Linear Differential Equations

In this lecture we discuss the methods

of solving first order linear differential
equations and reduction of order.
Linear Equations
A linear first order equation is an equation
that can be expressed in the form
1 0
( ) ( ) ( ), (1)
dy
a x a x y b x
dx
+ =
where a
1
(x), a
0
(x), and b(x) depend only
on the independent variable x, not on y.
We assume that the function a
1
(x), a
0
(x),
and b(x) are continuous on an interval
and a
1
(x) = 0 on that interval. Then, on
dividing by a
1
(x), we can rewrite equation
(1) in the standard form
( ) ( ) (2)
dy
P x y Q x
dx
+ =
where P(x), Q(x) are continuous functions
on the interval.
Lets express equation (2) in the differential
form
| |
( ) ( ) 0 (3) P x y Q x dx dy + =
If we test this equation for exactness, we find
( ) 0
M N
P x and
y x
c c
= =
c c
Consequently, equation(3) is exact only
when P(x) = 0. It turns out that an
integrating factor , which depends only
on x, can easily obtained the general
solution of (3).
Multiply (3) by a function (x) and try to
determine (x) so that the resulting equation
| |
( ) ( ) ( ) ( ) ( ) 0 (4) x P x y x Q x dx x dy + =
is exact.
( ) ( ) ( )
M N d
x P x and x
y x dx

c c
= =
c c
We see that (4) is exact if satisfies the DE
( ) ( ) ( ) (5)
d
x P x x
dx

=
( )
( ) exp ( ) (6) x P x dx =
}
Which is our desired IF
In (2), we multiply by (x) defined in (6) to obtain
( ) ( ) ( ) ( ) ( ) (7)
dy
x P x x y x Q x
dx
+ =
We know from (5)
( ) ( )
d
P x x
dx

=
and so (7) can be written in the form
( )
( ) ( ) ( ) (8)
( ) ( ) ( ) ( )
d
x y x Q x
dx
dy d
x x y x Q x
dx dx

=
+ =
Integrating (8) w.r.t. x gives
( ) ( ) ( ) x y x Q x dx C = +
}
and solving for y yields
( ) ( )( )
exp ( ) ( ) ( ) y P x dx x Q x dx C = +
} }
Working Rule to solve a LDE:
1. Write the equation in the standard form
( ) ( )
dy
P x y Q x
dx
+ =
2. Calculate the IF (x) by the formula
( )
( ) exp ( ) x P x dx =
}
3. Multiply the equation in standard form
by (x) and recalling that the LHS is just
( )
( ) ,
d
x y
dx

obtain
( )
( ) ( ) ( )
d
x y x Q x
dx
=
4. Integrate the last equation and solve
for y by dividing by (x).
Example 1. Solve
( ) 1 cos sin cos = + +
|
.
|

\
|
x x x y
dx
dy
x x
Dividing by x cos x, throughout, we get
x
x
y
x
x
dx
dy sec 1
tan =
|
.
|

\
|
+ +
1
tan
( )
( )
x dx
P x dx
x
x e e
| |
+
}
|
\ . }
= =
lnsec ln lnsec ln
( ) sec
x x x x
x e e e x x
+
= = =
( )
sec
sec sec
d x
dx x
y x x x x =
2
sec sec tan yx x xdx C x C = + = +
}
Multiply by
sec x x yields
Integrate both side we get
Problem (2g p. 62): Find the general
solution of the equation
. 0 ' cot = + + xy x xy x y
Ans.: . cos sin sin c x x x x xy + =
The usual notation implies that x is
independent variable & y the dependent
variable. Sometimes it is helpful to replace
x by y and y by x & work on the resulting
equation.
* When diff equation is of the form
dx
dy
( ) ( ) y Q x y P
dy
dx
= + .
( )
&
P y dy
IF e solution is x IF Q IF dy c
}
= = +
}
Q. 4 (b) Solve
2 y
dx
y x y e
dy
=> = +
y
e y x
y dy
dx
=
1
2
' '
y
y xy y y e =
y
e e IF
y
dy
y
1
ln
1
= =
}
=

