Beruflich Dokumente
Kultur Dokumente
Leonard Susskind
George Hrabovsky
2014
Table of Contents
Interlude 1 – Exercise 1, 2, 3, 4, 5, 6
Lecture 2 – Exercise 1, 2, 3, 4, 5, 6, 7, 8
Interlude 2 – Exercise 1, 2, 3, 4
Lecture 3 – Exercise 1, 2, 3, 4
Interlude 3 – Exercise 1, 2
Lecture 5 – Exercise 1, 2, 3
Lecture 6 – Exercise 1, 2, 3, 4, 5, 6
Lecture 7 – Exercise 1, 2, 3, 4, 5, 6, 7
Lecture 8 – Exercise 1, 2
Lecture 10 – Exercise 1, 2, 3
Lecture 11 – Exercise 1, 2, 3, 4, 5
Appendix 1 – Exercise 1
Graphically we have already defined vectors as objects with a certain length and a certain direction.
We can imagine them to be arrows of that length and pointing in that direction.
𝐴⃗
We have also defined multiplication of a vector with a number. Multiplying by 2 makes the length
twice as long. And multiplying with a negative number −1 changes the direction.
Then we have defined addition of vectors as bringing the tail of the second vector to the top of the
first one. The new vector then is the one that starts at the tail of the first and ends at the top of the
second one.
𝐴⃗ + 𝐵
⃗⃗
Subtraction can be defined as first multiplying the second vector by −1, which changes the direction.
And consequently you can add the resulting vectors in the same way as the addition.
𝐴⃗ − 𝐵
⃗⃗ = 𝐴⃗ + (−𝐵
⃗⃗)
Since the dot product of a vector with itself is always a positive number, we can rewrite this.
|𝐴⃗| = √𝐴⃗ ∙ 𝐴⃗
𝐴⃗ ∙ 𝐵
⃗⃗ = 𝐴𝑥 𝐵𝑥 + 𝐴𝑦 𝐵𝑦 + 𝐴𝑧 𝐵𝑧 = −8 + 9 + 2 = 3
𝐴⃗ ∙ 𝐵
⃗⃗ 𝐴⃗ ∙ 𝐵
⃗⃗
𝐴⃗ ∙ 𝐵
⃗⃗ = |𝐴⃗||𝐵
⃗⃗| cos 𝜃 ⇔ cos 𝜃 = ⇔ 𝜃 = arccos ( )
|𝐴⃗||𝐵⃗⃗| |𝐴⃗||𝐵⃗⃗|
3
⇔ 𝜃 = arccos = 1,421253189 𝑟𝑎𝑑 = 81,43753893°
√14√29
𝐴⃗ (1,1,1)
⃗⃗ (2, −1,3)
𝐵
𝐶⃗ (3,1,0)
⃗⃗ (−3,0,2)
𝐷
Whenever the dot product of the vectors is equal to 0, they are orthogonal.
𝐴⃗ ∙ 𝐵
⃗⃗ = 2 − 1 + 3 = 4
𝐴⃗ ∙ 𝐶⃗ = 3 + 1 + 0 = 4
𝐴⃗ ∙ 𝐷
⃗⃗ = −3 + 0 + 2 = −1
⃗⃗ ∙ 𝐶⃗ = 6 − 1 + 0 = 5
𝐵
⃗⃗ ∙ 𝐷
𝐵 ⃗⃗ = −6 + 0 + 6 = 0
𝐶⃗ ∙ 𝐷
⃗⃗ = −9 + 0 + 0 = −9
⃗⃗ and 𝐷
The vectors 𝐵 ⃗⃗ are orthogonal.
𝐴⃗ ∙ 𝐵
⃗⃗ = |𝐴⃗||𝐵
⃗⃗| cos 𝜃
If the length of the two vectors is not 0, then the only possibility for the right side to be 0 is that the
cosine of the angle between them is 0.
𝜋 3𝜋
In what cases is cos 𝜃 = 0? Well, only in the cases where 𝜃 = 2
or 𝜃 = 2
.
𝜽(𝜶) = 𝒆𝜶 + 𝜶 𝐥𝐧 𝜶
𝑑𝜃 𝛼
= 𝑒 𝛼 + ln 𝛼 + = 𝑒 𝛼 + ln 𝛼 + 1
𝑑𝛼 𝛼
𝑓 = sin 𝑡
𝑑𝑓
= cos 𝑡
𝑑𝑡
𝑔 = 𝑓2
𝑑𝑔
= 2𝑓 = 2 sin 𝑡
𝑑𝑓
𝑑𝑔 𝑑𝑔 𝑑𝑓
= = 2 sin t cos 𝑡
𝑑𝑡 𝑑𝑓 𝑑𝑡
𝑑𝑥
= 2 sin t cos 𝑡 + sin 𝑡
𝑑𝑡
𝜽(𝜶) = 𝒆𝜶 + 𝜶 𝐥𝐧 𝜶
𝑑𝜃 𝛼
= 𝑒 𝛼 + ln 𝛼 + = 𝑒 𝛼 + ln 𝛼 + 1
𝑑𝛼 𝛼
𝑑2 𝜃 1
2
= 𝑒𝛼 +
𝑑𝛼 𝛼
𝑓 = 𝑡2
𝑑𝑓
= 2𝑡
𝑑𝑡
𝑑𝑔 𝑑 sin 𝑓 𝑑 cos 𝑓
= − = cos 𝑓 + sin 𝑓 = cos 𝑡 2 + sin 𝑡 2
𝑑𝑓 𝑑𝑓 𝑑𝑓
𝑑𝑔 𝑑𝑔 𝑑𝑓
= = 2𝑡 cos 𝑡 2 + 2𝑡 sin 𝑡 2
𝑑𝑡 𝑑𝑓 𝑑𝑡
𝑓 = 𝑡2
𝑑𝑓
= 2𝑡
𝑑𝑡
𝑔 = sin 𝑓
𝑑𝑔
= cos 𝑓 = cos 𝑡 2
𝑑𝑓
ℎ = 𝑔2
𝑑ℎ
= 2𝑔 = 2 sin 𝑡 2
𝑑𝑔
𝑑ℎ 𝑑ℎ 𝑑𝑔 𝑑𝑓
= = 2 sin 𝑡 2 cos 𝑡 2 2𝑡 = 4𝑡 sin 𝑡 2 cos 𝑡 2
𝑑𝑡 𝑑𝑔 𝑑𝑓 𝑑𝑡
𝑓 = 𝑡2
𝑑𝑓
= 2𝑡
𝑑𝑡
𝑘 = − cos 𝑓
𝑑𝑘
= sin 𝑓 = sin 𝑡 2
𝑑𝑓
𝑑𝑘 𝑑𝑘 𝑑𝑓
= = 2𝑡 sin 𝑡 2
𝑑𝑡 𝑑𝑓 𝑑𝑡
For the chain rule we should first be sure about the Leibniz notation.
