Math 350 Notes 3
Math 350 Notes 3
Contents
Exact Equations 2
1
Exact Equations
We consider an equation of the form
M (x, y)dx + N (x, y)dy = 0
Recall a differential equation of the first order can be written in the form
dy
= f (x, y) (1)
dx
If we can express f (x, y) in the form
−M (x, y)
f (x, y) = then
N (x, y)
dy −M (x, y)
= from 1
dx N (x, y)
Then,
M (x, y)dx + N (x, y)dy = 0 (2)
Note that we must have + between M and N components, and the equation must be equal to zero on rhs.
Since a general solution of a first order D.E. is expected to contain one arbitrary constant, we can assume
that solution of equation 2 will be of the form U (x, y) = c where c is a constant.
∂U ∂U
dU = dx + dy = 0 (3)
∂x ∂y
Comparing equations 2 and 3
∂U ∂U
M (x, y) = and N (x, y) = (4)
∂x ∂y
From 4
∂2U
∂M ∂ ∂U
= =
∂y ∂y ∂x ∂y∂x
2
∂N ∂ ∂U ∂ U
= =
∂x ∂x ∂y ∂x∂y
If we assume continuity of U, and continuity of its first differential, then
∂2U ∂2U
=
∂x∂y ∂y∂x
∂M ∂N
∴ = (5)
∂y ∂x
Equation 5 is the necessary and sufficient condition for Equation 2 to be exact. We can now write
dU is an exact differential.
Generally, an equation is said to be exact if it can be written in the form
such that
∂M ∂N
=
∂y ∂x
2 Chisara P. Ogbogbo
Steps in solving exact DE
∂M ∂N
If = then the equation is exact
∂y ∂x
(iv) We integrate this equation w.r.t the 2 variables taking one at a time
Z Z
dU = M (x, y)dx + f (y) = U (6)
Z Z
dU = N (x, y)dy + g(x) = U (7)
Alternative to (iv)
As we integrate w.r.t x, we ignore the arbitrary constant of integration. Also, as we integrate with respect
to y, we ignore the constant of integration. We construct the desired function U (x, y) = C. We ’merge’ the
two resulting expressions obtained from the separate integrations. To ’merge’ these two expressions, write down
each term exactly once, even if a particular term appears in both results.
Example 1
3 Chisara P. Ogbogbo
Since
∂M ∂N
= , the equation is exact
∂y ∂x
(ii). Let u(x, y) = c be a solution
M (x, y)dx + N (x, y)dy = du(x, y) = 0
(2xy − y + 2x)dx + (x2 − 2xy + 2y)dy = du(x, y) = 0
2
Z
(2xy − y 2 + 2x)dx + f (y) = x2 y − xy 2 + x2 + f (y) (8)
Z
(x2 − 2xy + 2y)dy + g(x) = x2 y − xy 2 + y 2 + g(x) (9)
Example 2
dy e−y − yex
= −y y(0) = 4
dx xe + ex
Solution
We rearrange, to obtain
Then
4 Chisara P. Ogbogbo
Integrating (M ) w.r.t x, we have
Z
(yex − e−y ) dx + f (y) = yex − e−y x + f (y) (10)
y(1) − 0 = C
4=C
∴ U = yex − xe−y = 4
Example 3
M = y 2 − 2x
N = 2xy + 1
∂M ∂N
= 2y and = 2y
∂y ∂x
∂M ∂N
Since ∂y = ∂x , the D.E. is exact. We assume the solution to be U (x, y) = C
Z Z
M dx = (y 2 − 2x) dx = y 2 x − x2
Z Z
N dy = (2xy + 1) dy = xy 2 + y
∴ U = xy 2 − x2 + y = C
Example 4
5 Chisara P. Ogbogbo
Writing all terms that appear in both 12 and 13 without repeating terms, we have
Example 5
Solution
M (x, y) = 3x2 y − 1
N (x, y) = x3 + 6y − y 2
6 Chisara P. Ogbogbo
IF works sometimes for a non exact D.E. to make it exact and then solve it the usual way.
Let µ(x, y) be that IF when obtained is used to multiply the D.E. Then
becomes exact.
Recall criteria for exactness ∂M ∂N
∂y = ∂x
The differential equation 14 will be exact if
i.e. differentiate coefficients of dx (first component) w.r.t y and coefficient of dy (second component) w.r.t x
By the product rule, equation 15 gives
µx N − µy M = (My − Nx )µ (16)
M, N, My and Nx are all functions of x and y. In that case, we need to solve a P.D.E to obtain µ. In order to
circumvent this, we do the following:
We suppose µ is a function of one variable.i.e. µ depends on x alone or µ depends on y alone.
Then µx = dµ
dx and µy = 0
Then equation 16 becomes
dµ M y − Nx
= µ (17)
dx N
If
M y − Nx
N
still depends on both x and y, then we have not resolved the issue. However, if we make all necessary algebraic
simplifications and
M y − Nx
N
now depends on x alone, then equation 17 is a first order D.E. which is linear.
My − Nx
Let = P (x)
N
Then µ′ − P (x)µ = Q(x)
Then R My −Nx
dx
IF = µ(x) = e N
7 Chisara P. Ogbogbo
is a function of only y, then equation λ can be solved for µ. We can have the IF as follows:
R Nx −My
dy
e M
Example 1
My = 4xy 3 Nx = 4xy 3
∴ DE is now exact
We now solve as follows:
Z Z
Obtain M dx and N dy Merge to obtain U = C
Z
1
xy 4 dx = x2 y 4
2
x2 y 4 y6
Z
(2x2 y 3 + 3y 5 − 20y 3 )dy = + − 5y 4
2 2
1 1
∴ U = x2 y 4 + y 6 − 5y 4 = C
2 2
is the solution of the D.E.(by merging the two expressions without repeating terms)
Example 2
8 Chisara P. Ogbogbo
Consider the D.E
x−y 1
=
x(x − y) x
1
R
∴ IF = µ = e x dx = elnx = x
M = 3x2 y − xy 2 N = x3 − x2 y
3x3 x2
Z
1
(3x2 y − xy 2 )dx = y − y 2 = x3 y − x2 y 2
3 2 2
x2 y 2
Z
(x3 − x2 y)dy = x3 y −
2
Merging the two expressions without repeating terms, we have:
1
U = x3 y − x2 y 2 = C
2
Example 3
Students to try in class
Solve the DE
9 Chisara P. Ogbogbo
(x + y)sinydx + (xsiny + cosy)dy = 0
This DE is not exact since
My = (x + y)cosy + siny and Nx = siny
Notice that
Nx − M y M y − Nx
=
M −M
And
M y − Nx
−M
is given as
(x + y)cosy cosy
=−
−(x + y)siny siny
cosy
− siny is used to obtain the IF
R cosy −1
− siny
IF = e dy
= e−ln(siny) = eln(siny) = (siny)−1
Z
1
(x + y)dx = x2 + xy
2
Z
cosy
and x+ dy = xy + ln|siny|
siny
1
∴ U = x2 + xy + ln|siny| = C
2
Example 4
10 Chisara P. Ogbogbo
We multiply both sides of the DE by e2x to obtain
1 1
e0 (1) + (0)e0 − e0 = C
2 4
1
1− =C
4
3
C=
4
Students to try
11 Chisara P. Ogbogbo