00 Differential Equations Notes

Unauthorized distribution of any kind is prohibited.

Summary

A Differential Equation (DE) is an equation with a function and one or

more of its derivative
An Ordinary Differential Equation (ODE) is an equation that involves an
unknown function of a single variable, its independent variable, and one
or more of its derivatives
A Partial Differential Equation (PDE) is an equation that involves an
unknown function of a single variable, its independent variable, and one
or more of its derivatives

Classification
The degree is the exponent of the highest derivative
The order is the highest order derivative present in the equation
4
d b db
5 10 7 5
5 ⋅ ( ) + 7 ⋅ ( ) + b − b = p
4
dp dp

This has a degree of 5 and an order of 4

Linear and Homogenous differential equations

An nth order ODE is called linear if it has the format:

n n−1 ′
an(x)y + an−1(x)y + ⋯ + a1(x)y + a0(x)y = F (x)

An equation is called homogeneous if each term contains the function

or one of its derivatives.
′
y + y = 0 is a 1st order linear homogenous DE
If there are no independent variables by themselves, it's homogenous
′
y + y = e
x
is nonhomogenous
In y ′′ ′
+ ay + by = f (t) , f (t) is the forcing term

Systems of Differential Equations

dx
= −3x + y
dt
xy
= x − 3y
dt

Note

Any DE of order 2 or higher will be handled as a system

Problem 1
Often scientists use rate of change equations in their study of population growth
for one or more species. In this problem we study systems of rate of change
equations designed to inform us about the future populations for two species
that are either competitive or cooperative

1. Which system of rate of change equations below describes a situation

where the two species compete and which describe cooperative species?

dx/dt = −5x + 2xy

1.
dy/dt = −4y + 3xy

1. this is cooperative, since they go to zero if you remove the term of

the other variable

dx
= 4x − 2xy

2. dt
dy
= 2y − xy
dt

1. this is competitive, since removing either of the xy terms makes

and
dx

dt
go to infinity
dy

dt
ODE Classification
y
′′ ′
is 2nd order, linear, homogenous ODE (variable coefficients)
− 2xy + 2y = 0

y(2 − 3x) + x(3y − 1)y = 0 is 1st order nonlinear homogenous ODE

′

′
y(1 + (y ) ) = c
2
is 1st order nonlinear nonhomogenous ODE (2nd degree)

General Form DE
′ ′′ (n)
F (x, y, y , y , … , y ) = 0

Ex.
2 ′′ ′
x y + 3xy − y + 3x = 0

Normal Form DE
(n) ′ ′′ (n−1)
y = f (x, y, y , y , … , y )

Ex.
′′ ′ 7x
y = −y + 4xy + e

Another way to write 1st order DEs:

dy M (x, y)
′
y = = −
dx N (x, y)

or

M (x, y)dx + N (x, y)dy = 0

Solutions of DE
′
y = y

Solutions: y = e , y = 0 (trivial solution)

x

General solution: y = ce x

This is a one-parameter family of solutions

′′
y = 0

General solution: y ′
= c1 ⇒ y = c1x + c2

This is a two-parameter family of solutions

n
y = 0

Solution: c n+1x
n
+ cnx
n−1
+ ⋯ + c2x + c1

This is a n-parameter family of solutions

ϕ(x) = x
2
− x
−1
(f or x != 0) is a solution of y ′′ −2
− 2x y = 0

ϕ(x) = 2x
3
is a solution of xy ′
= y
1
′
y − xy 2 = 0

is the general solution (one-parameter family of

2
x 2
y = ( + C)
4

solutions)
4

y =
x

16
is a particular solution
y = 0 is a trivial solution and a singular solution (no value of C would
return this solution)
′
y(2 − 3x)dx + x(3y − 1)dy = 0 ⇒ y(2 − 3x) + x(3y − 1)y = 0

y = 0 is the trivial solution

! (y )′ 2
+ 1 = 0 has no real solutions

Initial Value Problems (IVP)

for an nth order problem, we need n initial values

Show that y = x +c
1
2
is a one parameter family of solutions of the DE
′
y + 2xy
2
= 0 w/ initial condition y(0) = 1

y = 0 (trivial, singular)
y = 2
x +2
1
(particular solution)

An IVP w/ more than one solution

4 1

Both y = 0 and y = of the IVP y x

16
′
= xy 2 ; y(0) = 0
⇒ When do solutions exist?
⇒ When are the solutions unique?