Infinitely small increments of 𝑥 and 𝑦 can be represented by 𝑑𝑥 and 𝑑𝑦. So these are two numbers
that tend to get very close to 0, but never are 0. According to Leibniz the derivative of 𝑦 with respect
to 𝑥 then is a quotient of 𝑑𝑦 over 𝑑𝑥.
Now let 𝑦 be a function of a function 𝑦 = 𝑓(𝑔(𝑥)) and we want to find the derivative of 𝑦 with
respect to 𝑥. Then we can define 𝑧 to be the second function 𝑧 = 𝑔(𝑥).
𝑑𝑦 𝑑𝑓(𝑔(𝑥)) 𝑑𝑓(𝑧) 𝑑𝑓(𝑧) 𝑑𝑧 𝑑𝑓(𝑧) 𝑑𝑔(𝑥) 𝑑𝑓(𝑧) 𝑑𝑔(𝑥)
= = = = =
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑧 𝑑𝑥 𝑑𝑧 𝑑𝑧 𝑑𝑥
Throughout the proofs here we will use some definitions and properties worth mentioning.
lim cos ∆𝑡 = 1
∆𝑡→0
sin ∆𝑡
lim =1
∆𝑡→0 ∆𝑡
𝑡
1 ∆𝑡 ∆𝑡
𝑒 = lim (1 + 𝑥)𝑥 = lim (1 + )
𝑥→0 ∆𝑡→0 𝑡
ln 𝑒 = 1
ln 𝑒 𝑡 = 𝑡 ln 𝑒 = 𝑡
𝒅 𝒔𝒊𝒏 𝒕
= 𝒄𝒐𝒔 𝒕
𝒅𝒕
𝑑 𝑠𝑖𝑛 𝑡 sin(𝑡 + ∆𝑡) − sin 𝑡 sin 𝑡 cos ∆𝑡 + cos 𝑡 sin ∆𝑡 − sin 𝑡
= lim = lim
𝑑𝑡 ∆𝑡→0 ∆𝑡 ∆𝑡→0 ∆𝑡
sin 𝑡 1 + cos 𝑡 sin ∆𝑡 − sin 𝑡 cos 𝑡 sin ∆𝑡 sin ∆𝑡
= lim = lim = cos 𝑡 lim = cos 𝑡 1 = cos 𝑡
∆𝑡→0 ∆𝑡 ∆𝑡→0 ∆𝑡 ∆𝑡→0 ∆𝑡
𝒅 𝒄𝒐𝒔 𝒕
= 𝒔𝒊𝒏 𝒕
𝒅𝒕
𝑑 𝑐𝑜𝑠 𝑡 cos(𝑡 + ∆𝑡) − cos 𝑡 cos 𝑡 cos ∆𝑡 − sin 𝑡 sin ∆𝑡 − cos 𝑡
= lim = lim
𝑑𝑡 ∆𝑡→0 ∆𝑡 ∆𝑡→0 ∆𝑡
cos 𝑡 1 − sin 𝑡 sin ∆𝑡 − cos 𝑡 − sin 𝑡 sin ∆𝑡 sin ∆𝑡
= lim = lim = − sin 𝑡 lim = − sin 𝑡 1 = − sin 𝑡
∆𝑡→0 ∆𝑡 ∆𝑡→0 ∆𝑡 ∆𝑡→0 ∆𝑡
𝒅 𝒍𝒏 𝒕 𝟏
=
𝒅𝒕 𝒕
𝑑 𝑙𝑛 𝑡 ln(𝑡 + ∆𝑡) − ln 𝑡 1 (𝑡 + ∆𝑡) 𝑡 1 ∆𝑡 1 𝑡 ∆𝑡
= lim = lim ln = lim ln (1 + ) = lim ln (1 + )
𝑑𝑡 ∆𝑡→0 ∆𝑡 ∆𝑡→0 ∆𝑡 𝑡 ∆𝑡→0 𝑡 ∆𝑡 𝑡 ∆𝑡→0 𝑡 ∆𝑡 𝑡
𝑡 𝑡
1 ∆𝑡 ∆𝑡 1 ∆𝑡 ∆𝑡 1 1
= lim ln (1 + ) = ln lim (1 + ) = ln 𝑒 =
∆𝑡→0 𝑡 𝑡 𝑡 ∆𝑡→0 𝑡 𝑡 𝑡
𝑑 ln 𝑒 𝑡 𝑑𝑡 ln 𝑒 𝑑𝑡 1
= = =1
𝑑𝑡 𝑑𝑡 𝑑𝑡
Now we use the derivative of ln 𝑡 and the chain rule to calculate the same thing. We let the function
𝑓 = ln 𝑒 𝑡 and the intermediary function 𝑔 = 𝑒 𝑡 .
𝑑 ln 𝑒 𝑡 𝑑𝑓 𝑑𝑔 1 𝑑𝑒 𝑡
= =
𝑑𝑡 𝑑𝑔 𝑑𝑡 𝑒 𝑡 𝑑𝑡
We already know the result of the left side.
1 𝑑 𝑒𝑡 𝑑𝑒 𝑡
1= ⇔ = 𝑒𝑡
𝑒 𝑡 𝑑𝑡 𝑑𝑡
𝑥(𝑡) = sin 𝜔𝑡
Let’s take a closer look at the sine of an angle in radians. If you start at angle 0, it’s clear that you
need to go around the unit circle once to be back at 0 = 2𝜋. You can read off the values of the sine
on the 𝑦 axis.
𝜔𝑡 = 0, 𝑥(𝑡) = 0
𝜋
𝜔𝑡 = , 𝑥(𝑡) = 1
2
𝜔𝑡 = 𝜋, 𝑥(𝑡) = 0
3𝜋
𝜔𝑡 = , 𝑥(𝑡) = −1
2
𝜔𝑡 = 2𝜋, 𝑥(𝑡) = 0
If 𝜔 = 1, then the time for one revolution, called 𝑇, would be just equal to 2𝜋.
2𝜋
If 𝜔 = 2, then 𝑇 would be equal to .
2
1
If 𝜔 = 2, then 𝑇 would be 4𝜋.
It’s clear that the larger 𝜔 is, the faster one cycle is finished. 𝑇 is inversely proportional to 𝜔. Since 𝜔
is called the frequency, this seems logical.