1 1
y y
x ye dy e dy
y y
= =
} }
c e
y
x
y
+ = =>
cy e y x
y
+ = =>
A first order equation that can be written in
the form
is known as Bernoullis equation .
It is clearly linear when n=0 or 1.
For n > 1, it can be reduced to a linear
equation by change of variable
( ) ( ) (*)
n
dy
P x y Q x y
dx
+ =
.
1 n
y z

=
Divide equation (*) be y
n
yields
1
( ) ( ) (**)
n n
dy
y P x y Q x
dx

+ =
Taking z = y
1-n
, we find that
(1 )
n
dz dy
n y
dx dx

=
and so (**) becomes
1
( ) ( )
1
dz
P x z Q x
n dx
+ =

Since 1/(1-n) is just a constant,

Example: Solve the DE 3
5
5
2
dy
y xy
dx
=
This is a Bernoulli equation with n =3,
P(x) = -5, and Q(x) = -5x/2
We first divide by y
3
to obtain


3 2
5
5
2
dy
y y x
dx

=
Next we make the substitution z = y
-2

Since dz/dx = -2y
-3
dy/dx, then above
equation reduces to
1 5
5
2 2
dz
z x
dx
=
or
10 5
dz
z x
dx
+ =
which is linear. Hence
10
1
2 20
x
x
z Ce

= +
Substituting z = y
-2
gives the solution
2 10
1
2 20
x
x
y Ce

= +
Reduction of Order:
A general second order DE has the
form
In this section we consider two special
types of second order equations that
can be solved by first order methods.
(1) . 0 ) " , ' , , ( = y y y x F
Type A: Dependent variable missing.
When y is not explicitly present, (1) can
be written as
Then (2) transforms into
If we can solve (3) for p, then (2) can
be solved for y.
) 2 ( . 0 ) " , ' , ( = y y x f
. " and ' Let dx dp y p y = =
) 3 ( . 0 ) , , ( = dx dp p x f
Problem (1 (g) p. 65): Solve
( ) 1 4x y y x =
'
+
' '
The variable y is missing
let &
dp
y p y
dx
' ''
= =
( )
1 , 4
dp
from x p x
dx
+ =
which is linear,
1
4 (2)
dp
p
dx x
+ =
1
log
e
dx
x
x
IF e e x
}
= = =
4 p x xdx c = +
}
c x c
x
px + = + =
2
2
2
2
4
2
dy c
p x
dx x
= = +
2
c
dy x dx
x
(
= +
(

} }
dx
x
c
x dy
}
(

+ =
}
2
2
2
2
ln
2
x
y c x c = + +
2
2
ln y x c x c = + +
Problem (1b p. 65): Solve the DE
. ) (
3
y y y x
'
+
'
=
' '
Ans.:
. ) (
2
1
2
2
x c c y =
Type B: Independent variable missing.
When x is not explicitly present, then
(1) can be written as
Then (4) becomes
If we can solve (5) for p, then (4) can
be solved for y.
) 4 ( . 0 ) " , ' , ( = y y y g
= = = =
dy
dp
p
dx
dy
dy
dp
dx
dp
y p y " then , ' Let
( , , ) 0. (5) g y p pdp dy =
Problem (1(e), P 65): Solve
( ) ( ) 1 1 2
2
y y y
'
+ =
' '
Here x is not explicitly present
p y let =
'
dy
dp
p y =
' '
&
(1) can be writen as,
2
1 2 p
dy
dp
p y + =
} }
=
+

y
dy
p
dp p
2
1
2
( )
( )
2
1
ln 1 ln p c y + =
( )
2
1
ln 1 ln ln p y c + = +
2
1
1 p c y + = 1
1
2
= y c p
1
1
2
=
|
.
|

\
|
y c
dx
dy
1
1
= y c
dx
dy
1
( 1)
dy
dx
c y
=

} }
1 1 2
2 ( 1) c y c x c = +
Problem (2b p. 65): Find the specified
particular solution of the DE
. 0 when 1 and
; ) ' ( "
2
1
2 2
= =
'
=
+
'
=

x y y
y y y yy
Ans.:
. 8 3 2
2
3
x
ye y =