Math 432 HW 2.

6 Solutions
Assigned: 1-8, 11, 19, 25, 29, 32, 34, 39, and 40.

Selected for Grading: 2, 7, 34, 39

Solutions:
1. Given: (y – 4x – 1)2dx – dy = 0.
Solving this for dy/dx gives dy/dx = (y – 4x – 1)2, which is of the form dy/dx = G(ax + by).

2. Given: 2tx dx + (t 2 – x2) dt = 0.

In the "straight forward" way of reading this equation, x is the independent variable and t is the
dependent. So, solving for dt/dx, we get
(t 2 – x2) dt = – 2tx dx
dt/dx = –2tx/(t 2 – x2) = 2tx/(x2 – t 2)

This is homogeneous:

In the less straight forward way of reading the given equation we'd have t the independent variable and x
the dependent. So, this time solving for dx/dt, we get
2tx dx = (x2 – t 2) dt

which shows again that the equation is homogeneous.

Note also that since

then

So the equation is Bernoulli with n = –1.

3. Given: (t + x + 2) dx + (3t – x – 6) dt = 0.
This is an equation with linear coefficients.

4. Given: dy/dx + y/x = x3y2.

Rewriting it slightly gives dy/dx + (1/x)y = x3y2, which shows that it is a Bernoulli equation with n = 2.

5. Given: (ye –2x + y3) dx – e –2x dy = 0.

Trying to solve for dy/dx I got dy/dx = y + y3e2x.
Modifying slightly gives dy/dx – y = e2xy3, which is a Bernoulli equation with n = 3.
6. Given: .
In correct differential form: .
Since M(tθ, ty) = tθ = tM(θ,y) then M is homogeneous of degree α = 1.
Similarly, since N(tθ, ty) = ty + t = tN(θ, y) then N is homogeneous of the same degree.
So the differential equation is homogeneous.

Solving for dy/dθ: .

So we have , which shows that this is a Bernoulli equatio with n = 1/2.

7. Given: (y3 – θy2) dθ + 2θ 2y dy = 0.

This one is homogeneous:
M(tθ, ty) = t 3y3 – t 3θy2 = t 3M(θ, y), so M is homogeneous of degree α = 3 and
N(tθ, ty) = 2t 3θy2 = t 3N(θ, y), so N is homogeneous of the same degree.

8. Given: cos(x + y) dy = sin(x + y) dx.

Manipulation gives dy/dx = sin(x + y)/cos(x + y) = tan(x + y), which is of the form dy/dx = G(x + y).

11. (y2 – xy) dx + x2dy = 0.

Here's a check before assuming that it's homogeneous.
M(tx, ty) = t 2y2 – txty = t 2(y2 – xy) = t 2M(x, y)
N(tx, ty) = t 2x2 = t 2N(x, y).

Before proceeding, let me put this equation into a more manageable form: dy/dx = (xy – y2)/x2.

Now let v = y/x.

Then y = xv and so dy/dx = x(dv/dx) + v.
Substituting gives x(dv/dx) + v = (x2v – x2v2)/x2 = v – v2.
So x(dv/dx) = – v2 which is separable. (The first move divides by v2 = y2/x2, which assumes y ≠ 0.)

–1/v = –ln|x| + C
1/v = ln|x| + C
x/y = ln|x| + C

Solution: y = x/(ln|x| + C).

Also, consider the solution y(x) ≡ 0. It is a solution of the differential equation above, so we need to
include it here.

Solutions: y = x/(ln|x| + C) and y ≡ 0.

19. dy/dx = (x – y + 5)2.

Let z = x – y.
Then y = x – z and so dy/dx = 1 – dz/dx.
Substituting gives
1 – dz/dx = (z + 5)2
dz/dx = 1 – (z + 5)2 = 1 – z2 – 10 z – 25 = –(z2 + 10z + 24) = –(z + 4)(z + 6), which is separable.