𝒗𝒙 = −𝑹𝝎 𝒔𝒊𝒏 𝝎𝒕
𝒗𝒚 = 𝑹𝝎 𝒄𝒐𝒔 𝝎𝒕
𝒂𝒙 = −𝑹𝝎𝟐 𝒄𝒐𝒔 𝝎𝒕
𝒂𝒚 = −𝑹𝝎𝟐 𝒔𝒊𝒏 𝝎𝒕
𝑣⃗ ∙ 𝑎⃗ = 0
⇔ 𝑣𝑥 𝑎𝑥 + 𝑣𝑦 𝑎𝑦 = 0
𝑑 cos (𝜔𝑡 − 𝜙) 𝑑 (cos 𝜔𝑡 cos 𝜙 + sin 𝜔𝑡 sin 𝜙) 𝑑 (cos 𝜔𝑡 cos 𝜙) + 𝑑 (sin 𝜔𝑡 sin 𝜙)
𝑣𝑥 = = =
𝑑𝑡 𝑑𝑡 𝑑𝑡
= −𝜔 sin 𝜔𝑡 cos 𝜙 + 𝜔 cos 𝜔𝑡 sin 𝜙 = −𝜔 sin(𝜔𝑡 − 𝜙)
𝑑 sin (𝜔𝑡 − 𝜙) 𝑑 (sin 𝜔𝑡 cos 𝜙 − cos 𝜔𝑡 sin 𝜙) 𝑑 (sin 𝜔𝑡 cos 𝜙) − 𝑑 (cos 𝜔𝑡 sin 𝜙)
𝑣𝑦 = = =
𝑑𝑡 𝑑𝑡 𝑑𝑡
= 𝜔 cos 𝜔𝑡 cos 𝜙 + 𝜔 sin 𝜔𝑡 sin 𝜙 = 𝜔 cos(𝜔𝑡 − 𝜙)
|𝑣⃗| = √(−𝜔 sin(𝜔𝑡 − 𝜙))2 + (𝜔 cos(𝜔𝑡 − 𝜙))2 = 𝜔 √𝑠𝑖𝑛2 (𝜔𝑡 − 𝜙) + 𝑐𝑜𝑠 2 (𝜔𝑡 − 𝜙) = 𝜔
−𝜔 𝑑 sin(𝜔𝑡 − 𝜙)
𝑎𝑥 = = −𝜔 𝜔 cos(𝜔𝑡 − 𝜙) = −𝜔2 cos(𝜔𝑡 − 𝜙)
𝑑𝑡
𝜔 𝑑 cos(𝜔𝑡 − 𝜙)
𝑎𝑦 = = 𝜔 (−𝜔 sin(𝜔𝑡 − 𝜙)) = −𝜔2 sin(𝜔𝑡 − 𝜙)
𝑑𝑡
⃗⃗ = (𝒄 𝒄𝒐𝒔𝟑 𝒕, 𝒄 𝒔𝒊𝒏𝟑 𝒕)
𝒓
𝑐 𝑑 𝑐𝑜𝑠 3 𝑡
𝑣𝑥 = = 𝑐 3 𝑐𝑜𝑠 2 𝑡 (− sin 𝑡) = −3𝑐 sin 𝑡 𝑐𝑜𝑠 2 𝑡
𝑑𝑡
𝑐 𝑑 𝑠𝑖𝑛3 𝑡
𝑣𝑦 = = 𝑐 3 𝑠𝑖𝑛2 𝑡 (cos 𝑡) = 3𝑐 𝑠𝑖𝑛2 𝑡 cos 𝑡
𝑑𝑡
|𝑣⃗| = √(−3𝑐 sin 𝑡 𝑐𝑜𝑠 2 𝑡)2 + (3𝑐 𝑠𝑖𝑛2 𝑡 cos 𝑡 )2 = 3𝑐 √𝑠𝑖𝑛2 𝑡 𝑐𝑜𝑠 4 𝑡 + 𝑠𝑖𝑛4 𝑡 𝑐𝑜𝑠 2 𝑡
𝑐 𝑑 (𝑡 − 𝑠𝑖𝑛 𝑡)
𝑣𝑥 = = 𝑐 (1 − 𝑐𝑜𝑠 𝑡)
𝑑𝑡
𝑐 𝑑 (1 − 𝑐𝑜𝑠 𝑡)
𝑣𝑦 = = 𝑐 𝑠𝑖𝑛 𝑡
𝑑𝑡
𝒇(𝒕) = 𝒕𝟒
𝑡5
∫ 𝑡 4 𝑑𝑡 = +𝑐
5
𝒇(𝒕) = 𝒄𝒐𝒔 𝒕
∫ cos 𝑡 𝑑𝑡 = sin 𝑡 + 𝑐
𝒇(𝒕) = 𝒕𝟐 − 𝟐
𝑡3
∫(𝑡 2 − 2) 𝑑𝑡 = − 2𝑡 + 𝑐
3
𝒇(𝒕) = 𝒕𝟒
𝑡5
∫ 𝑡 4 𝑑𝑡 = +𝑐
5
𝑇
𝑇 5 05 𝑇 5
∫ 𝑡 4 𝑑𝑡 = − =
0 5 5 5
𝒇(𝒕) = 𝒄𝒐𝒔 𝒕
∫ cos 𝑡 𝑑𝑡 = sin 𝑡 + 𝑐
𝑇
∫ cos 𝑡 𝑑𝑡 = sin 𝑇 − sin 0 = sin 𝑇
0
𝒇(𝒕) = 𝒕𝟐 − 𝟐
2
𝑡3
∫(𝑡 − 2) 𝑑𝑡 = − 2𝑡 + 𝑐
3
𝑇
𝑇3
∫ (𝑡 2 − 2) 𝑑𝑡 = − 2𝑇
0 3
𝒂(𝒕) = 𝒕𝟒
𝑇
𝑇 5 05 𝑇 5
𝑣(𝑇) = ∫ 𝑡 4 𝑑𝑡 = − =
0 5 5 5
𝑇 5
𝑡 𝑇6
𝑟(𝑇) = ∫ 𝑑𝑡 =
0 5 30
𝒂(𝒕) = 𝒄𝒐𝒔 𝒕
𝑇
𝑣(𝑇) = ∫ cos 𝑡 𝑑𝑡 = sin 𝑇 − sin 0 = sin 𝑇
0
𝑇
𝑟(𝑇) = ∫ sin 𝑡 𝑑𝑡 = − cos 𝑇 + cos 0 = 1 − cos 𝑇
0
𝒂(𝒕) = 𝒕𝟐 − 𝟐
𝑇
𝑇3
𝑣(𝑇) = ∫ (𝑡 2 − 2) 𝑑𝑡 = − 2𝑇
0 3
𝑇
𝑡3 𝑇4
𝑟(𝑇) = ∫ ( − 2𝑡) 𝑑𝑡 = − 𝑇2
0 3 12
𝜋
2
∫ 𝑥 𝑐𝑜𝑠 𝑥 𝑑𝑥
0
𝜋
2 𝑑 sin 𝑥
=∫ 𝑥 𝑑𝑥
0 𝑑𝑥
𝜋
𝜋
2 𝑑𝑥
= 𝑥 sin 𝑥 |02 −∫ sin 𝑥 𝑑𝑥
0 𝑑𝑥
𝜋
𝜋 𝜋 2
= sin − ∫ sin 𝑥 𝑑𝑥
2 2 0
𝜋 𝜋
= − (− cos + cos 0)
2 2
𝜋
= −1
2
𝐹⃗ = 𝑚𝑣⃗
𝐹⃗
𝑣⃗ =
𝑚
𝑑𝑥(𝑡) 𝐹(𝑡)
=
𝑑𝑡 𝑚
𝑑𝑥(𝑡) 𝐹(𝑡)
∫ 𝑑𝑡 = ∫ 𝑑𝑡
𝑑𝑡 𝑚
𝑥(𝑡) 𝑡
𝑑𝑥(𝑡) 𝐹(𝑡)
∫ 𝑑𝑡 = ∫ 𝑑𝑡
𝑥(0) 𝑑𝑡 0 𝑚
𝑡
2𝑡 2
𝑥(𝑡) − 𝑥(0) = ∫ 𝑑𝑡
0 𝑚
2𝑡 3
𝑥(𝑡) − 𝜋 =
3𝑚
2𝑡 3
𝑥(𝑡) = +𝜋
3𝑚
𝐹𝑧
𝑣𝑧̇ =
𝑚
𝑑𝑣𝑧 𝐹𝑧
=
𝑑𝑡 𝑚
𝑡 𝑡
𝑑𝑣𝑧 𝐹𝑧
∫ 𝑑𝑡 = ∫ 𝑑𝑡
0 𝑑𝑡 0 𝑚
𝐹𝑧
𝑣𝑧 (𝑡) − 𝑣𝑧 (0) = 𝑡
𝑚
𝐹𝑧
𝑣𝑧 (𝑡) = 𝑣𝑧 (0) + 𝑡
𝑚
We will differentiate this equation twice. The first derivative will give us the velocity.
