763310A ANALYTICAL MECHANICS Exercise 7 Autumn 2016

1. Solution:

Let us introduce two new auxiliary variables a(t) and b(t). Because the
length of the rope is constant, we have in the system constraint

a(t) + b(t) + d(t) = constant. (1)

the altitude of the monkey is h − a(t) and that of the banana h − b(t).
Now the kinetic energy of the system is
1 1
T = mḃ2 + M ȧ2
2 2
and the total potential energy of the system is
V = mg(h − b) + M g(h − a).

Thus the system has Lagrange function

1 1
L = T − V = mḃ2 + M ȧ2 + mg(b − h) + M g(a − h). (2)
2 2
Now we have two ways to do the calculations.

with dierential constraint

From equation 1 we get a dierential constraint

da + db + dd = 0. (3)
The equations of motion are attained by the Lagrange equation
  m
d ∂L ∂L X
− = λl ali , i = 1 . . . m.
dt ∂ q̇i ∂qi l=1

Because we have two variables qi = a, b, then n = 2 and only one con-

straint meaning m = 1. The coecients in the Lagrange equation are
achieved by comparing the sum and the dierential constraint 3. The
coecients are a11 = 1, a12 = 1 and a1t = d(t)
˙ (d is some known function
of time, and dd(t) = d(t)dt
˙ ). Now we can make calculations:
 
d ∂L ∂L
− = a11 λ1 = λ1
dt ∂ ȧ ∂a
 
d ∂L ∂L
− = a12 λ1 = λ1
dt ∂ ḃ ∂b

M ä − M g = λ1
mb̈ − mg = λ1
M ä − M g = mb̈ − mg, (4)
where we eliminated the Lagrange multiplier λ1 . Because from the dif-
ferential constraint
da + db + dd = 0
⇒ ȧ + ḃ + d˙ = 0
⇒ ä + b̈ + d¨ = 0
⇒ b̈ = −ä − d¨
 we get from equation 4
¨ − mg
M ä − M g = −m(ä + d)
(M + m)ä + md¨ = (M − m)g

(M + m)z̈ − md¨ = (m − M )g, (5)

where we noticed z = h − a ⇒ ä = −z̈ .

with holonomic constraint

If we denote the length of the rope with l0 , we get from the equation 1
a holonomic constraint

a(t) + b(t) + d(t) = l0 = constant. (6)

We can use this constraint to solve e.g. b(t):

b(t) = l0 − a(t) − d(t).

Now we can substitute the solved b to Lagrange function (eq. 2)

1 ˙ 2 + 1 M ȧ2 + mg(l0 − a − d − h) + M g(a − h)
L = L(a, ȧ) = m(ȧ + d)
2 2
and the Lagrange equation (note that we have now only one coordinate
a)
 
d ∂L ∂L
− =0
dt ∂ ȧ ∂a
 
d ˙ 2
m(ȧ + d) + M ȧ + mg − M g = 0
dt
(m + M )ä + md¨ + (m − M )g = 0, z = h − a
(M + m)z̈ − md¨ = (m − M )g

So we did get the same result. Let's solve the equation of the motion:
(M + m)z̈ − md¨ = (m − M )g
m ¨ m−M
z̈ = d+ g
m+M m+M
ż(t) − ż(0) =
m ˙ ) + m − M gt
˙ − d(0)
(d(t)
|{z} m + M |{z} m+M
=0 =0
m 1m−M 2
z(t) − z(0) = d(t) + gt ,
m+M 2m+M

where we used the initial values d(0)

˙ = ż(0) = 0. If M = m, then it
holds that z(t) = 2 d(t) (+z0 that is irrelevant for now), and thus the
1
 vertical distance is
z(t) − (h − b(t)) = z(t) − h + l0 − a(t) − d(t)
= z(t) − a(t) − d(t) + (l0 − h)
1 1
= d(t) − (h − d(t)) − d(t) + (l0 − h)
2 2
= l0 − 2h = constant.

2. Solution:
The Lagrangian is familar
1 1
L = (m1 + m2 )ẋ2 + m2 lẋφ̇ cos φ + m2 l2 φ̇2 + m2 gl cos φ.
2 2
It is easy to notice that the Lagrangian does not depent on the coordinate
x. This means that the corresponding momentum px is conserved:
∂L
px = = (m1 + m2 )ẋ + m2 lφ̇ cos φ.
∂ ẋ
and ṗx = 0 (check the Lagrange equation). Because there is explicit
dependence on the coordinate φ, the momentum
∂L
pφ = = ml2 φ̇ + m2 lẋ cos φ
∂ φ̇
is not a constant of motion. The second important conserved quantity
is the Hamiltonian H :
X
H= q̇i pi − L
i

= (m1 + m2 )ẋ2 + 2m2 lẋφ̇ cos φ + ml2 φ̇2

1 1
− [ (m1 + m2 )ẋ2 + m2 lẋφ̇ cos φ + m2 l2 φ̇2 + m2 gl cos φ]
2 2
1 1
= (m1 + m2 )ẋ2 + m2 lẋφ̇ cos φ + m2 l2 φ̇2 − m2 gl cos φ.
2 2
Because we do not have explicit time dependence in the Lagrangian, our
Hamiltonian is constant:
∂L dH
=0⇒ = 0.
∂t dt
The last thing to notice is that the constant Hamiltonian is equivalent
with the conservation law for energy. Because our potential (and con-
strains) is independent of velocity and time the Hamiltonian is the same
as the total energy in the system. One can conrm this from previous
exercise 4.1 where we have
1 1
T = T1 + T2 = m1 ẋ + m2 [ẋ2 + 2lẋφ̇ cos φ + l2 φ̇2 ]
2 2
 and
V = −m2 gl cos φ.

