Solutions 2

3.4.11 (Page 83)

Part a) Fixed points are integral multiples of π.
Vector Field

−2π −π 0 π 2π

Part b) When r > 1, the absolute value of x is always greater than the absolute value of
sin x unless x = 0, which is the only fixed point. The derivative of (rx − sin x) is r − 1 at
point x = 0, since it’s positive x = 0 is unstable.
Vector Field

Part c) As r decreases, the graph of y = rx has more intersections with the graph of
y = sin x, i.e. more fixed points are created. At an intersection point x = c, if (y = sin x)
crosses (y = rx) from the below, then x = c is unstable, and vice versa. When (y = rx)
touches (y = sin x) at a new point, a bifurcation occurs, and after the bifurcation the smaller
fixed point will be unstable.

Notice that 0 is always a fixed point and it changes from unstable to stable as r passes 1.
We conclude that when r decreases from ∞ to 0, there is a subcritical pitchfork bifurcation
at r = 1 and saddle-node bifurcations when 0 < r < 1.
Part d) When r ≪ 1, y = rx touches y = sin x at approximately the peaks of its graph,
2π
i.e. x = π2 + 2kπ, where k is a positive integer. Therefore bifurcations occur near r = 4k+1 .
Part e) When r further decreases, two loci of fixed points will merge and vanish, which
is clear if you stare the figure in part c for a while. These are also saddle-node bifurcations,
shown below.

3.6.2 (Page 86)

When h = 0 this system is the same as the one in Section 3.2. As h varies, the curves in
the following pictures move vertically.
1
2

r<0 r=0 r>0

V V V

x x x

Part a) If h < 0, the above curves are shifted down. This could affect the number of
fixed points when r is close to 0. In the bifurcation diagram below, the system has no fixed
point in the middle.
Bifurcation Diagram, h<0
x

unstable
r

stable

If h = 0, this system has a transcritical bifurcation, please see Section 3.2.

Bifurcation Diagram, h=0
x

stable r
unstable
 3

If h > 0, the curves are shifted up, there will always be two zeros. Notice that the curves
crosse the x-axis from below at the smaller fixed point, so it’s unstable.

Bifurcation Diagram, h>0

x

stable
r

unstable

Part b) Bifurcation occurs precisely if (h + rx − x2 ) has only one solution. The quadratic
has discriminant (r 2 + 14 h), which distinguishes qualitatively different vector fields. When
it’s positive, the quadratic has two solutions, of which the smaller one is an unstable fixed
point; when it’s 0, bifurcation occurs; when it’s negative, there’s no fixed point.

0.5

−0.5

−1

−1.5

−2

−2.5

−3
−3 −2 −1 0 1 2 3

Transcritical

Part c) There are two regions: the region above the curve h = − 14 r 2 , and the region
below it.
4

Upper Region Lower Region

V V

x x

3.7.3 (Page 89)

N
Part a) First let x = K, then the system becomes
d(Kx)
= rKx(1 − x) − H,
dt
notice that d(Kx) = Kdx, (rK) is a constant,
1 dx H
= x(1 − x) − .
r dt rK
H
Let τ = rt, h = rK , we get the desired dimensionless form.
Part b) The maximum of x(1 − x) is 41 , thus x(1 − x) − h can be viewed as a parabola
y = x(1 − x) intersecting y = h. The vector fields are
Vector Fields

h<0.25

h>0.25

Part c) This is equivalent to solving x(1 − x) − h = 0, which can be done either by looking
at the graph of y = x(1 − x), or by writing the above as
1 1
(x − )2 = − h.
2 4
1
It’s then clear that there are two fixed points when h < 4 and no fixed points when h > 14 .
The critical value hc = 14 is a saddle-node bifurcation.
 5

Part d) If h < hc , let xu be the unstable fixed point and xs the stable one. To avoid the
model’s silliness mentioned at the end of the problem, we assume xu > 0. h < hc means
the amount of fishing is moderate. If initially the fish population is small (x < xu ), then
eventually there will be no fish left. If x > xu , the population will stabilize at xs , i.e. both
the fish and the fishermen are happy. If h > hc , then the fishermen are asking too much,
eventually there will be no fish left.