𝑑 𝐹𝑧 2 𝐹𝑧
𝑧̇ (𝑡) = 𝑣𝑧 (𝑡) = (𝑧0 + 𝑣𝑧 (0)𝑡 + 𝑡 ) = 𝑣𝑧 (0) + 𝑡
𝑑𝑡 2𝑚 𝑚
𝑑
𝑥̇ (𝑡) = (𝐴 cos 𝜔𝑡 + 𝐵 sin 𝜔𝑡)
𝑑𝑡
= −𝐴𝜔 sin 𝜔𝑡 + 𝐵𝜔 cos 𝜔𝑡
𝑑
𝑥̈ (𝑡) = (−𝐴𝜔 sin 𝜔𝑡 + 𝐵𝜔 cos 𝜔𝑡)
𝑑𝑡
= −𝐴𝜔2 cos 𝜔𝑡 − 𝐵𝜔2 sin 𝜔𝑡
= −𝜔2 𝑥(𝑡)
⇔ sin(𝑥𝑦)−𝑥 2 − 𝑦 2 = 0
𝜕𝑉
: 𝑦 cos(𝑥𝑦) − 2𝑥 = 0
𝜕𝑥
𝜕𝑉
: 𝑥 cos(𝑥𝑦) − 2𝑦 = 0
𝜕𝑦
𝜕2𝑉
: −𝑦 2 sin(𝑥𝑦) − 2 = 0 ⇔ 𝑦 2 sin(𝑥𝑦) + 2 = 0
𝜕𝑥 2
𝜕2𝑉
: −𝑥 2 sin(𝑥𝑦) − 2 = 0 ⇔ 𝑥 2 sin(𝑥𝑦) + 2 = 0
𝜕𝑦 2
𝜕2𝑉
: cos(𝑥𝑦) − 𝑥𝑦 sin(𝑥𝑦) = 0
𝜕𝑥 𝜕𝑦
𝒙 (𝒙𝟐+𝒚𝟐)
𝑽(𝒙, 𝒚) = 𝒆
𝒚
2 +𝑦 2 )
First we work on 𝑒 (𝑥 with the chain rule to get some intermediary results.
𝑓 = 𝑥2 + 𝑦2
𝜕𝑓 𝜕𝑓
= 2𝑥 𝑎𝑛𝑑 = 2𝑦
𝜕𝑥 𝜕𝑦
2 +𝑦 2 )
𝑔 = 𝑒 (𝑥 = 𝑒𝑓
𝜕𝑔 2 2
= 𝑒 𝑓 = 𝑒 (𝑥 +𝑦 )
𝜕𝑓
𝜕𝑔 𝜕𝑔 𝜕𝑓 2 2 𝜕𝑔 𝜕𝑔 𝜕𝑓 2 2
= = 2𝑥 𝑒 (𝑥 +𝑦 ) 𝑎𝑛𝑑 = = 2𝑦 𝑒 (𝑥 +𝑦 )
𝜕𝑥 𝜕𝑓 𝜕𝑥 𝜕𝑦 𝜕𝑓 𝜕𝑦
Now we continue with the complete equations and use the above results.
2𝑥 2 + 3 (𝑥2 +𝑦2 )
= 2𝑥 ( )𝑒
𝑦
𝜕 2 𝑉 𝜕𝑉 𝑥 2 2 2𝑥 2 2 𝑥 2 2
2
= [(2𝑥 − 2 ) 𝑒 (𝑥 +𝑦 ) ] = 3 𝑒 (𝑥 +𝑦 ) + (2𝑥 − 2 ) 2𝑦 𝑒 (𝑥 +𝑦 )
𝜕𝑦 𝜕𝑦 𝑦 𝑦 𝑦
2𝑥 2𝑥 (𝑥 2 +𝑦2 )
= ( 3 + 4𝑥𝑦 − ) 𝑒
𝑦 𝑦
1 1 2 2
= 2𝑥 ( 3 + 2𝑦 − ) 𝑒 (𝑥 +𝑦 )
𝑦 𝑦
𝜕2𝑉 𝜕𝑉 2𝑥 2 + 1 (𝑥 2 +𝑦2 ) −(2𝑥 2 + 1) (𝑥 2 +𝑦2 ) 2𝑥 2 + 1 2 2
= [( )𝑒 ]= 𝑒 +( ) 2𝑦 𝑒 (𝑥 +𝑦 )
𝜕𝑥 𝜕𝑦 𝜕𝑦 𝑦 𝑦2 𝑦
𝑽(𝒙, 𝒚) = 𝒆𝒙 𝒄𝒐𝒔 𝒚
𝜕𝑉
= 𝑒 𝑥 cos 𝑦
𝜕𝑥
𝜕𝑉
= −𝑒 𝑥 sin 𝑦
𝜕𝑦
𝜕2𝑉
= 𝑒 𝑥 cos 𝑦
𝜕𝑥 2
𝜕2𝑉
= −𝑒 𝑥 cos 𝑦
𝜕𝑦 2
𝜕2𝑉 𝜕𝑉 𝑥
= (𝑒 cos 𝑦) = −𝑒 𝑥 sin 𝑦
𝜕𝑥 𝜕𝑦 𝜕𝑦
Are these points stationary points of the following functions? If so, of what type?
