Math 432 HW 1.

2 Solutions

Assigned: 1, 2, 3, 5, 8, 9, 12, 13, 15, 18, 19, 21, 22, 23, 27, 28, 30, and 31

NOTE: For #18, the text from "Note that" to the end is just for your edification.

Selected for Grading: 3, 9, 12, 23, 28

Solutions:
1. (a) We have the differential equation x(dy/dx) = 2y, the proposed solution (x) = x
2
, and the interval
I = (, ).
'(x) = 2x
So x'(x) = x(2x) = 2x
2
= 2(x).
So (x) is a solution, and it is defined on the given interval.
(b) Here we have the differential equation dy/dx + y
2
= e
2x
+ (1 2x)e
x
+ x
2
1, the proposed solution
(x) = e
x
x, and the interval (, ).
'(x) = e
x
1.
So
'(x) + (x)
2
= e
x
1 + (e
x
x)
2

= e
x
1 + e
2x
2xe
x
+ x
2

= e
x
(1 2x) + e
2x
1 + x
2

= e
2x
+ (1 2x)e
x
+ x
2
1, as needed.
That much verifies that (x) is a solution. And since this function is defined on all of (, ) then we're
finished with this part.
(c) Finally, we have x
2
d
2
y/dx
2
= 2y and (x) = x
2
x
1
on the interval (0, ).
'(x) = 2x + x
2

''(x) = 2 2x
3

So x
2
''(x) = x
2
(2 2x
3
) = 2x
2
2x
1
= 2(x
2
x
1
) = 2(x).
Thus this function (x) is indeed a solution to the given differential equation.
And, since the function is defined on the interval (0, ), then we've finished with this part too.

NOTE: This function is also a solution to the differential equation on the interval (, 0).

2. (a) The differential equation is dy/dx = 1/(2y) and the solution is supposedly given implicitly by the
relation y
2
+ x 3 = 0 on the interval (, 3).
Differentiating both sides of the relation gives 2y(dy/dx) + 1 = 0.
Solving this for dy/dx gives dy/dx = 1/(2y), so this relation does give a solution implicitly.

What about the interval? Take an algebraic look at the relation y
2
+ x 3 = 0. Solving this for x we
get x = 3 y
2
. Since y
2
0, then this shows that x 3. Since our interval must be an open one, we
throw out the endpoint 3 and are left with the interval x < 3 or (, 3).

The following graph of the given relation make things a little easier to see.


(b) The relation xy
3
xy
3
sin x = 1 can be solved for y.
xy
3
(1 sin x) = 1
y
3
= 1/[x(1 sin x)] {Note that making that step assumes that x 0 and that sin x 1.}
y = {1/[x(1 sin x)]}
1/3
= [x(1 sin x)]
1/3

Differentiating:

This much verifies that the implicitly given y is a solution to the differential equation. As for the
interval of definition, any interval that does not include x = 0 and that does not include any x-value for
which sin x = 0 would do. The given interval (0, /2) is one such interval.

3. Given: y = sin x + x
2
.
dy/dx = cos x + 2x and
2
y/dx
2
= sin x + 2.
So d
2
y/dx
2
+ y = sin x + 2 + sin x + x
2
= 2 + x
2
= x
2
+ 2.
Yes, this is a solution to the given differential equation. [On the interval (, ).]

5. Given: = 2e
3t
e
2t
.
d/dt = 6e
3t
2e
2t
and d
2
/dt
2
= 18e
3t
4e
2t
.
So
d
2
/dt
2
(d/dt) + 3 = 18e
3t
4e
2t
(2e
3t
e
2t
)(6e
3t
2e
2t
) + 3(2e
3t
e
2t
)
= 18e
3t
4e
2t
12e
6t
+ 10e
5t
2e
4t
+ 6e
3t
3e
2t

= 12e
6t
+ 10e
5t
2e
4t
+ 24 e
3t
7e
2t

e
2t

No, this is not a solution to the given differential equation.

8. Given: y = 3 sin 2x + e
x
.
y' = 6 cos 2x e
x
and y'' = 12 sin 2x + e
x
.
So y'' + 4y = 12 sin 2x + e
x
+ 4(3 sin 2x + e
x
) = 5e
x
.
Yes, this is a solution to the given differential equation.