(We might have just divided by zero. We'll come back to this later.)
The integral on the left is done using partial fractions:

From which we get two equations:

a+b=0
6a + 4b = 1
This system has solution a = 1/2, b = –1/2.
Returning to ...

ln|(z + 4)/(z + 6)| = –2x + C

|(z + 4)/(z + 6)| = Ce –2x for some C > 0
(z + 4)/(z + 6) = Ae –2x for some A ≠ 0.
z + 4 = Ae –2x(z + 6)
z(1 – 6Ae –2x) = 6Ae –2x – 4
x – y = z = (6Ae –2x – 4)/(1 – 6Ae –2x)

Solution: y = x – (6Ae –2x – 4)/(1 – 6Ae –2x).

Now, returning to the issue of dividing by zero. . . This could have happened if z = –4 or if z = –6.

The constant function z ≡ –4 would yield x – y = –4, and hence y = x + 4. In that case, dy/dx = 1 (not
zero) and (x – y + 5)2 = (x – (x + 4) + 5)2 = 12 = 1. So this function is a solution to the given differential
equation.

And the constant function z = –6 would give x – y = –6, and hence y = x + 6. Thus dy/dx = 1 again and
(x – y + 5)2 = (x – (x + 6) + 5)2 = (–1)2 = 1. So this is another solution to the differential equation.

So then, the solutions are y = x – (6Ae –2x – 4)/(1 – 6Ae –2x) for some constant, A, and y = x + 4.

(That one solution y = x + 4 is included in the more general solution when A = 0.)
25. dx/dt + tx3 + x/t = 0.
Writing this in the form dx/dt + (1/t)x = –tx3 shows that this is Bernoulli with n = 3.
So we make the substitution v = x –2. (But note also that x(t) ≡ 0 is a solution.)
Then dv/dt = –2x –3(dx/dt) and multiplying both sides of the equation dx/dt + (1/t)x = –tx3 by –2x –3 gives
–2x –3(dx/dt) – 2x –2t –1 = 2t
dv/dt – 2t –1v = 2t, which is linear with P(t) = –2/t and Q(t) = 2t.
The integrating factor:
∫ P(t) dt = ln(1/t 2)
μ(t) = 1/t 2
v = t 2[∫ (1/t 2)(2t) dt + C] = t 2[2 ln|t| + C] = Ct 2 + 2t 2ln|t|
So x –2 = Ct 2 + 2t 2ln|t| for some constant C, or more simply (but still implicitly)

The solutions are given by x 2 = t 2/[C + 2 ln|t|] and x(t) ≡ 0.

29. Given: (–3x + y – 1) dx + (x + y + 3) dy = 0.

In standard form: dy/dx = (3x – y + 1)/(x + y + 3).

This is not homogeneous but it is an equation with linear coefficients. So we'll want to make the
substitution x = u + h, y = v + k, where h and k satisfy these two equations:
–3h + k – 1 = 0
h+k+3=0
The solution to this system is h = –1, k = –2. So we let x = u – 1 and y = v – 2. These give dx = du and
dy = dv. So when we substitute into the original equation we get
(–3u + v) du + (u + v) dv = 0.
In standard form: dv/du = (3u – v)/(u + v), which is homogeneous. So, let z = v/u. Then v = uz, and dv/du
= u(dz/du) + z. And substituting gives
u(dz/du) + z = (3u – uz)/(u + uz) = (3 – z)/(1 + z)

This is separable:

NOTE: We just assumed that u ≠ 0, z ≠ 1, and z ≠ –3.