Picard's Theorem (not important)

Answers when solutions will exist or be unique
y = f (x, y) (normal form): both f (x, y) and (x, y) must be continuous
′ ∂f

∂y
Solving DEs
Analytical Methods (for 1st order ODE)
Separable
Exact
Exact w/ I.F.
Linear
Substitution
Euler homogeneous
Bernoulli
G(ax+by)
? linear combinations

Separable

Note

a first order DE is separable if y ′

= g(x) ⋅ f (y)

!y ′
= 2 − x y
2
is not separable
y =
′ t

y
2
is separable
y =
′ x−2xy

y2−3
is separable

Find the general solution of

dy 2
= y
dt

dy
∫ 2
= ∫ dt
y

1 1
− = t ⇒ y = −
y t+c

y(t) = −
1

t+c
when t ≠ c, 0 when t = c
Example

Solve the IVP:

y − 1
′
y = ; y(−1) = 2
2 − x

1 1
∫ dy = ∫ dx
y−1 2−x

ln|y − 1| = −ln|2 − x| + c

|y − 1| =
c1

|2−x|
where c1 = e
c
> 0

y − 1 =
±c1

|2−x|
⇒y−1= c2

|2−x|
where c 2 = ±c1

y = 1 is a constant solution (not singular)

Exact Differential Equations

2 2
dy x + y
=
dx x − y
2 2
(x + y )dx + (y − x)dy = 0

Total differential

dF (x, y) dF (x, y)
dF (x, y) = dx + dy
dx dy

Ex.
2 3
F (x, y) = xy + 2x y

dF 2 2
= y + 6x y
dx

dF 3
= 2xy + 2x
dy

2 3
=> dF = (y + 6x y)dx + (2xy + 2x )dy

Definition

The expression M(x,y)dx+N(x,y)dy is called an exact differential in a domain

D if ∃ a function F of two real variables ∈ dF (x,y)

dx
= M (x, y) and
dF (x,y)
= N (x, y)
dy
Ex.
2
y dx + 2xydy = 0

Is this an exact differential?

-$M y = 2y = 2y = Nx , so yes
2
F (x, y) = xy

∂F

∂x
= y
2
, ∂F

∂y
= 2xy

Ex.
Is this exact?
ydx + 2xdy = 0

Not exact
!M y = 1 ≠ 2 = Nx

Theorem

Consider the DE M (x, y)dx + N (x, y)dy = 0

M, N are continuous first partial derivatives at all (x, y) ∈ D

1. If the D.E. is exact in D, then

∂M (x,y) ∂N (x,y)
=
dy dx

2. Conversely if , then the D.E. is exact

∂M (x,y) ∂N (x,y)
=
dy dx

Ex.
3 x 2 2 x
(2x sin y + y e )dx + (x cos y + 3y e )dy = 0

M (x, y) N (x, y)

2 x 2 x
My = 2x cos y + 3y e = Nx = 2x cos y + 3y e

This is exact

Ex.
ydx = xdy ⇒ ydx − xdy = 0

! No. −1 ≠ 1

How to solve exact differentials

Theorem

Suppose the DE M (x, y)dx + N (x, y)dy = 0 satisfies the previous theorem and
is exact in domain D. Then a one parameter family of solutions of this DE is
given by F (x, y) = c where F is a function ∈ F x = M and F y = N

ex.
2 2
(3x + 4xy)dx + (2x + 2y)dy = 0

1. Check for exactness

My = 4x = 4x = Nx

2
M (x, y) = 3x + 4xy

2 3 2
∫ M (x, y)dx = ∫ (3x + 4xy)dx = x + 2x y + g(y)

