2.

4 Substitution Methods
The first-order ode may not yield at once (or at all) to the methods we have studied so far. Even then
it may be possible to transform the equation into a type that we know how to solve by just changing the
variables.

2.4a Equations with Homogeneous Coefficients

The function ℎ(𝑥, 𝑦) is said to be homogeneous of degree k in x and y if, and only if,
ℎ(𝜆𝑥, 𝜆𝑦) = 𝜆𝑘 ℎ(𝑥, 𝑦) . (1)
For example, lets determine the homogeneity of
𝑦 𝑥4
ℎ(𝑥, 𝑦) = 2𝑦 3 𝑒 𝑥 − (2)
𝑥 + 3𝑦
Substitute x with 𝜆𝑥 and y with 𝜆𝑦.
𝜆𝑦 (𝜆𝑥)4
ℎ(𝜆𝑥, 𝜆𝑦) = 2(𝜆𝑦)3 𝑒 𝜆𝑥 −
𝜆𝑥 + 3𝜆𝑦
𝜆 𝑦 𝜆4 𝑥 4
= 2𝜆3 𝑦 3 𝑒 𝜆 ∙ 𝑥 −
𝜆(𝑥 + 3𝑦)
𝑦 𝜆3 𝑥 4
= 2𝜆3 𝑦 3 𝑒 𝑥 −
(𝑥 + 3𝑦)
𝑦 𝑥4
= 𝜆3 (2𝑦 3 𝑒 𝑥 − ) (3)
(𝑥 + 3𝑦)
Since the terms inside the parenthesis is equal to function (2), function (3) becomes
ℎ (𝜆𝑥, 𝜆𝑦) = 𝜆3 ℎ(𝑥, 𝑦) (4)
By comparing (1) and (4), we can conclude that the function (2) is homogeneous of degree 3 in x and
y.
The following theorems will prove the use of substitution method in solving de with homogeneous
coefficients.

Theorem 1. If M(x, y) and N(x, y) are both homogeneous of the same degree, the function M(x, y)⁄N(x, y)
is homogeneous of degree zero.
Proof:
λk M(x, y) M(x, y) M(x, y)
k
= λk−k = λ0
λ N(x, y) N(x, y) N(x, y)
( )
Theorem 2. If h x, y is homogeneous of degree zero in x and y, h(x, y) is a function of y/x alone.
Proof. Let us put y = vx. Then Theorem 2 states that if h(x, y) is homogeneous of degree zero, h(x, y)
is a function of v alone.
h(x, y) = h(x, vx)
If we now let x play the role of λ and knew beforehand that f(x, y) is homogeneous of degree zero,
h(x, vx) = x 0 h(1, v) = h(1, v) , (5)
By (5), h(x, y) depends on v alone, as stated in Theorem 2.

Suppose that the coefficients M and N in a 1st-order ODE,

M(x, y)dx + N(x, y)dy = 0 , (6)
are both homogeneous functions and are of the same degree in x and y. By Theorem 1 and 2, the ratio
M/N is a function y/x alone. Hence equation (1) may be put in the form
 dy M(x, y)
+ =0
dx N(x, y)
dy y
+ g( ) = 0 , (7)
dx x
This suggests the introduction of a new variable v by putting y = vx. Then (7) becomes
dv
x + v + g(v) = 0 , (8)
dx
in which the variables are separable. After solving (8) using variable separable method, insert y⁄x for
v, and thus arrive at the solution of (6). We have shown that the substitution y = vx will transform
equation (6) into an equation in v and x in which the variables are separable.
The method above would have been equally successful had we used x = vy to obtain from (6) an
equation in y and v.

Example 1 Variable substitution of y wih vx

Solve the equation
(x 2 − xy + y 2 ) dx − xy dy = 0 . (9)
Let’s check if the coefficients in (9) are both homogeneous and of the same degree.
M(x, y) = x 2 − xy + y 2 , N(x, y) = xy (10)
Substituting x with λx and y with λy in (10)
M(λx, λy) = (λx)2 − λx λy + (λy)2 , N(λx, λy) = λx λy
2 2 2 2 2
= λ x − λ xy + λ y , N(λx, λy) = λ2 xy
2 2 2
= λ (x − xy + y ) , N(λx, λy) = λ2 (xy) (11)
Since the functions inside the parenthesis of (11) are equal to the functions (10), we can rewrite (11)
M(λx, λy) = λ2 M(x, y), N(λx, λy) = λ2 N(x, y) , (12)
Hence the coefficients (10) of de (9) are both homogeneous and of the same degree in x and y, let us
put y = vx. Then dy = v dx + x dv), and equation (9) becomes
(x 2 − x 2 v + x 2 v 2 ) dx − x 2 v(v dx + x dv) = 0 ,
Simplify to make it variable separable equation. x 2 can be removed.
(1 − v + v 2 )dx − v(vdx + xdv) = 0 ,
or
(1 − v)dx − xv dv = 0 .
Hence we separate variable to get
dx vdv
+ =0.
x v−1
dx 1
+ [1 + ] dv = 0
x v−1
a family of solution is seen to be
ln|x| + v + ln|v − 1| = ln|c| ,
Exponentiation on both sides, (Note that 𝑒 ln 𝑎 = 𝑎, 𝑎𝑛𝑑 𝑒 (𝑎+𝑏) = 𝑒 𝑎 𝑒 𝑏 )
𝑥(𝑣 − 1)𝑒 𝑣 = 𝑐 .
In terms of the original variables, these solutions are given by
𝑦 𝑦
𝑥 ( − 1) 𝑒 𝑥 = 𝑐 ,
𝑥
or
𝑦
(𝑦 − 𝑥)𝑒 𝑥 = 𝑐 .

Example 2 Variable substitution of x with vy.

Solve the equation
 xy dx + (x 2 + y 2 )dy = 0 . (13)
Again the coefficients in the equation are homogeneous and of the same degree. We could use y =
vx, but the relative simplicity of the dx term in (13) suggests that we put x = vy. Then dx = v dy + y dv,
and equation (13) is replaced by
vy 2 (v dy + y dv) + (v 2 y 2 + y 2 )dy = 0,
Simplify to make it variable separable equation. Remove y 2 term.
v(v dy + y dv) + (v 2 + 1)dy = 0 ,
or
vy dv + (2v 2 + 1)dy = 0 ,
Separating variables
v dv dy
2
+ = 0,
2v + 1 y
or
1 4v dv dy
∙ 2 + =0,
4 2v + 1 y
Simplify by multiplying both sides by 4. Integrate to get the family of solution.
ln(2v 2 + 1) + 4 ln |y| = ln c ,
Exponentiation on both sides, (Note that 𝑒 ln 𝑎 = 𝑎, 𝑏 ln 𝑎 = ln 𝑎𝑏 , 𝑎𝑛𝑑 𝑒 (𝑎+𝑏) = 𝑒 𝑎 𝑒 𝑏 )
y 4 (2v 2 + 1) = c .
In terms of the original variables, these solutions are given by
2x 2
y 4 ( 2 + 1) = c;
y

y 2 (2x 2 + y 2 ) = c .

Frequently in equations with homogeneous coefficients, it is quite immaterial whether one uses y =
vx or x = vy. However, it is sometimes easier to substitute for the variable whose differential has the
simpler coefficient.