You are on page 1of 7

MATH 1280 HOMEWORK 2 SOLUTIONS

Section 2.5
Problem 5: Consider the initial value problem x = |x|p/q , x(0) = 0, where p and q are positive integraers
with no common factors.
(a) Show that there are an infinite number of solutions if p < q.
(b) Show that there is a unique solution if p > q.
Solution: First note that since x 0 and x(0) = 0, it is only necessary to consider the case x 0 in our
calculations. The given system is separable, so we can integrate it directly to find


Z
p
xp/q+1

(t t0 )q/(qp)
t + C = |x|p/q dx =
1 pq
q
(a) If p < q, then a piecewise x defined as
(
0

x= 
1 pq (t t0 )q/(qp)

if t < t0
if t t0

is a solution for any nonnegative value of t0 . Thus, there are an infinite number of solutions. (This is unsurprising, since the vector field is not Lipschitz for p < q, so we know that we couldnt guarantee uniqueness through
the regular means.)
(b) If q < p, the solution we found above by integrating is negative, and thus not a valid solution. Alternatively, we can note that the vector field is Lipshitz in this case, and we can apply the Picard-Lindelof theorem
to get uniqueness.

Date: February 23, 2014.


1

MATH 1280 HW2

Section 2.7
3

Problem 6: For x = r + x x , plot the potential function V (x) and identify all the equilibrium points and
their stability for various values of r.
dV
x4
x2
Solution: Since x = f (x) =
, we find V (x) =

rx. There are 6 qualitatively different


dx
4
2
2
2
2
situations: |r| < , |r| = , |r| > .
3 3
3 3
3 3
A cartoon of the first three situations is sketched below, with the opposite inequalities corresponding to the
mirror image of the cartoons. The minima of V (x) correspond to stable fixed points of the system, and maxima
correspond to unstable fixed points.
V

2
|r| <
3 3
x

2
r=
3 3
x

2
r>
3 3
x

MATH 1280 HW2

Section 3.1
Problem 1: Collect the bifurcation information for x = 1 + rx + x2 .
Solution: Fixed points occur at the roots of 1 + rx + x2 , giving x =
that there are three cases of interest.

r2 4

. We see immediately

Case (i): |r| < 2. In this case there are no real values for x .
Case (ii): |r| = 2. In this case there is a single fixed point at x = 2r . This is where the saddle node
bifurcation occurs.
Case (iii): |r| > 2. In this case both possibilities for x are valid. Checking the derivative of 1 + rx + x2 will
show that one is stable and one is unstable, so I leave that to you.
These cases are shown below for r > 0.
|r| < 2

|r| = 2

|r| > 2

The bifurcation diagram is shown below. Stable branches are shown as solid lines, unstable branches are
shown as dashed lines.
x

r
2

MATH 1280 HW2

x
x
Problem 4: Collect the bifurcation information for x = r +
.
2 1+x

1 2r 4r2 12r + 1
Solution: In terms of r, the fixed points are x =
. We see that the number of solu2
2
tions will be dependent on the discriminant, 4r 12r+1. The three qualitatively different cases are shown below.
x

r chosen so the discriminant is


negative no solutions

r chosen so the discriminant is


zero a single solution, saddle
node bifurcation
x

r chosen so the discriminant is


positive two real solutions

3 2
The saddle node bifurcations occur precisely when 4r 12r +1 = 0 rc =
. The bifurcation diagram
2
is given below.
x
2

r
rc

rc

MATH 1280 HW2

Section 3.2
Problem 2: Collect the bifurcation information for x = rx ln(1 + x).
Solution: It is simplest to view the fixed points as intersections of rx and ln(1 + x). There are three
qualitatively different cases of interest.
Case (i): r < 0. There is only a single intersection, at the origin. It is stable.
Case (ii): 0 < r < 1. There are two intersections, one at the origin and one for a positive value of x. The
origin is stable, the positive point is unstable.
Case (iii): r > 1. There are again two intersections, one at the origin and one for a negative value of x. The
origin is unstable, the negative point is stable. We see that the origin has undergone a transcritical bifurcation
at r = 1.
Vector fields for the three cases are shown below, as is a bifurcation diagram.
r<0

0<r<1

r>1

MATH 1280 HW2

Problem 4: Collect the bifurction information for x = x(r ex ).


Solution: We notice that x = 0 is always a fixed point, and the other fixed point will occur at x = ln r.
Immediately we see that x = 0 is the only fixed point for r 0.
This is one of the few cases where I think its easiest to see stability directly from the calculation rather than
from a sketch of the vector field, so lets calculate directly.
If we let f (x) = x(r ex ), then f 0 (x) = r ex (1 + x).
For x = 0, f 0 (0) = r 1, so this will be stable for r < 1 and unstable for r < 1.
For x = ln r, f 0 (ln r) = r ln r. This will be unstable (f 0 (x ) > 0) when r < 1, and stable (f 0 (x ) < 0) when
r > 1.
The qualitatively different vector fields are below, as well as the bifurcation diagram.
r<0

0<r<1

r>1

x=0

x=0

x=0

Problem 6: The problem statement is fairly involved, so I wont reproduce it here.


Solution:
(a) Given x = X + bX 3 + O(X 4 ), we have x3 = X 3 + O(X 4 ). Then x + cx3 = X + bX 3 + cX 3 + O(X 4 ), so
choosing c = b gives the appropriate inversion.
(b) Using the result from part (a) and collecting to O(x4 ) we have

X = x bx3 + O(x4 )
X 2 = x2 + O(x4 )
X 3 = x3 + O(x4 )

Then from the problem statement we have

X = RX X 2 + aX 3 + O(X 4 )
= Rx Rbx3 x2 + ax3
Applying the chain rule to x and writing everything in terms of x, we have

MATH 1280 HW2

x = X + 3bX 2 X + O(X 4 )
= Rx Rbx3 x2 + ax3 + 3bx2 (Rx Rbx3 x2 + ax3 ) + O(x4 )
= Rx x2 + (2bR + a)x3 + O(x4 )
a
gives k = 0.
(c) Choosing b =
2R
(d) The assumption that R 6= 0 was definitely needed for this procedure to work. As we can see in part (c),
our choice of b has an R in the denominator, so R = 0 causes a problem. Put another way, if R = 0 the cubic
term in our transformation is ax3 , and we no longer have a free parameter to try and eliminate the a term.

Jeff Dunworth, Thackeray 517


E-mail address: jbd20@pitt.edu

You might also like