Lecture 05, Math 441

Aldo Manfroi

University of Illinois at Urbana-Champaign

Substitutions Methods (cont.)
y (x)
y ′ = F ( yx ) Try v (x) = x → separate

y2 y2
Consider x y ′ = x + y and rewrite as y ′ = x2
+ y
x → y ′ − x1 y = 1 2
x2
y

This is an example of a Bernoulli equation. Use the substitution

v (x) = y 1−2 = y1 :

y = v1 ; y ′ = − v12 v ′ and substituting in the equation:

1 ′ 1 1 1 1
− 2
v − = 2 2 ; → v′ + v = − 2
v xv x v x x

which is linear and can be solved with an integrating factor. timestamp: 00:00
 1
The integrating factor is µ(x) = e∫ x
dx
= eln(x) = x

Multiplying the equation for v by µ we get:

1 1
x v′ + v = − → (x v )′ = − → x v = − ln (x) + A →
x x

A − ln (x) x
v (x) = → y (x) =
x A − ln (x)

timestamp: 04:30
Bernoulli Equations

In general, a Bernoulli equation has the form

y ′ + p(x) y = q(x) y α

where α ≠ 0, 1 (If α is 0 or 1, this would be a linear equation)

We can use the substitution v (x) = y 1−α and the equation becomes

1st order linear for v (x) and we can use an integrating factor.

timestamp: 06:44
 √
Ex: y ′ = y → y ′ − x1 y = y 2
1
y
x +

and we use v (x) = y 1− 2 = y 2 ; y = v 2 ; y ′ = 2v v ′

1 1
So α = 1
2

Using this substitution to get

1 1 1
2v v ′ − v 2 = v → v ′ − v =
x 2x 2

(assuming v ≠ 0. If v = 0 then y = 0 and you can verify this is

another solution)

timestamp: 08:00
The integrating factor is µ(x) = e− ∫ = e− 2 ln(x) = x − 2 =
1 1 1
2x
dx √1 and
x

use this to get:

′
v′ 1 1 1 1 1 √
√ − √ v = √ → (√ v) = √ → √ v = x + A
x 2x x 2 x x 2 x x

√ √ √ 2
v = x + A x = y → y (x) = (x + A x)

to which we need to add the singular solution y (x) = 0 .

timestamp: 11:13
Theorem of Existence and Uniqueness for 1st Order Linear
IVP

Consider the 1st order linear initial value problem

⎧
⎪
⎪
⎪
⎪
⎪
⎪ y ′ + p(x)y = q(x)
⎪
⎨
⎪
⎪
⎪
⎪
⎪
⎪ y (a) = b
⎪
⎩

Theorem: If ∃ an interval I around x = a where p and q are

continuous, then ∃ one and only one solution to the IVP over all of

I . (∃ means ”it exists” or ”there is”)

timestamp: 14:20
Note: the interval I can be determined without solving the

equation, it depends only on the interval of continuity of p and q.

So we can a priori know where the solution is guaranteed to exist

and be unique.

timestamp: 16:20
y

( ) x
a
I
timestamp: 17:00
We can prove this by just finding the solution of the IVP using the

method of integrating factor.

µ(x) = e∫ pdx

µ y ′ + p µ y = µ q → (µ y )′ = µ q → µ y = ∫ µqds + A

1
y (x) = (∫ µqds + A) THE solution to the IVP
µ

All these integrals can be done where p and q are continuous. A

can be determined uniquely from the initial condition. timestamp: 17:50

Ex: Consider the IVP: x 2 y ′ − y = 3; y (1) = 0

First rewrite the equation in the form y ′ − 1

x2
y = 3
x2

So p(x) = 1
x2
and q(x) = 3
x2

x
0 1

They both have a single discontinuity at x = 0. The IC is given at

x = 1, so the interval I where a unique solution is guaranteed to

exist is (0, +∞)

timestamp: 22:50
Let’s find the solution. Use the IF µ = e− ∫
1 1
dx
x2 = ex :

1 1 3 1 1 ′ 3 1
y ′e x −
1 1 1
e xy = e x → (y e x ) = e x → e x y = −3e x + A
x2 x2 x2

y (x) = A e− x − 3
1
(general solution)

Applying the initial condition: y (1) = A e−1 − 3 = 0 → A = 3 e

So THE unique solution to the IVP is y (x) = 3 e(1− x ) − 3

1

timestamp: 26:08
Theorem of Existence and Uniqueness for 1st Order IVP

Given the 1st order IVP

⎧
⎪
⎪
⎪
⎪
⎪
⎪ y ′ = f (x, y ) (where f may be nonlinear)
⎪
⎨
⎪
⎪
⎪
⎪
⎪
⎪ y (a) = b
⎪
⎩

Theorem: If ∃ a simply connected region R in the x, y plane that

includes (a, b) and if f (x, y ) is continuous in R then ∃ a solution

∂f (x,y )
in an interval around x = a. And if ∂y is also continuous in R

then the solution is unique. timestamp: 32:00

Note: the solution is guaranteed to exist somewhere around the

initial condition, but the theorem does not say how far it can go

(as opposed to the linear case).

Also, if the conditions are not met, a solution may or may not

exist, may or may not be unique. It is not an if and only if result.

timestamp: 35:00
 y

f is continuous Ô⇒ existence
fy is continuous Ô⇒ uniqueness

( ) x
a

timestamp: 35:22
Ex: dy
dx = − yx

Here f (x, y ) = − yx ; ∂f
∂y = − x1

They are both continuous for x ≠ 0, so if the initial condition is

given anywhere but at x = 0 the theorem guarantees one and only

one solution near the initial condition.

If the initial condition is given at x = 0 then the theorem doesn’t

help.

timestamp: 38:50
y

( ) x
a

timestamp: 39:40
Let’s solve this separable equation:

dy dx C
=− → ln (y ) = − ln (x) + A → y =
y x x

which exists as long as x ≠ 0 and C can be determined if the IC is

not at x = 0.

timestamp: 40:55