3. Solution:
The Lagrangian is
s
ṙ2 + r2 φ̇2 k
L = −mc2 1− + .
c2 r
It is good to notice that this Lagrangian is not similar to the others in
this course L 6= T − V . This is because in the relativistic case one has
to go a little bit dierent way to variate the action when one uses the
Hamilton principle. More about this in the theoretical course, Classical
Field Theory. Now the momenta are
∂L mc2 r2 φ̇ mr2 φ̇
pφ = =q · 2 =q
∂ φ̇ 2 2 φ̇2 c 2 2 φ̇2
1 − ṙ +r
c2
1 − ṙ +r
c2

∂L mc2 ṙ mṙ
pr = =q · 2 =q .
∂ ṙ 2 2 2 c ṙ2 +r2 φ̇2
1 − ṙ +r
c2
φ̇
1 − c2

Because of ∂φ L = 0, momentum pφ is a constant of motion but pr is not:

∂L mc2 rφ̇2 k
=q · 2 − 2.
∂r 2 2 φ̇2 c r
1 − ṙ +r
c2

Let's denote E0 = mc2 and

"s #−1
ṙ2 + r2 φ̇2
γ= 1− .
c2

Thus
ṙ r2 φ̇
pr = γE0 · and pφ = γE0 ·
c2 c2
and also
∂L rφ̇2 V (r)
= γE0 · 2 + .
∂r c r
Also the Lagrangian looks more simple
k
L = −γ −1 E0 + .
r
The Lagrange equations are
dpφ
=0
dt
 dpr rφ̇2 V (r)
= γE0 · 2 + .
dt c r
Because ∂t L = 0 ⇒ Ḣ = 0, our Hamiltonian is conserved. By denition
X
H= q̇i pi − L
i

ṙ2 r2 φ̇2 k
= γE0 2
+ γE 0 2
− [−γ −1 E0 + ]
c c r
 ṙ2 + r2 φ̇2  k
= γE0 2
+ γ −1 E0 −
c r
−2 −1 k
= γE0 (1 − γ ) + γ E0 −
r
k
= γE0 −
r
that agrees with relativity, like the momenta do. In the classical limit
v << c the Hamilton reduces to the classical case
k 1 k
H = γE0 − ≈ mc2 + mv 2 − .
r 2 r
Because the potential is independent on time and velocity, the classical
Hamiltonian should be the total energy. This agrees with the relativistic
Hamiltonian when one gets to classical limit: kinetic energy is the normal
1
T = mv 2
2
but the potential energy also include the rest energy
k
Vclass = mc2 − = E0 + V.
r
4. Solution:
Kinetic energy in spherical coordinates is
1
T = m(ṙ2 + r2 φ̇2 sin2 θ + r2 θ̇2 ).
2
In the system it holds
r = a = constant ⇒ ṙ ≡ 0
φ̇ = ω = constant ⇒ φ = ωt + φ0

meaning that the system has kinetic energy

1 1 1
T = m(a2 ω 2 sin2 θ + a2 θ̇2 ) = ma2 θ̇2 + ma2 ω 2 sin2 θ
2 2 2
Potential energy is
V = mga cos θ.

Now the Lagrangian is

1 1
L = T − V = ma2 θ̇2 + ma2 ω 2 sin2 θ − mga cos θ
2 2
The total energy is
1 1
E = T + V = ma2 θ̇2 + ma2 ω 2 sin2 θ + mga cos θ.
2 2
For the Hamiltonian one needs to dene canonical momentum
∂L
p= = ma2 θ̇.
∂ θ̇
From the denition the Hamiltonian is
X
H= q̇i pi − L
i

= θ̇p − L
1 1
= ma2 θ̇2 − ( ma2 θ̇2 + ma2 ω 2 sin2 θ − mga cos θ)
2 2
1 2 2 1 2 2 2
= ma θ̇ − ma ω sin θ + mga cos θ.
2 2
Now one can clearly see that the Hamiltonian is not the same as the
total energy in the system H 6= E . Because the Lagrangian does not
explicitly depend on time, the Hamiltonian is conserved :
∂L dH
=0⇒ = 0.
∂t dt
Due to the constraint in the system, r = a = constant and φ̇ = ω =
constant one ends up with the Lagrangian that does not depend on time
(explicitly). As mentioned in the lectures, in this kind of case one has a
conservation law H = constant, but the Hamilton does not have to be
same as the total energy. Let's prove this conclusion:
dH d 1 1
= ( ma2 θ̇2 − ma2 ω 2 sin2 θ + mga cos θ)
dt dt 2 2
= ma2 θ̈θ̇ − a2 ω 2 sin θ cos θθ̇ − mga sin θθ̇
= θ̇[ma2 θ̈ − a2 ω 2 sin θ cos θ − mga sin θ]
| {z }
=0
= 0,

where we use the Lagrange equation

 
d ∂L ∂L
− =0
dt ∂ θ̇ ∂θ
⇔
ma2 θ̈ − a2 ω 2 sin θ cos θ − mga sin θ = 0.

It is easy to see that

dE
6= 0.
dt
This implies that the energy in the system is not conserved. We still have
the conservation law of energy? The solution is that our system is not
isolated but it exchanges energy with the environment. The environment
is acting to the system with the force that keeps the angular velocity
as constant (note that our Hamiltonian regards this). Notice that the
combined system (environment + our system) is isolated meaning

Esystem + Eenviroment = constant + H.

So our conservation law of energy holds in the combined system.