9. Given: x
2
+ y
2
= 4.
Differentiating: 2x + 2yy' = 0.
Solving for y': y' = x/y.
No, this is not a solution to the differential equation dy/dx = x/y.

x
y
12. Given: x
2
sin(x + y) = 1.
Differentiating: 2x cos(x + y)(1 + y') = 0.
Solving for y':
1 + y' = 2x/cos(x + y)
y' = 2x/cos(x + y) 1 = 2x sec(x + y) 1.
Yes, this relation is an implicit solution to the given differential equation.

13. Differentiate both sides of the equation sin y + xy x
3
= 2 and you get
(cos y) y' + xy' + y 3x
2
= 0.
Solving this equation for y' gives y' = (3x
2
y)(cos y + x).
Differentiating both sides of this second equation you get
(cos y) y'' + y'(sin y) y' + xy'' + y' + y' 6x = 0.
Solving this for y'':
(cos y + x) y'' = 6x 2y' + (sin y)(y')
2

So the answer is "yes".


15. Given: (x) = 2/(1 ce
x
) = 2(1 ce
x
)
1
for some arbitrary constant, c.
'(x) = 2(1 ce
x
)
2
(ce
x
) = 2ce
x
/(1 ce
x
)
2

And so

So each such (x) is indeed a solution to the given differential equation.

Here are the requested graphs.


18. Given: c > 0; (x) = (c
2
x
2
)
1
; and dy/dx 3y = 3, y(0) = 1/c
2
on (c, c).
Differentiating gives '(x) = (c
2
x
2
)
2
(2x) = 2x/(c
2
x
2
)
2
= 2xy
2
.
This much shows that this (x) is a solution to the differential equation on some interval.
It is also a solution to the initial-value problem since (0) = (c
2
0
2
)
1
= 1/c
2
.
Finally, note that (x) is defined on any interval that excludes both x = c and x = c, for example the
given interval.

19. The equation (dy/dx)
2
+ y
2
+ 4 = 0 can't have any real-valued solution because if there were a solution then
we'd have a square, namely (dy/dx)
2
, equal to a negative number, namely 4 y
2
. Since that can't be, then
no such solution could exist.

21. For both parts (a) and (b) we start with (x) = x
m
for some yet-to-be-determined m. Also for both parts
we'll need the first and second derivatives of (x), so I'll calculate them here:
'(x) = mx
m 1
and ''(x) = m(m 1)x
m 2
.
(a) I'll just make the substitution and see whether it tells me what m must be equal to.
3x
2
''(x) + 11x'(x) 3(x) = 3x
2
m(m 1)x
m 2
+ 11xmx
m 1
3x
m

= x
m
[3m(m 1) + 11m 3]
= x
m
(3m
2
+ 8m 3)
For that to be equal to zero, we'd need for that second factor to equal zero:
3m
2
+ 8m 3 = 0
(3m 1)(m + 3) = 0
m = 1/3, 3.
So there are two possible solutions: y = x
1/3
and y = x
3
.

{Note that this second one's interval would have to exclude x = 0.}

(b) x
2
''(x) x'(x) 5(x) = x
2
m(m 1)x
m 2
xmx
m 1
5x
m

= x
m
[m(m 1) m 5]
= x
m
(m
2
2m 5)

x
y
c = 2
c = 1
c = 1
c = 2
c = 0
So, using the quadratic formula, we get that

.
So there are two possible solutions:

and

.

{Here again, that second solution is defined on either (, 0) or (0, ), not for all real numbers.}

22. Given: the differential equation d
2
y/dx
2
+ dy/dx 2y = 0 and a proposed solution (x) = c
1
e
x
+ c
2
e
2x
.
Verification:
'(x) = c
1
e
x
2c
2
e
2x

''(x) = c
1
e
x
+ 4c
2
e
2x

''(x) + '(x) 2(x) = c
1
e
x
+ 4c
2
e
2x
+ c
1
e
x
2c
2
e
2x
2(c
1
e
x
+ c
2
e
2x
)
= e
x
(c
1
+ c
1
2c
1
) + e
2x
(4c
2
2c
2
2c
2
) = 0 + 0 = 0
(a) From the initial conditions y(0) = 2 and y'(0) = 1 we get two equations:
(0) = c
1
+ c
2
= 2 and
'(0) = c
1
2c
2
= 1
Solving the first for c
1
gives c
1
= 2 c
2
.
Substituting that into the second equation gives 2 c
2
2c
2
= 1, and hence c
2
= 1/3.
Then c
1
= 2 1/3 = 5/3.