If u ≡ 0 then x ≡ –1. For that to make sense we'd have to think of y as the independent variable and x
as the dependent. In that case we'd have the differential equation dx/dy = (x + y + 3)/(3x – y + 1). In
that equation we'd have dx/dy = 0 and (x + y + 3)/(3x – y + 1) = (4 + y)/(4 – y) ≠ 0. So x(y) ≡ –1 is not
a solution.
If z ≡ 1 then v/u = (y + 2)/(x + 1) ≡ 1 from which it follows that y = x – 1. In that case we'd have
dy/dx = 1 and (3x – y + 1)/(x + y + 3) = (3x – (x – 1) + 1)/(x + (x – 1) + 3) = (2x + 2)/(2x + 2) = 1, so
this is a solution: y = x – 1.
Finally, if z ≡ –3 then v/u = (y + 2)/(x + 1) ≡ –3 from which we'd get y = –3x – 5. In that case we'd
have dy/dx = –3 and (3x – y + 1)/(x + y + 3) = (3x – (–3x – 5) + 1)/(x – 3x – 5 + 3) = (6x + 6)/(–2x – 2),
which is identically equal to –3. So we have a second solution: y = –3x – 5.
Having dealt with the special cases, we can proceed with the separable equation.
 Using partial fractions, we get

ln|(z – 1)(z + 3)| = ln(1/u2) + C

|(z – 1)(z + 3)| = A/u2 for some A > 0.
(z – 1)(z + 3) = B/u2 for some B ≠ 0.
Note that allowing B = 0 would include the two extra solutions we found above.

Back substituting z = v/u gives

(v – u)(v + 3u) = B
v2 + 2uv – 3u2 = B
Back substituting x = u – 1 and y = v – 2 gives

Implicit solution: (y + 2)2 + 2(x + 1)(y + 2) – 3(x + 1)2 = B for some constant, B.

32. Given: (2x + y + 4) dx + (x – 2y – 2) dy = 0.

Remark: This is exact, but the directions are to solve it using the method for equations with linear
coefficients. OK, here we go.

In standard form(s) we have either dy/dx = (2x + y + 4)/(2y – x + 2) or dx/dy = (2y – x + 2)/(2x + y + 4).

We want to make the substitution x = u + h and y = v + k where h and k satisfy the equations
2h + k + 4 = 0
h – 2k – 2 = 0
This system's solution is h = –6/5, k = –8/5.
Here's the new differential equation:
(2u + v) du + (u – 2v) dv = 0.
In standard form: dv/du = (2u + v)/(2v – u), which is homogeneous. So we substitute z = v/u or v = uz,
which gives dv/du = u(dz/du) + z. And we get this new differential equation:
u(dz/du) + z = (z + 2)/(2z – 1).
Which can be manipulated into the separable differential equation
u(dz/du) = (z + 2)/(2z – 1) – z = –2(z2 – z – 1)/(2z – 1).
Separating variables:

(And we assumed that z2 – z – 1 ≠ 0.)

Integrating both sides gives
ln|z2 – z – 1| = –2ln|u| + C = ln(1/u2) + C
|z2 – z – 1| = A/u2 for some A > 0.
z2 – z – 1 = B/u2 for some B ≠ 0.
 Note that allowing B = 0 will enable us to include whatever solutions we may have omitted above when we
assumed that z2 – z – 1 ≠ 0. So the solution of the current differential equation is given implicitly by
z2 – z – 1 = B/u2.
Back substituting:
(v/u)2 – v/u – 1 = B/u2
(v2 – vu – u2)/u2 = B/u2
v2 – vu – u2 = B
And back substituting again:
(y + 8/5)2 – (y + 8/5)(x + 6/5) – (x + 6/5)2 = B2
(5y + 8)2/25 – (5y + 8)(5x + 6)/25 – (5x + 6)2/25 = B2
(5y + 8)2 – (5y + 8)(5x + 6) – (5x + 6)2 = 25B2

Solution: (5y + 8)2 – (5y + 8)(5x + 6) – (5x + 6)2 = C for some C ≥ 0.

34. The given differential equation 2tx dx + (t 2 – x2) dt = 0 is both homogeneous and Bernoulli. As was noted
above (see #2), we can view x as either the dependent variable or the independent variable. It is when we
view x as the dependent variable that we have the option of viewing this equation as either homogeneous
or Bernoulli.
=========================================================================
Let me handle it first as a Bernoulli equation.
In the form that shows it to be such we have

[We assumed that t ≠ 0 to get this far. You should note that the constant function t ≡ 0 is a solution to the
original equation.]
Preparing to make the substitution:
v = x1 – (–1) = x2 so dv/dt = 2x ∙ dx/dt
Multiplying both sides of the equation by 2x:

This is linear, with integrating factor

So the solution is given by

And so we have the implicit solution: x2 = –t 2 + Ct or x2 + t 2 = Ct for some constant, C.

