MSE, HS 2021

OrdDiff, 31.01.2022

Solution Final exam

Problem 1 a) Polynomial disturbance, 0 ∈ R, so Ax + Bx2 + Cx3

b) Trig function, j ∈ R, so ys = x · (A cos x + B sin x)

c) Exponential function, −1 ∈ R, double, so ys = Ax2 exp(−x)

d) Combination of polynomial and exponential, 0 ∈

/ R, 1 ∈
/ R, so ys = A + Bx +
2
Cx + D exp(x)

Problem 2 a) The homogeneous solution is yh = c1 exp(2t) + c2 exp(−2t). The trial

function gives

y 0 = exp(jx) (ja + b + jbt) , y 00 = − exp(jx) (a + b t − 2jb)

y 00 − 4y = − exp(jx) (5 a + 5 b t − 2jb)

from which a = −2j and b = −5, so

y = yh − exp(jx) (5 t + 2j)

b) Taking the imaginary part on both sides of the original equation gives

y = yh − 2 cos (t) − 5 t sin (t)

Problem 3 a) Since ẋ + ẏ + ż = 0, the quantity n = x + y + z is constant and we

can replace z by z = n − x − y. The reduced system is

ẋ = −αx + γ · (n − x − y) = −(α + γ)x − γy + γn

ẏ = αx − βy

or in matrix form

   
−α − γ −γ γn
~x˙ = ~x +
α −β 0
 b) From x˙∗ = ~0, the constant solution is
 
γN β
x=
αβ + αγ + βγ α

Since det(A) > 0 and tr(A) < 0, the unique equilibrium ~0 of the homogeneous
equation is asymptotially stable, the homogeneous part of the solution will even-
tually converge to ~0.

c) Eigenvalues are λ = −4 ± j, . . .

Problem 4 a) v = y −1/3 means y = v −3 and so y 0 = −3v −4 v 0 . substituting, we get

−3x(v −4 v 0 ) + 6v −3 = 3xv −4 and after some manipulations

2v
v0 =
−1
x
R R
Using the variation of constants formula with 2/xx. = 2 ln |x| and exp 2/xdx =
x2
Z
v(x) = x 2
x−2 x. = x2 (1/x + c) = cx2 + x

So
1
y=
(c x2 + x)3

b) We need a potential, so

Z
2x2 y + 4y 3 y. = x2 y 2 + y 4 + c(x)
Z
2xy 2 + 3x2 x. = x2 y 2 + x3 + d(y)

Thus, the potential is F (x, y) = x2 y 2 + x3 + y 4 and solutions are contour lines of

F , i.e. x2 y 2 + x3 + y 4 = c

Problem 5 a) To determine fixed points, we have to solve the equation rx = sin(x),

resp. we have to find the intersection of the curves y = rx and y = sin(x). For r > 1,
the origin x = 0 is clearly the only intersection point of the two curves (since the
slope of y = rx is y 0 = r > 1 and the slope of the sine curve is y 0 = cos(x) ∈ [−1, 1]).
Stability: For f (x) = rx − sin(x), we have f 0 (x) = r − cos(x) and therefore f 0 (0) =
r − cos(0) = r − 1 > 0. Therefore, x∗ = 0 is a unstable fixed point of the system.
b) By inspection of the two graphs y = sin(x) and y = x2 (or by applying the interme-
diate value theorem to y = sin(x) − x2 ), we see that there are 3 fixed points x∗1 = 0,
x∗2 > 0 and x∗3 = −x∗2 < 0, with π2 < x∗2 < π.
Graphically:

Stability: f (x) = 1
2 − sin(x), we have f 0 (x) = 1
2 − cos(x) and therefore:
• x∗1 = 0: f 0 (0) = 1
2 − cos(0) = 12 − 1 = − 12 < 0, i.e. stable fixed point.
• x∗2 > 0: f 0 (0) = 1
2 − cos(x∗2 ) > 21 > 0, i.e. unstable fixed point.
• x∗3 > 0: f 0 (0) = 1
2 − cos(x∗3 ) > 21 > 0, i.e. unstable fixed point.
c) The transition from the situation described in a) to the situation described in b),
i.e. a change from 1 stable fixed point to 3 fixed points with 1 unstable and 2 stable
ones is a subcritical pitchform bifurcation.
Bifurcation diagram:

d) As r decreases further towards 0, the number of intersection points of the two

curves y = rx and y = sin(x) grows infinitely. There is no upper limit for the
number of fixed points, and at every bifurcation two pairs of new fixed points are
created, i.e. these are all subcritical pitchfork bifurcations. The system thus has
infinitely many bifurcation points in the range 0 < r < 1.
Problem 6 a) We classify the fixed point using the quantities τ , ∆ and τ 2 − 4∆. We
have
τ = −2, ∆ = 1 + α, τ 2 − 4∆ = −4α.
We therefore get

(i) α < −1: saddle point

(ii) α = −1: non-isolated fixed point
(iii) −1 < α ≤ 0: stable node
(iv) α > 0: stable spiral

b) To find the fixed points, we solve the system of equations

x(1 − x − y) = 0
y(0.8 − 0.6y − x) = 0
and obtain the solutions
4 1 1
P1 = (0, 0), P2 = (0, ), P3 = (1, 0), P4 = ( , )
3 2 2
The Jacobi matrix of the system is in general
 
1 − 2x − y −x
J=
−y 0.8 − 1.2y − x
and for the fixed points
 
1 0
P1 : J(P1 ) = , i.e. unstable node
0 0.8
 1 
−3 0
P2 : J(P2 ) = 4 , i.e. stable node
3 − 54
 
−1 −1
P3 : J(P3 ) = , i.e. stable node
0 − 15
 1
− 2 − 12

P4 : J(P4 ) = 3 , i.e. saddle
− 12 − 10

Problem 7 a) To find fixed points, we have to solve the equation x = f (x) for
f (x) = e−x . Looking at the graphs of f1 (x) = x and f2 (x) = e−x , it is clear that
there is exactly one intersection point x∗ satisfying 0 < x∗ < 1, since f1 (0) < f2 (0)
and f1 (1) > f2 (1); moreover, f1 (x) and f2 (x) are monotonically increasing resp.
decreasing.

b) To determine the stabilty of the fixed point x∗ found in part a), we compute
∗
f 0 (x∗ ) = −e−x = −x∗ ∈ (−1, 0)

hence in particular |f 0 (x∗ )| < 1, and x∗ is a stable fixed point.

Problem 8 a) To find the fixed points, we solve the system of equations

y(1 − x2 ) = 0
1 − y2 = 0

and obtain the solutions

P1 = (1, 1), P2 = (1, −1), P3 = (−1, 1), P4 = (−1, −1)

The Jacobi matrix of the system is in general

−2xy 1 − x2
 
J=
0 −2y

and for the fixed points

 
−2 0
P1 : J(P1 ) = , i.e. stable node
0 −2
 
2 0
P2 : J(P2 ) = , i.e. unstable node
0 2
 
2 0
P3 : J(P3 ) = , i.e. saddle
0 −2
 
−2 0
P4 : J(P4 ) = , i.e. saddle
0 2

b) The system cannot have a conserved quantity because in systems with a conserved
quantity, all fixed points all of the types saddle or center. In this system, we also
have nodes, which contradicts the restriction on systems with a conserved quantity.
d
c) To check the claimed identity dt M (x(t), y(t)) = 0, we use the chain rule (assuming
the argument of | . . . | is positive; otherwise similarly):

d ∂M ∂M
M (x(t), y(t)) = · ẋ + · ẏ
dt ∂x ∂y
2 2y
= · y(1 − x2 ) −
(1 + x)(1 − x) 1 − y2
= 0

d) There is no contradiction between the results of b) and c) because the domain of

R R
the quantity M (x, y) considered in c) is not all of 2 , but rather 2 \ {±1, ±1}.
R
A conserved quantity in the sense of b) is assumed to have all of 2 as domain
R R
(as indicated in b): M : 2 → ). So, the quantity M (x, y) is not a conserved
quantity of the system in the precise meaning of the term.