Answer: c
1
= 5/3, c
2
= 1/3.

(b) From the initial conditions y(1) = 1 and y'(1) = 0 we get two equations:
(1) = c
1
e + c
2
e
2
= 1 and
'(1) = c
1
e 2c
2
e
2
= 0
Solving the second for c
1
gives c
1
= 2c
2
e
3
.
Substituting that into the first equation gives 2c
2
e
2
+ c
2
e
2
= 1, which yields c
2
= e
2
/3.
That means that c
1
= 2(e
2
/3)e
3
= 2/(3e).

Answer: c
1
= 2/(3e), c
2
= e
2
/3.

23. For the initial-value problem dy/dx = y
4
x
4
, y(0) = 7 we have
f (x, y) = y
4
x
4
and f /y = 4y
3
, both of which are also continuous on the entire plane.
So Theorem 1 does imply the existence of a unique solution.

27. For the initial-value problem y(dy/dx) = x, y(1) = 0 we first solve for dy/dx: dy/dx = x/y.
But the function f (x, y) is not even defined for the initial-value point (1, 0).
So Theorem 1 does not imply the existence of a unique solution.

NOTE: Theorem 1 does not say that there is not a unique solution to the initial-value problem in this case;
it merely does not say there is one. {Difference: I say there isn't vs. I don't say there is.}

28. This time we have the initial-value problem

, y(2) = 1.
The function f (x, y) is continuous in an open rectangle containing (2, 1).
The first partial

, however, is not continuous at any point whose y-cordinate is 1.

So Theorem 1 says nothing it does not guarantee the existence and uniqueness of a solution.

29. (a) First consider the constant function
1
(x) 0.
This function's first derivative is
1
'(x) 0 as well and so we do get
1
'(x) = 3
1
(x)
2/3
.
Thus this function is a solution to the differential equation.
And it certainly satisfies the initial condition:
1
(2) = 0.
So y =
1
(x) is a solution to the IVP.

For the second function
2
(x) = (x 2)
3
we have
2
'(x) = 3(x 2)
2
= 3
2
(x)
2/3
.
And
2
(2) = (2 2)
3
= 0.
So this too is a solution to the given IVP.

(b) The very similar IVP y' = 3y
2/3
, y(0) = 10
7
does have a unique solution in a neighborhood of x = 0.
You can be sure of this by Theorem 1 since both f (x, y) = 3y
2/3
and the partial f /y = 2y
1/3
are
continuous on the rectangle R: < x < , 0 < y < which contains the point (0, 10
7
).

30. First I want to check that (0, 1) is on the curve given by x + y + e
xy
= 0.
So I'll plug in: 0 + (1) + e
0(1)
= 0 1 + 1 = 0. Good.
Next, for the function G(x, y) = x + y + e
xy
we have the partial derivatives:
G/x = 1 + ye
xy
is continuous on the entire plane, and
G/y = 1 + xe
xy
is continuous on the entire plane as well.
Next we check the value of G
y
(0, 1) = 1 + 0 = 1 0.
So all the conditions of the Implicit Function Theorem are met.
Therefore there exists a differentiable function y = (x) defined on some interval I = (, ) that satisfies
the relation G(x, (x)) = 0 for < x < .

31. Given: y(dy/dx) 4x = 0.
(a) Solving for dy/dx gives dy/dx = 4x/y (which, you will note, assumes that y 0).
For any initial condition y(x
0
) = 0, the function f (x, y) is not continuous at (x
0
, 0), so Theorem 1 does
not guarantee anything.
(b) For x
0
0 and y(x
0
) = 0 we'd need y(x
0
)y'(x
0
) 4x
0
= 0 4x
0
= 4x
0
to be equal to zero, which is
impossible. So the IVP couldn't possibly have a solution in any neighborhood of x
0
.
(c) One solution to the IVP y(dy/dx) 4x = 0, y(0) = 0 is y = 2x.
And another is y = 2x.
Verifications:
For y = 2x, y' = 2 and so yy' 4x = (2x)(2) 4x = 0, and also y(0) = 0,
For y = 2x, y' = 2 and so yy' 4x = (2x)(2) 4x = 0, and also y(0) = 0.