Since this form doesn't include it, I'll also mention the constant solution t ≡ 0.
=========================================================================
Thinking of the equation as a homogeneous one, the form we want in this case is

Preparing for the substitution:

v = x/t so x = tv and dx/dt = t ∙ dv/dt + v
Substituting:

This is separable:

ln(v2 + 1) = –ln|t| + C = ln|1/t| + C

v2 + 1 = A/t for some A ≠ 0
v2 = A/t – 1 = (A – t)/t
Back substituting:
x2/t 2 = (A – t)/t
x2 = t (A – t) = At – t 2
Implicit solution: x2 + t 2 = At for some constant, A ≠ 0.
=========================================================================
If you take the other view of the given equation, namely that x is independent and t is dependent, then you
start from the homogeneous equation

For the substitution:

v = t/x so t = xv and dt/dx = x ∙ dv/dx + v
Substituting:

Separating (and this time that assumes that v ≠ 0):

(That last step was made using partial fraction decomposition.)

ln(v2 + 1) – ln|v| = –ln|x| + C = ln|1/x| + C

v2 + 1 = Bv/x
Back substituting: t 2/x2 + 1 = Bt/x2

Solution: t 2 + x2 = Bt for some B ≠ 0.

39. As we know from above, the equation (y3 – θy2) dθ + 2θ 2y dy = 0 is homogeneous. It also has (by
inspection) an equilibrium solution y(θ) ≡ 0. We'll keep this solution in mind, but now on to the method for
homogeneous equations.

So we're looking at this equation:

Preparing for the substitution:

v = y/θ so y = θv and dy/dθ = θ ∙ dv/dθ + v
Making the substitution:

Separating the variables (which assumes that v ≠ 0 and v +1 ≠ 0, and the constant functions, v(θ) ≡ 0, –1,
are solutions):

ln |v| – ln |v + 1| = (–1/2)ln |θ| + C1 (I'm going to multiply by –2.)

ln (1/v2) + ln(v + 1)2 = ln|θ| + C2
(v + 1)2/v2 = Cθ for some C ≠ 0.
(v + 1)2 = Cθv2 for some C and v ≡ 0.
(Allowing C = 0 picks up one of the constant solutions mentioned above.)
Back substituting:
(y/θ + 1)2 = Cθ(y/θ)2
(y + θ)2/θ2 = Cθy2/θ2
(y + θ)2 = Cθy2

Solutions: (y + θ)2 = Cθy2 for some C, and the constant function y ≡ 0.

NOTE: This cannot be solved for y without additional assumptions about the sign of y and the sign of θ. So
it should be left in implicit form.
40. The differential equation cos(x + y) dy = sin(x + y) dx can be put into the form dy/dx = tan(x + y).
So we make the substitution v = x + y.
That gives dv/dx = 1 + dy/dx.
So dy/dx = dv/dx – 1, and we get the new differential equation

Dividing by tan v + 1 (which we'll assume for now is not zero) and "multiplying" by dx gives something to
integrate:

That first integral takes some doing. It begins with some straightforward manipulation:

Then there's this tricky maneuver:

I did that so that that fraction inside the brackets would be du/u – do you see it? Anyway, we can now
integrate.

for some C ≠ 0.

Back substituting: cos(x + y) + sin(x + y) = Ce2x – (x + y) = Cex – y for some C ≠ 0.

And note that allowing C = 0 would give cos(x + y) + sin(x + y) = 0 which, after you divide through by
cos(x + y), simplifies to 1 + tan(x + y) = 0. So that picks up that special case created when we separated
variables.

Answer: cos(x + y) + sin(x + y) = Cex – y for some constant, C.