so the function is going down. If y < 3 then the slope is positive, so the function
is going up. So in the long-run all solutions get closer and closer to the horizontal line y = 3.

Ans. to 1.1.1.a: y → 3 as t → ∞

Solution of 1.1.1.b: Solving 6 + 2y = 0, we get the constant function y = −3. If y > −3 then the
slope is positive so the function is going up. If y < −3 then the slope is negative, so the function
is going down. So in the long-run all solutions diverge from the horizontal line y = −3.

Ans. to 1.1.1.b: y diverges from −3 as t → ∞

Problem 1.1.2: Draw a direction field for the differential equation

y 0 = (1 − y)(y − 2) .

Based on the direction field, determine the behavior of y as t → ∞. If it depends on the initial
condition, then state the behavior accordingly.

2
Solution of 1.1.2: Using algebra, we solve the algebraic equation (1 − y)(y − 2) = 0. We get the
two solutions y = 1 and y = 2. If y > 2 then the slope is negative so the function is going down.
If 1 < y < 2 then the slope is positive, so the function is going up. So in the long-run all solutions
for which y0 > 1 converge to the line y = 2. If y < 1 (e.g. y = 0) then the slope is negative and
the function is going down to −∞.

Ans. to 1.1.2: For y > 1, y → 2 as t → ∞. For y < 1, y diverges from y = 1.

1.2: What does it mean to be an explicit solution of a DiffEq

Problem 1.2.1: Verify that y(t) = et + e2t is a solution of the initial value differential equation

y 00 − 3y 0 + 2y = 0 , y(0) = 2 , y 0 (0) = 3 .

Solution to 1.2.1: Using calculus,

y(t) = et + e2t , y 0 (t) = et + 2e2t , y 00 (t) = et + 4e2t .

Plugging-in,

y 00 (t) − 3y 0 (t) + 2y(t) = et + 4e2t − 3(et + 2e2t ) + 2(et + e2t ) = et (1 − 3 + 2) + e2t (4 − 6 + 2) = 0 .

So the diff.eq. is OK. Now,

y(0) = e0 + e0 = 1 + 1 = 2 , y 0 (0) = e0 + 2e0 = 1 + 2 = 3 .

So the initial conditions are OK.

Problem 1.2.2: Verify that y(t) = t2 is a solution of the initial value differential equation

y 0 (t)3 − ty(t) = 7t3 , y(1) = 1 .

Solution to 1.2.2: Using calculus

y(t) = t2 , y 0 (t) = 2t ,

Using algebra
(y 0 (t))3 − t y(t) = (2t)3 − t(t2 ) = 8t3 − t3 = 7t3 .

So the diff.eq. is OK. Now

y(1) = 12 = 1 ,

so the initial condition is OK too.

3
1.3: Classification of Differential Equations

If the function we are looking for, y(t), and all its derivatives, y 0 (t), y 00 (t), . . . that show up, are all
by themselves, (i.e. not raised to a power, or inside a function!,or multiplied by each other), then
the diff.eq. is linear. The highest derivative that shows up is the order.

The format of a homogeneous linear diff.eq. or order r is

fr (t)y (r) (t) + fr−1 (t)y (r−1) (t) + . . . + f0 (t)y(t) = 0,

where fr (t), fr−1 (t), . . . f0 (t) are functions of t. The format of an inhomogeneous linear diff.eq. or
order r is
fr (t)y (r) (t) + fr−1 (t)y (r−1) (t) + . . . + f0 (t)y(t) = R(t) ,

where R(t) is yet another function of t.

If a diff.eq. does not have such a format, e.g. y 0 (t)y(t) + t = 5 or y 0 (t) + y(t)2 = 0 then it is
non-linear.

Problem 1.3.1: For each of of the following diff.eq. state whether there are linear or non-linear,
and find the order.

a: y 00 (t) + 6y 0 (t) + y(t) = 7 ,

b: y 000 (t) + (sin t) y 0 (t) + y(t) = 7 ,

c: y 00 (t)2 + y 0 (t) + y(t) = 0 ,

d: y (4) (t) + y 0 (t)y(t) = 6 .

Solution of 1.3.1: In (a) the highest derivative that shows up is the second, so it is a second-order
diff.eq. Since each of y(t), y 0 (t), y 00 (t) are by themselves, this diff.eq. is linear.

In (b) the highest derivative that shows up is the third (y 000 (t)), so it is a third-order diff.eq. Since
each of y(t), y 0 (t), y 00 (t) are by themselves it is linear. Note that even though the coeff. of y 0 (t) is
not a number (it is a function, sin t), the diff.eq. is still linear (but it is not constant-coefficients).

In (c) the highest derivative that shows up is the second (y 00 (t)), so it is a second-order diff.eq.
Since y 00 (t) is squared, it is non-linear.

In (d) the highest derivative that shows up is the fourth (y (4) (t)), so it is a fourth-order diff.eq.
Since y(t) and y 0 (t) are multiplied by each other, it is non-linear.

Ans. to 1.3.1: (a) is a second-order linear diff.eq; (b) is a third-order linear diff.eq. ; (c) is a
second-order non-linear diff.eq. ; (d) is a fourth-order non-linear diff.eq.