𝜋 𝜋
For (𝑥 = − 2 , 𝑦 = 2 ) we have 𝐷𝑒𝑡 𝐻 = −1, so this is a saddle point.
𝜋 𝜋
For (𝑥 = − 2 , 𝑦 = − 2 ) we have 𝐷𝑒𝑡 𝐻 = 1 and 𝑇𝑟 𝐻 = 2, so this is a local minimum.
It’s obvious that we get the same values for the other two points, so these are no stationary points
either.
𝑑𝑣 2
𝑑𝑡
𝑑𝑣𝑣
=
𝑑𝑡
𝑑𝑣 𝑑𝑣
= 𝑣+𝑣
𝑑𝑡 𝑑𝑡
𝑑𝑣 𝑑𝑣
=𝑣 +𝑣
𝑑𝑡 𝑑𝑡
𝑑𝑣
= 2𝑣
𝑑𝑡
= 2𝑣𝑣̇
𝜕𝑉(𝑥𝑖 )
𝑚𝑖 𝑥𝑖̈ = −
𝜕𝑥𝑖
This is the general form for all dimensions, so first we start with 𝑎𝑥 .
1
𝜕 [2 𝑘(𝑥 2 + 𝑦 2 )] 𝑘
𝑚𝑎𝑥 = − ⇔ 𝑚𝑎𝑥 = −𝑘𝑥 ⇔ 𝑎𝑥 = − 𝑥
𝜕𝑥 𝑚
𝑘
Obviously we also have 𝑎𝑦 = − 𝑦.
𝑚
𝑘
In the above equations of motion we can set 𝜔2 = 𝑚. The equations then become 𝑥̈ 𝑖 = −𝜔2 𝑥𝑖 .
We recognize this as the equations of motion of the harmonic oscillator.
Show that there are circular orbits and that they all have the same period.
To have uniform circular motion in the 𝑥, 𝑦 plane, we know that our position coordinates need to
satisfy
𝑥(𝑡) = 𝑅 cos 𝜔𝑡 and 𝑦(𝑡) = 𝑅 sin 𝜔𝑡.
We also know that our equations of motion need to satisfy
𝑥̈ (𝑡) = −𝑅𝜔2 cos 𝜔𝑡 and 𝑦̈ (𝑡) = −𝑅𝜔2 sin 𝜔𝑡.
If we combine both these transformations, we get exactly 𝑥̈ 𝑖 = −𝜔2 𝑥𝑖 .
2𝜋
The period or the time it takes to go one full revolution is, as we know, equal to 𝑇 = .
𝜔
1 1 1
𝑇 = ∑ 𝑚𝑥𝑖̇ 2 = 𝑚𝑥̇ 2 + 𝑚𝑦̇ 2
2 2 2
𝑖
𝜕𝑉(𝑥𝑖 )
𝑚𝑖 𝑥𝑖̈ = −
𝜕𝑥𝑖
Let’s make a transformation to polar coordinates 𝑥 = 𝑟 cos 𝜃 and 𝑦 = 𝑟 sin 𝜃.
𝑟 = √𝑥 2 + 𝑦 2 or 𝑟 2 = 𝑥 2 + 𝑦 2
𝜕 𝑘 𝑘 𝜕 1 𝑘 −2 𝑘
𝑚𝑟̈ = − ( 2) = − ( 2) = − ( 3 ) = 3
𝜕𝑟 2𝑟 2 𝜕𝑟 𝑟 2 𝑟 𝑟
𝑘
⇔ 𝑟̈ =
𝑚𝑟 3
Further, since 𝜃̈ = 0, there is no acceleration for the 𝜃 component.
Are there circular orbits? If so, do they all have the same period? We haven’t found any.
𝑑(𝑟 cos 𝜃)
𝑥̇ = = 𝑟̇ cos 𝜃 − 𝑟𝜃̇ sin 𝜃
𝑑𝑡
2
𝑥̇ 2 = (𝑟̇ cos 𝜃 − 𝑟𝜃̇ sin 𝜃) = 𝑟̇ 2 cos 2 𝜃 − 2𝑟𝑟̇ 𝜃̇ cos 𝜃 sin 𝜃 + 𝑟 2 𝜃̇ 2 sin2 𝜃
𝑑(𝑟 sin 𝜃)
𝑦̇ = = 𝑟̇ sin 𝜃 + 𝑟𝜃̇ cos 𝜃
𝑑𝑡
2
𝑦̇ 2 = (𝑟̇ sin 𝜃 + 𝑟𝜃̇ cos 𝜃) = 𝑟̇ 2 sin2 𝜃 + 2𝑟𝑟̇ 𝜃̇ sin 𝜃 cos 𝜃 + 𝑟 2 𝜃̇ 2 cos 2 𝜃
𝑥̇ 2 + 𝑦̇ 2 = 𝑟̇ 2 + 𝑟 2 𝜃̇ 2
1 𝑚
𝑇 = 𝑚(𝑥̇ 2 + 𝑦̇ 2 ) = (𝑟̇ 2 + 𝑟 2 𝜃̇ 2 )
2 2
𝑚 2 𝑘 1
𝐸 =𝑇+𝑉 = (𝑟̇ + 𝑟 2 𝜃̇ 2 ) + 2
2 2𝑟
𝑑𝐸 𝑚 𝑑𝑟̇ 2 𝑑𝑟 2 𝜃̇ 2 𝑘𝑑 1 𝑚 𝑘 −2
= ( + )+ 2
= (2𝑟̈ 𝑟̇ + 2𝑟̇ 𝑟𝜃̇ 2 + 2𝑟 2 𝜃̈𝜃̇ ) + ( 3 𝑟̇ )
𝑑𝑡 2 𝑑𝑡 𝑑𝑡 2 𝑑𝑡 𝑟 2 2 𝑟
From later lectures we have learned that the Hamiltonian of this system doesn’t depend on 𝜃 itself.
We call 𝜃 a cyclic coordinate. Hamilton’s equation of motion gives us 𝑝̇𝜃 = 0. This means in this case
that angular momentum is conserved.