2 ′ 2
N (x, y) = Fy = 2x + g (y) = 2x + 2y

⇒ g (y) = 2y => g(y) = y

′ 2

Our solution: x 3
+ 2x y + y
2 2
= c

Method of grouping (shortcut)

2 2
(3x + 4xy)dx + (2x + 2y)dy = 0

⇒ 3x dx + 4xydx + 2x dyU + 2ydy = 0 (distribute differentials)

2 2

⇒ d(x ) + d(2x y) + d(y ) = d(c)

3 2 2

⇒x 3 2
+ 2x y + y
2
= c

ex.
(2x cos y + 3x y)dx + (x
2 3
− x
2
sin y − y)dyC = 0 , y(0) = 2


⇒
2x cos ydx − x sin ydy + 3x ydx + x dy − ydy = 0

2 2 3







       
2 3 2 d(c)
d(x cos y) d(x y) y
−d( )
2
Integrating factors

Question

What happens if ∂M

∂y
≠
∂N

∂x
?

Ex:
ydx + 2xdy = 0 (not exact)
if we multiply by y:

2
y dx + 2xydy = 0 (exact)

Since this equation is integrable, we call y the integrating factor

If the D.E. M (x, y)dx + N (x, y)dy = 0 is not exact but

μ(x, y)M (x, y)dx + μ(x, y)N (x, y)dy = 0 is exact, then μ(x, y) is the integrating

factor

Every first-order DE can be solved using an integrating factor (allegedly)

Example

2
dy 3x + xy
=
2
dx y + x y

2 2
(y + x y)dy = (3x + xy )dx

My = 2xy ≠ Nx = −2xy (not exact)

As it turns out,
1
μ(x, y) = 2 2
(3+y )(1+x )

multiplying everything by μ(x, y)

x y
2
dx − 2
dy = 0
1+x 3+y

We can now integrate

Question

How do we find the integrating factor?

If M dx + N dy = 0 is not exact/separable, suppose μM dx + μN dy = 0 is exact

⇒The equation is exact if and only if (μM ) y = (μN )x . (the subscripts denote
partial derivatives)

(μM )y = (μN )x (9)

By the product rule of differentiation:

μMy + μyM = μNx + μxN

or

μxN − μyM = (My − Nx)μ

Although M , N , M , and N are known functions of x and y, the difficulty here in

y x

determining μ(x, y) from (9) is that we must solve a partial differential equation.
Instead, we will make a simplifying assumption. Suppose μ is a function of one
variable; for example, say that μ depends only on x. In this case, μ = and x
dμ

dx

μy = 0 , so (9) can be written as

dμ My − Nx
= μ (10)
dx N
We still have a problem if the quotient μ depends on both x and y.
dμ My−Nx
=
dx N

However, if after all obvious algebraic simplifications are made, the quotient
turns out to depend solely on the variable x, then (10) is a first-
dμ My−Nx
= μ
dx N

order ordinary differential equation. We can finally determine μ because (10)

is separable and linear. It follows that μ(x) = e . Similarly, it follows
∫ ((My−Nx)/N ) dx

from (9) that if μ depends only on the variable y, then

dμ Nx − My
= μ (11)
dy M

In this case (like the case above), if is a function of y only, the we can
Nx−My

solve (11) for μ:

1 Nx − My
1. ∫ dμ = ∫ dy
μ M

Nx − My
2. ln |μ| = ∫ dy
M
Nx−My
∫ dy
3. μ = e M

Finding the integrating factor

For the differential equation

M (x, y)dx + N (x, y)dy = 0. (12)

If is a function of x alone, then an integrating factor for (12) is

My−Nx

My−Nx
∫( ) dx
μ(x) = e N
(13)

