2.

4 EXACT
EQUATIONS
We know how to solve DEs of the form

dy (1)
 f (x)
dx
and
dy f (x)
 (2)
dx g(y)

Let us now consider

dy f (x, y) (3)
 
dx g(x, y)
 Equation (3) can also be written as
dy
g(x, y)  f (x, y) (4)
or dx
dy
M(x, y)  N(x, y) 0 (5)
or dx

M(x, y)dx  N(x, y)dy  0 (6)

How do we find the solution?


Recall from calculus, that if
z  f (x, y)

then
dz f f dy
  
dx x y dx

or, in differential form

 f f
dz  dx  dy
x y
 dz
We know that if  0
dx
or dz  0

z  f (x, y)  c (a constant)
then

So, if there exists a function f such that
f
  M(x, y)
x
and f
 N(x, y)
y

 dz
Then equation (5) can be written as 0
dx

and equation (6) can be written as dz 

, 0

and the solution of the DE is f (x, y) . c

It can be shown that such afunction f, exists if

M N
 
y x
The differential equation (6) [or (4) or (5)] is called
an exact differential equation.

Example 1

Solve
dy
x  2xe  y  6x
x 2

dx


 Solution There exists a function
dy f (x, y) such that
x  2xe x  y  6x 2 f
dx  M(x, y)  y  2xe x  6x 2
x
 x 2

y  2xe  6x dx  xdy  0
f
 and  N(x, y)  x
Let M(x, y)  y  2xe x  6x 2 y
 Let us integrate the first of
and N(x, y)  x these equations
M N   f dx   y  2xe x  6x 2 dx
 
 1, 1 x
y x
f (x, y)  xy  2x  2xe  2e
3 x x
M N The DE is exact.
   h(y)
y x

 f xy  2x 3  2xe x  2e x  C,
Then  x  h(y)
y C  c  c1
But f  N(x, y)  x,
Note
f
y We may also start with  N(x, y)
 y
 h(x, y)  0  f
i.e x
y
h(y)  c1, a constan t Then 
f
dy   x dy
y 
 f (x, y)  xy  2x  2xe
3 x

  f (x, y)  xy  h(x)
 2e  c1
x
f
  y  h(x)
A solution of the DE is x
and so on.
f (x, y)  c
xy  2x  2xe  2e  c1 c
3 x x
 Example 2

Solve the IVP

 
x  y
2
dx  2xy  x 2
1dy  0, y(1)  1


 Solution

LetM(x, y)  x  y   x 2  2xy  y 2
2

andN(x, y)  2xy  x 1
2

M N
 2 x  y ,  2 x  y
y x
 M N
 
y x

The DE is exact.
There exists a function f (x, y) such that
f
 M  x, y   x 2  2 xy  y 2
x
f
 N  x, y  2 xy  x 2  1
y
 f x3
 x dx   x2
 2 xy  y 
2
dx
3
 x 2 y  xy 2  y  c1  c

x3 x3
f ( x, y )   x 2 y  xy 2  h( y )  x 2 y  xy 2  y  C , C  c  c1
3 3
f
 x 2  2 xy  h( y ) The initial condition y (1)  1 gives
y
1 4
f  1  1  1  c,  c 
But  N ( x, y )  2 xy  x 2  1 3 3
y
h( y )  1 The solution of the IVP is
h( y )   y  c x3 4
 x y  xy  y 
2 2

x3 3 3
f ( x, y )   x 2 y  xy 2  y  c1
3
A solution of the DE is
f ( x, y )  c
INTEGRATING FACTOR

M(x, y)dx  N(x, y)dy

Let 0
M N
If 
y x
Then the DE is NOT exact .
To make it exact, we need to multiply the DE by an
integrating factor .
My  Nx
It can be shown that if is a function of x alone
N
(or a constant),
 then an integrating factor is
M y N x

(x)  e  N
dx
Nx  My
If is a function of y alone (or a constant), then
M
an integrating factor is
N x M y

(y)  e
dy
M

Example 3

Solve

6xydx  4 y  9x 2 dy  0
Solution

LetM(x, y)  6xy and N(x, y)  4 y  9x 2

M N
 6x,  18x
y x
M N

y x (except when x = 0)
The DE is not exact.

My  Nx 6x 18x 12x
 2  
N 4 y  9x 4 y  9x 2

M y  N x is not a function of x alone.

N
Nx  My 18x  6x 2
  which is a function of y alone.
M 6xy y

An integrating factor is
2
 dy
ln y 2
e y
e 2 ln y
e  y2
6xy 3 dx  (4 y 3  9x 2 y 2 )dy  0 is exact. (Please verify!)

P(x,
Let y)  6xy 3
and Q(x, y)  4 y 3
 9x 2 2
y

There exists a function f (x, y) such that

f f
 3
 P(x, y)  6xy and  Q(x, y)  4 y 3  9x 2 y 2
x y

 f
Then  x dx   dx
6xy 3

 f (x, y)  3x 2 y 3  h(y)
f
  9x 2 y 2  h(y)
y
But f
 Q(x, y)  4 y 3  9x 2 y 2
y
 h(y)  4 y 3
h(y)  y 4  c1
 f (x, y)  3x y  y  c1
2 3 4
 A solution of the DE is

f (x, y)  c
3x 2 y 3  y 4  c1  c
3x 2 y 3  y 4  C, C  c  c1