1
𝐿 = 𝑇 − 𝑉 = 𝑚𝑥̇ 2 − 𝑉(𝑥)
2
𝜕𝐿
= 𝑚𝑥̇
𝜕𝑥̇
𝑑 𝜕𝐿
= 𝑚𝑥̈
𝑑𝑡 𝜕𝑥̇
𝜕𝐿 𝜕𝑉(𝑥)
=− = 𝐹(𝑥)
𝜕𝑥 𝜕𝑥
𝑑 𝜕𝐿 𝜕𝐿
− =0
𝑑𝑡 𝜕𝑥̇ 𝜕𝑥
𝑑 𝜕𝐿 𝜕𝐿
⇔ =
𝑑𝑡 𝜕𝑥̇ 𝜕𝑥
⇔ 𝑚𝑥̈ = 𝐹(𝑥)
⇔ 𝐹 = 𝑚𝑎
1
𝐿 = 𝑇 − 𝑉 = ∑ ( 𝑚𝑥𝑖̇ 2 ) − 𝑉({𝑥})
2
𝑖
𝜕𝐿
= ∑(𝑚𝑥𝑖̇ )
𝜕𝑥𝑖̇
𝑖
𝑑 𝜕𝐿
= ∑(𝑚𝑥𝑖̈ )
𝑑𝑡 𝜕𝑥𝑖̇
𝑖
𝜕𝐿 𝜕𝑉({𝑥})
=− = 𝐹𝑖 ({𝑥})
𝜕𝑥𝑖 𝜕𝑥𝑖
𝑑 𝜕𝐿 𝜕𝐿
− =0
𝑑𝑡 𝜕𝑥𝑖̇ 𝜕𝑥𝑖
𝑑 𝜕𝐿 𝜕𝐿
⇔ =
𝑑𝑡 𝜕𝑥𝑖̇ 𝜕𝑥𝑖
⇔ ∑(𝑚𝑥𝑖̈ ) = 𝐹𝑖 ({𝑥})
𝑖
⇔ 𝐹𝑖 = 𝑚𝑖 𝑥𝑖̈
This is the general form of the Euler-Lagrange equations for all dimensions, in this case x and y.
𝑑 𝜕𝐿 𝜕𝐿
− =0
𝑑𝑡 𝜕𝑥𝑖̇ 𝜕𝑥𝑖
We start with the x coordinate first and split up the equation in some parts.
𝑑 𝜕𝐿 𝜕𝐿
− =0
𝑑𝑡 𝜕𝑥̇ 𝜕𝑥
𝜕𝐿
= 𝑚𝑥̇ + 𝑚𝜔𝑦
𝜕𝑥̇
𝑑 𝜕𝐿
= 𝑚𝑥̈ + 𝑚𝜔𝑦̇
𝑑𝑡 𝜕𝑥̇
𝜕𝐿
= 𝑚𝜔2 𝑥 − 𝑚𝜔𝑦̇
𝜕𝑥
𝑑 𝜕𝐿 𝜕𝐿
− = 0 ⇔ 𝑚𝑥̈ + 𝑚𝜔𝑦̇ − 𝑚𝜔2 𝑥 + 𝑚𝜔𝑦̇ = 0 ⇔ 𝑥̈ + 2𝜔𝑦̇ − 𝜔2 𝑥 = 0
𝑑𝑡 𝜕𝑥̇ 𝜕𝑥
⇔ 𝑥̈ = 𝜔2 𝑥 − 2𝜔𝑦̇
⇔ 𝑦̈ = 𝜔2 𝑦 + 2𝜔𝑥̇
𝑥 2 + 𝑦 2 = 𝑅 2 𝑐𝑜𝑠 2 𝜃 + 𝑅 2 𝑠𝑖𝑛2 𝜃 = 𝑅 2
= 𝑅̇ 2 + 𝑅 2 𝜃̇ 2
𝑥̇ 𝑦 − 𝑦̇ 𝑥 = 𝑅 𝑠𝑖𝑛 𝜃(𝑅̇ cos 𝜃 − 𝑅𝜃̇ sin 𝜃) − 𝑅 𝑐𝑜𝑠 𝜃(𝑅̇ sin 𝜃 + 𝑅𝜃̇ cos 𝜃)
= −𝑅 2 𝜃̇ sin2 𝜃 − 𝑅 2 𝜃̇ cos2 𝜃
= −𝑅 2 𝜃̇
Now we can easily plug these results into George’s Lagrangian.
𝑚 2 𝑚𝜔2 𝑅2
𝐿= (𝑅̇ + 𝑅 2 𝜃̇ 2 ) + − 𝑚𝜔𝑅 2 𝜃̇
2 2
Just for fun we work out the same exercise for Lenny’s Lagrangian.
𝒎 𝟐
𝑳= (𝒙̇ + 𝒚̇ 𝟐 )
𝟐
In polar coordinates this becomes:
𝑚 2
𝐿= (𝑅̇ + 𝑅 2 𝜃̇ 2 )
2
Now the equations of motion for the R coordinate.
𝑑 𝜕𝐿 𝜕𝐿
− = 0 ⇔ 𝑚𝑅̈ − 𝑚𝑅𝜃̇ 2 = 0 ⇔ 𝑅̈ = 𝑅𝜃̇ 2
𝑑𝑡 𝜕𝑅̇ 𝜕𝑅
We continue with the 𝜃 coordinate.
𝑑 𝜕𝐿 𝜕𝐿 −2𝑅̇
− = 0 ⇔ 2𝑚𝑅̇ 𝑅𝜃̇ + 𝑚𝑅 2 𝜃̈ = 0 ⇔ 2𝑅̇ 𝜃̇ + 𝑅𝜃̈ = 0 ⇔ 𝜃̈ = 𝜃̇
𝑑𝑡 𝜕𝜃̇ 𝜕𝜃 𝑅
The above Lagrangian describes a single particle of mass m moving in the 𝑥, 𝑦 plane but with no
𝑚
forces acting on it. In Cartesian coordinates it is 𝐿 = 2 (𝑥̇ 2 + 𝑦̇ 2 ).
The bob of the pendulum has mass 𝑚. We take 𝜃 to be the angle to the vertical axis and 𝑙 the length
of the rod. We can then replace 𝑅 in the above equations with the constant 𝑙.
So we only have the kinetic energy 𝑇 in the above Lagrangian. For a hanging pendulum we have to
bring in the potential energy 𝑉. This depends on the height ℎ which is 𝑙 − 𝑙 cos 𝜃.
𝑚𝑙 2 𝜃̇ 2
𝐿 =𝑇−𝑉 = − 𝑚𝑔𝑙 + 𝑚𝑔𝑙 𝑐𝑜𝑠 𝜃
2
We only need to derive the Euler-Lagrange equation of motion for the 𝜃 coordinate.