If is a function of y alone, then an integrating factor for (12) is

Nx−My

Nx−My
∫( ) dy
μ(x) = e M
(14)
Example

ydx + (3 + 3x − y)dy = 0

My = 1 ≠ Nx = 3 (not exact)
using (11),
3−1 2
=
y y

using (14),
∫ 2/y dy 2 ln y 2
μ = e = e = y

multiply by y 2

3 2 2 3
y dx + (3y + 3xy − y )dy = 0

2 2
My = 3y = 3y = Nx

3 3
F (x, y) = ∫ M (x, y) dx = ∫ y dx = xy + g(y)

2 ′ 2 2
Fy = 3xy + g (y) = Nx = 3y + 3xy

′ 2 3
g (y) = 3y − y

3 1 4
g(y) = y − y
4

2 3 1 4
3xy + y − y
4


  
F

Example

2 2
(2x + y)dx + (x y − x)dy = 0

My = 1 ≠ 2xy − 1 = Nx

Nx−Mx 2−2xy −2(xy−1) −2

= 2
= =
M x y−x x(xy−1) x

∫ −2/x dx −2 ln x 1
μ = e = e = 2
x

multiply by μ
y 1
(2 + 2
)dx + (y − )dy = 0
x x

1 1
2
= My = 2
= Nx
x x
y 1
2dx + 2
dx − dy + ydy = c


 


x x
 

   
d(2x) y y
2
d(− ) d( )
x 2

2
y y
F = 2x + − + c
2 x
Example

−1
(2x + yx )dx + (xy − 1)dy = 0

−1
My = x ≠ Nx = y
−1
My−Nx x −y 1−xy 1
= = = −
N xy−1 x(xy−1) x
−1
Nx−My y−x
= −1
M 2x+yx

∫ −1/x dx − ln x 1
μ = e = e =
x

multiply by μ
−2 1
(2 + yx )dx + (y − )dy = 0
x

−2 1
2dx + yx dx − dy + ydy = 0


 


x


   
d(2x) −y y
2
d( )
x 2

Linear
This is actually important to remember

General form of a linear DE

′
y + p(x)y = q(x)

dy
+ p(x)y − q(x) = 0
dx

(p(x)y − q(x))dx + dy = 0

My = p(x)

Nx = 0 (not exact)

My − Nx p(x) − 0
=
N 1
∫ p(x) dx
μ = e

∫ p(x) dx ′ ∫ p(x) dx ∫ p(x) dx

e y + p(x)e y = q(x)e


  
d ∫ p(x) dx ∫ p(x) dx
(e y)=∫ e q(x) dx
dx
Example

y
′
y = + 2x + 1
x

1
′
y − y = 2x + 1
x
1
∫ −1/x dx − ln |x|
μ = e = e =
x
1 ′
1 1
y = 2 + y −
2
x x x




     
d 1 1
∫ ( y) ∫ 2+ dx
dx x x

1 1
y = ∫ 2 + dx
x x

1
y = 2x + ln |x| + c
x
2
y = 2x + x ln |x| + cx
Example

′ −x
y + 4y − e = 0

4
y(0) =
3

′ −x
y + 4y = e

∫ 4 dx 4x
μ = e = e

4x ′ 4x 3x
e y + 4e y = e

d
4x 3x
∫ (e y ) dx = ∫ e dx

 dx  
μy

3x
e
4x
e y = + c


  3
μy

note: you can generally skip to μy = ∫ μq(x) dx

1
−4x
y = + ce
x
3e
4
y(0) =
3
4 1
= + c
3 3

c = 1

Bernoulli Equations
Used for things in the form:
′ n
y + P (x)y = Q(x)y

where n ≠ 0, 1
Solving Bernoulli Equations
dy n
+ P (x)y = Q(x)y
dx

let u = y 1−n

−n
du = (1 − n)y

du −n
= y
1−n

Multiplying by y : −n

−n ′ 1−n
y y + P (x)y = Q(x)