3.7.5 (Page 90)

Part a) First take g = k4 x, the system becomes

x2
k4 ẋ = k1 s0 − k2 k4 x + k3 ,
1 + x2

k1 s 0 k2 k4 k3
then let s = k3 , r= k3 and τ = k4 t.
x
Part b) If s = 0, the two fixed points correspond to the solutions of 1+x2 = r. The graph
x
of function y = 1+x 2 looks like

When r > 0 and r small, there are always two points on y = r. To determine rc , we can
look at the maximum of y. Since

1 − x2
y′ = ,
(1 + x2 )2

the maximum is achieved at x = 1, thus rc = 21 .

x 2
Part c) If r > rc ,the graph of y = −rx + 1+x 2 looks like the solid line in the picture below
6

When s increases, the graph becomes the dashed line. If we start at g(0) = 0, then g
increases as t increases. Now let s go back to 0, then we go back to the solid line, and g
will also return to 0 since it is the only fixed point and it’s stable.
If r < rc , we have the graphs
y

As s increases, the gene product g begins to accumulate. If s returns to 0 shortly after

it takes off, then x has a value between 0 and the smaller positive fixed point, which is
unstable. Therefore x will go back to 0 again, i.e. the gene will turn off. But if s stays
positive long enough, so that x can accumulate until it exceeds the smaller fixed point,
then even s goes back to 0, g will still be pushed to the larger fixed point, i.e. the gene is
switched on.
Part d) If r >> rc , there will always be one stable fixed point, no bifurcation occurs. If
x2
r < rc , notice that in the above graph, f (x) = −rx + 1+x 2 has a minimum fmin , a saddle-

node bifurcation will occur at s = −fmin . To get the parametric form of r and s, recall that
f achieves its minimum when
r + 2rx2 + rx4 − 2x
f ′ (x) = − = 0.
(1 + x2 )2
2x x2 (1−x2 )
Therefore r = (1+x2 )2 , and s = −fmin = (1+x2 )2 .
The interesting case is when r ≥ rc but not too large, which corresponds to the figure
below
 7

r = −0.55

At s = 0 there is only one fixed point x = 0, but as s increases, there will be three fixed
points.
All the bifurcations in this part are saddle-node bifurcations.
Part e) The graph of (r, s) looks like a triangle,
0.5

0
r

−0.5

−1
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1

1 fixed point

We’re only interested in the first quadrant of this diagram.

4.1.2 (Page 113)

The graph of θ̇ = 1 + 2 cos(θ) looks like

We can then draw its phase portrait

8

2π 4π
The stable fixed point is θ ∗ = 3 while the unstable fixed point is θ ∗ = 3 .

4.1.5 (Page 113)

The graph of θ̇ = sin(θ) + cos(θ) looks like

We can then draw its phase portrait

3π 7π
The stable fixed point is θ ∗ = 4 while the unstable fixed point is θ ∗ = 4 .

4.3.3 (Page 113)

The graphs of θ̇ = µ sin(θ) − sin(2θ) look like

θ
 9

where, in ascending order, µ = −2, −1, 0, 1, 2. We can draw the phase portraits as follow:

µ ≤ −2 −2 < µ < 0 µ=0

0<µ<2 µ≥2

Hence we can see that for µ = −2, the system undergoes a supercritical pitchfork bifur-
cation at θ ∗ = π while for µ = 2, the system undergoes a subcritical pitchfork bifurcation
at θ ∗ = 0 (see the bifurcation diagram below).
θ

θ=π

µ = −2 θ=0 µ=2 µ

θ = −π
10

sin(θ)
4.3.4 (Page 113) We can draw the phase portraits for θ̇ = µ+cos(θ) as follow:

µ = <−1 µ = −1 −1< µ < 0

0<µ<1 µ=1 µ>1

For −1 < µ < 1, there exists two angles θ1 , θ2 (denoted as crosses in the above figures)
such that θ̇|θ=θ1,2 are not defined. They are called attractors or finite time singularity since
the flow is toward them but NOT fixed points (as the dynamics is not well defined at
those points). As µ → −1− , the stable fixed point at θ ∗ = 0 undergoes a supercritical
bifurcation at µ = −1 and produce the two attractors. On the other hand, as µ → 1− ,
two attractors move toward θ ∗ = π, indicated by the lines with crosses. It undergoes a
subcritical bifurcation at µ = 1 and produce a stable fixed point at µ = 1. Notice that at
these critical values µ = ± , θ ∗ = π, 0 are not longer fixed points but attractors themselves
(see the bifurcation diagram below).
 11

θ=π

µ = −1 θ=0 µ=1 µ

θ = −π

4.5.3 (Page 116)

Part a) When µ is slightly less than 1, the graph of f (θ) = µ + θ looks like

We can then draw its phase portrait

rest state

The stable fixed point is the globally attracting rest state, when θ passes the unstable
fixed point, i.e. the ”threshold”, the system will go almost all the way around the circle
before it returns to the ”rest state”.
12

Part b) The dotted line in the figure below correponds to the threshold. If initally θ is
on the right of the threshold, where V is above the dotted line, then V will reach 1 before
it returns to the rest state.
t
1

0.5

−0.5

−1
V(t)
0 1 2 3 4 5 6 7 8