𝜕𝐿
= 𝑚𝑙 2 𝜃̇
𝜕𝜃̇
𝑑 𝜕𝐿
= 𝑚𝑙 2 𝜃̈
𝑑𝑡 𝜕𝜃̇
𝑑 𝜕𝐿 𝜕𝐿
− = 0 ⇔ 𝑚𝑙 2 𝜃̈ − (−𝑚𝑔𝑙 sin 𝜃) = 0 ⇔ 𝑙𝜃̈ + 𝑔 sin 𝜃 = 0
𝑑𝑡 𝜕𝜃̇ 𝜕𝜃
−𝑔 sin 𝜃
⇔ 𝜃̈ =
𝑙
𝑚
We have to prove that 2
(𝑥1̇ 2 + 𝑥2̇ 2 ) = 𝑚(𝑥+̇ 2 + 𝑥−̇ 2 )
𝑑 𝑥1 + 𝑥2 𝑥̇ 1 + 𝑥̇ 2
𝑥̇ + = ( )=
𝑑𝑡 2 2
𝑥̇ 1 2 + 𝑥̇ 2 2 + 2𝑥̇ 1 𝑥̇ 2
𝑥̇ + 2 =
4
𝑑 𝑥1 − 𝑥2 𝑥̇ 1 − 𝑥̇ 2
𝑥̇ − = ( )=
𝑑𝑡 2 2
𝑥̇ 1 2 + 𝑥̇ 2 2 − 2𝑥̇ 1 𝑥̇ 2
𝑥̇ − 2 =
4
𝑑 𝜕𝐿 𝜕𝐿
𝑝̇1 = = = −𝑉 ′ (𝑞1 − 𝑞2 )
𝑑𝑡 𝜕𝑞̇ 1 𝜕𝑞1
𝑑 𝜕𝐿 𝜕𝐿
𝑝̇2 = = = +𝑉 ′ (𝑞1 − 𝑞2 )
𝑑𝑡 𝜕𝑞̇ 2 𝜕𝑞2
In the expression −𝑉(𝑞1 − 𝑞2 ) there is a negative sign for 𝑉 and a positive sign for 𝑞1 . The product of
these signs gives a negative sign in front of the derivative.
Alternatively there is a negative sign for 𝑉 and a negative sign for 𝑞2 in −𝑉(𝑞1 − 𝑞2 ). So the product
of these signs becomes positive in front of the derivative.
We multiply the first equation with 𝑏 and the second one with 𝑎 as instructed.
𝑏𝑝̇1 + 𝑎𝑝̇ 2 = 0
We say that in this case the quantity 𝑏𝑝1 + 𝑎𝑝2 is conserved. Why?
If a quantity is conserved, then this means that it doesn’t change over time. This is the same as saying
that its time derivative is 0. We know that 𝑎 and 𝑏 are constants.
𝑑 𝑑𝑝1 𝑑𝑝2
(𝑏𝑝1 + 𝑎𝑝2 ) = 0 ⇔ 𝑏 +𝑎 = 0 ⇔ 𝑏𝑝̇1 + 𝑎𝑝̇2 = 0
𝑑𝑡 𝑑𝑡 𝑑𝑡
𝑎𝑞1 − 𝑏𝑞2 = 𝑎(𝑞1 − 𝑏𝛿) − 𝑏(𝑞2 − 𝑎𝛿) = 𝑎𝑞1 − 𝑎𝑏𝛿 − 𝑏𝑞2 + 𝑎𝑏𝛿 = 𝑎𝑞1 − 𝑏𝑞2
We know that 𝑎 and 𝑏 are constants and that 𝛿 doesn’t depend on time.
𝑚
= (𝑞 ̇ 2 + 𝑞2̇ 2 ) − 𝑉(𝑎𝑞1 − 𝑏𝑞2 )
2 1
𝒚 → 𝒚 − 𝒙𝜹 and 𝒚̇ → 𝒚̇ − 𝒙̇ 𝜹
𝜹𝒙 = 𝒚𝜹 and 𝜹𝒚 = −𝒙𝜹
The Lagrangian does not change to first order in 𝜹. Show this to be true.
𝑚 2
(𝑥̇ + 𝑦̇ 2 ) − 𝑉(𝑥 2 + 𝑦 2 )
2
𝑚
= (𝑥̇ 2 + 𝑦̇ 2 𝛿 2 + 2𝑥̇ 𝑦̇𝛿 + 𝑦̇ 2 + 𝑥̇ 2 𝛿 2 − 2𝑥̇ 𝑦̇𝛿) − [𝑉(𝑥 2 + 𝑦 2 ) + 𝛿𝑉(𝑥 2 + 𝑦 2 )]
2
The terms with 𝛿 2 are 0 for an infinitesimally small 𝛿.
𝑚 2 𝜕𝑉(𝑥 2 + 𝑦 2 ) 𝜕𝑉(𝑥 2 + 𝑦 2 )
= (𝑥̇ + 𝑦̇ 2 ) − [𝑉(𝑥 2 + 𝑦 2 ) + 𝛿𝑥 + 𝛿𝑦]
2 𝜕𝑥 𝜕𝑦
𝑚 2
= (𝑥̇ + 𝑦̇ 2 ) − [𝑉(𝑥 2 + 𝑦 2 ) + 2𝑥𝛿𝑥 + 2𝑦𝛿𝑦]
2
𝑚
= (𝑥̇ 2 + 𝑦̇ 2 ) − [𝑉(𝑥 2 + 𝑦 2 ) + 2𝑥𝑦𝛿 − 2𝑦𝑥𝛿]
2
𝑚
= (𝑥̇ 2 + 𝑦̇ 2 ) − 𝑉(𝑥 2 + 𝑦 2 )
2
We choose 𝜃 to be the angle from the vertical axis. The bob of the pendulum has mass 𝑚 and the
length of the rod is 𝑙. The coordinates of the bob are 𝑥 and 𝑦.
𝑚 2 𝑚 2 2 𝑚𝑙 2 𝜃̇ 2
𝑇= (𝑥̇ 2) ̇ 2 2 ̇2 2
+ 𝑦̇ = (𝑙 𝜃 𝑐𝑜𝑠 𝜃 + 𝑙 𝜃 𝑠𝑖𝑛 𝜃) =
2 2 2
The potential energy 𝑉 depends on the height ℎ which is 𝑙 − 𝑙 cos 𝜃.
𝑚𝑙 2 𝜃̇ 2
𝐿 =𝑇−𝑉 = − 𝑚𝑔𝑙 + 𝑚𝑔𝑙 cos 𝜃
2
The Euler-Lagrange equation of motion can be given in terms of the one degree of freedom, which is
𝜃.