Substituting in u and y , −n

1 ′
u + P (x)u = Q(x)
1−n
Example

′ 3
y + y = xy

−3 ′ −2
y y + y = x

 
u

−2
u = y

du
−3 ′
= −2y y
dx
1
′ −3 ′
− u = y y
2

Now we substitute:

−3 ′ −2
y y + y = x




   
1 ′ u
− u
2

′
u − 2u = −2x

The dif f erential equation is now linear

−2x
μ = e
′ −2x −2x
u − 2e u = −2xe

1 1
−2x −2x −2x −2x
e u = −2 ∫ xe dx = −2 ( xe − e ) + c
2 4

1
2x
u = x + +
2
1
−2 2x
y = x + + ce
2
1
y =
1 2x
√x + + ce
2

Euler homogeneous
An equation F (x, y) is Euler Homogeneous if it can be written in the form:

F (v)
n
F (tx, ty) = t F (x, y)
ex.
2 2
F (x, y) = x − y

2 2 2 2 2 2
F (tx, ty) = (tx) − (ty) = t (x − y ) = t F (x, y)

for example,
2 2
x −y
F (x, y) =
xy
2 2 2

'
t (x −y )
F (tx, ty) = 2
2t xy

Example

2 2
x + y
′
y =
xy
2
divide numerator and denominator by x

y 2
1 + ( )
′ x
y =
y

y
let v =
x

xv = y
′ ′
v + xv = y

substitute it in:
2
1 + v 1
′
v + xv = = + v
v v
1
′
xv =
v
1
′
v =
vx
1
∫ v dv = ∫ dx
x

1
2
v = ln |x| + c
2
2
v = 2 ln |x| + c1

2 2
v = ln x + c1
2 2
v = ln cx
2
y
2 2
( ) = x ln cx
x
Example

2 2
(x − 3y )dx + 2xydy = 0

2 2 ′
(x − 3y ) + 2xyy = 0

2 2
(3y − x )
′
y =
2xy

2
divide numerator and denominator by x

y 2
3( ) − 1
′ x
y =
y
2( )
x

y
let v =
x

xv = y
′ ′
v + xv = y

2
3v − 1
′
v + xv =
2v
′ 2
2v(v + xv ) = 3v − 1
′ 2
2xvv = v − 1
2
v − 1
′
v =
2vx

2v 1
∫ dv = ∫ dx
2
v − 1 x

skipping integration:
2 3 2
y = c1x + x

y = ±√ cx + x = ±x√ cx + 1
3 2

G(ax+by)
Example

′ −1
y = y − x − 1 + (x − y + 2)

v = x − y

y = x − v

′ ′
y = 1 − v

′ ′ −1
y = 1 − v = −v − 1 + (v + 2)

′ −1
v = −v − 2 + (v + 2)
2
dv v +4v+3
=
dx v+2

v+2
∫ 2
dv = ∫ dx
v +4v+3

1 2
x + c = ln |v + 4v + 3| (u-sub)
2

2x 2
e = (v + 2) − 1

2x
(x − y − 3)(x − y + 1) = ce

Review examples

Example

2 2
x + y
′
y =
2 2
x − y
y
v =
x

vx = y
′ ′
vx + v = y
2
1 + v
′
y =
2
1 − v
2 2 3
1 + v 1 + v − v + v
′
vx = − v =
2 2
1 − v 1 − v

unsolvable
Example

−2xy
′
y = , y(0) = 4
2
x + 1

f ind y(1) = 2

this is separable

1 −2x
∫ dy = ∫ dx
2
y x + 1
2
ln |y| = − ln |x + 1| + c
c
y =
2
x + 1
c
4 =
1

c = 4

4
y = = 2
2
1 + 1

Example

1
′
y = sin x (y − 2)
cos x
′
y = y tan(x) − 2 sin x
′
y − y tan(x) = −2 sin x

∫ − tan x dx
μ = e

μ = cos(x)
′
cos(x)y − y sin(x) = −2 sin x cos x

1
y cos(x) = cos(2x) + c
2

1
y = sec(x) ( cos(2x) + c)
2