𝑑 𝜕𝐿 𝜕𝐿
− =0
𝑑𝑡 𝜕𝜃̇ 𝜕𝜃
𝑑
⇔ 𝑚𝑙 2 𝜃̇ − (−𝑚𝑔𝑙 sin 𝜃) = 0
𝑑𝑡
⇔ 𝑚𝑙 2 𝜃̈ = −𝑚𝑔𝑙 sin 𝜃
⇔ 𝑙𝜃̈ = −𝑔 sin 𝜃
−𝑔 sin 𝜃
⇔ 𝜃̈ =
𝑙
3𝜃̇ 2 𝛼̇ 2
𝐿= + + 𝜃̇ 𝛼̇ + 𝜃̇ 2 𝑐𝑜𝑠 𝛼 + 𝜃̇ 𝛼̇ 𝑐𝑜𝑠 𝛼 + 2𝑔 cos 𝜃 + 𝑔 cos 𝜃 cos 𝛼 + 𝑔 sin 𝜃 sin 𝛼
2 2
𝑑 𝜕𝐿 𝜕𝐿
− =0
𝑑𝑡 𝜕𝜃̇ 𝜕𝜃
𝑑
⇔ (3𝜃̇ + 𝛼̇ + 2𝜃̇ 𝑐𝑜𝑠 𝛼 + 𝛼̇ 𝑐𝑜𝑠 𝛼) − (−2𝑔 sin 𝜃 − 𝑔 sin 𝜃 cos 𝛼 + 𝑔 cos 𝜃 sin 𝛼) = 0
𝑑𝑡
⇔ 3𝜃̈ + 𝛼̈ + 2𝜃̈ cos 𝛼 − 2𝜃̇ 𝛼̇ sin 𝛼 + 𝛼̈ cos 𝛼 − 2𝛼̇ sin 𝛼 + 2𝑔 sin 𝜃 + 𝑔 sin(𝜃 − 𝛼) = 0
𝑑 𝜕𝐿 𝜕𝐿
− =0
𝑑𝑡 𝜕𝛼̇ 𝜕𝛼
𝑑
⇔ (𝛼̇ + 𝜃̇ + 𝜃̇ 𝑐𝑜𝑠 𝛼) − (−𝜃̇ 2 sin 𝛼 − 𝜃̇ 𝛼̇ sin 𝛼 − 𝑔 cos 𝜃 sin 𝛼 + 𝑔 sin 𝜃 cos 𝛼) = 0
𝑑𝑡
3𝜃̇ 2 𝛼̇ 2
𝐿= + + 𝜃̇ 𝛼̇ + 𝜃̇ 2 𝑐𝑜𝑠 𝛼 + 𝜃̇ 𝛼̇ 𝑐𝑜𝑠 𝛼
2 2
𝜕𝐿
𝑝𝜃 =
𝜕𝜃̇
We have to check if the action or the Lagrangian changes when we rotate the complete system
around the origin by an infinitesimally small 𝛿. If we do that, then 𝜃 changes, but since 𝛼 is relative to
𝜃, it doesn’t change.
𝜃 →𝜃+𝛿
𝛼→𝛼
All the terms of the Lagrangian stay the same under this transformation, namely 𝜃̇, 𝛼̇ and 𝑐𝑜𝑠 𝛼.
Now we know from Noether’s theorem that the following quantity is conserved:
𝑄 = ∑ 𝑝𝑖 𝑓𝑖 (𝑞)
𝑖
Eq. (14)
𝟏 𝟐 𝝎 𝟐
𝑳= 𝒒̇ − 𝒒
𝟐𝝎 𝟐
𝑚𝑥̇ 2 𝑘 2 𝑚 𝑞̇ 2 𝑘 𝑞2
𝐿= − 𝑥 = −
2 2 2 √𝑘𝑚 2 √𝑘𝑚
To get to the form of Eq. (14) we see that:
√𝑘𝑚 𝑘
𝜔= 𝑎𝑛𝑑 𝜔 =
𝑚 √𝑘𝑚
For both expressions to be correct, the following should be true:
√𝑘𝑚 𝑘
=
𝑚 √𝑘𝑚
⇔ √𝑘𝑚√𝑘𝑚 = 𝑘𝑚
⇔ 𝑘𝑚 = 𝑘𝑚
√𝑘𝑚
𝜔=
𝑚
𝑘𝑚 𝑘
⇔ 𝜔2 = =
𝑚2 𝑚
This connection is still the same as in the Newtonian formulation of mechanics.
Eq. (14)
𝟏 𝟐 𝝎 𝟐
𝑳= 𝒒̇ − 𝒒
𝟐𝝎 𝟐
𝐻 = ∑ 𝑝𝑖 𝑞̇ 𝑖 − 𝐿
𝑖
𝜔 2 𝑝2 𝜔 2
= 𝑝𝜔𝑝 − + 𝑞
2𝜔 2
𝜔 𝜔
= (𝜔 − ) 𝑝2 + 𝑞 2
2 2
𝜔 𝜔
= 𝑝2 + 𝑞 2
2 2
𝜔
= (𝑝2 + 𝑞 2 )
2
We use the definition of Poisson brackets for the left side of the equation.
{𝐹(𝑞, 𝑝), 𝑝𝑖 }
𝜕𝑝𝑖 𝜕𝑝
𝜕𝑝𝑗
is equal to 1 if 𝑖 = 𝑗 and equal to 0 if 𝑖 ≠ 𝑗. So in fact 𝜕𝑝 𝑖 = 𝛿𝑖𝑗 .
𝑗
𝜕𝑝𝑖
𝜕𝑞𝑗
is always equal to 0 because 𝑝𝑖 ≠ 𝑞𝑗 .
𝑚𝑞̈
= 𝑝̇
= {𝑝, 𝐻}
1 2
= {𝑝, 𝑝 + 𝑉(𝑞)}
2𝑚
1 2
= {𝑝, 𝑝 } + {𝑝, 𝑉(𝑞)}
2𝑚
1
= {𝑝, 𝑝𝑝} − {𝑉(𝑞), 𝑝}
2𝑚
𝑝 𝑝 𝜕𝑉(𝑞)
= {𝑝, 𝑝} + {𝑝, 𝑝} −
2𝑚 2𝑚 𝜕𝑞
𝜕𝑉(𝑞)
=−
𝜕𝑞
= 𝐹𝑞
Using both the definitions of Poisson brackets and the axioms, work out the Poisson brackets in
the above equations. Hint: in each expression, look for things in the parentheses that have nonzero
Poisson brackets with the coordinate 𝒙, 𝒚 or 𝒛. For example in the first Poisson bracket 𝒙 has a
nonzero Poisson bracket with 𝒑𝒙 .
=0+0−𝑦−0
= −𝑦
=𝑥+0−0−0
=𝑥
=0+0−0−0
=0
⃗⃗ × ⃗𝑨⃗) .
Confirm this equation. Also prove that 𝑽𝒊 𝑨𝒋 − 𝑽𝒋 𝑨𝒊 = ∑𝒌 𝝐𝒊𝒋𝒌 (𝑽 𝒌
⃗⃗ × 𝐴⃗) = ∑ ∑ 𝜖𝑖𝑗𝑘 𝑉𝑗 𝐴𝑘
(𝑉 𝑖
𝑗 𝑘
If in the above equation 𝑖 is fixed to 𝑥, 𝑦 or 𝑧, then 𝑘 cycles through 𝑥, 𝑦 and 𝑧 with 𝑗 respectivally
cycling through 𝑥, 𝑦 and 𝑧. So this will give us one term on the left side and nine terms on the right
side.
If we let 𝑖 = 𝑥, then the indices of 𝜖 become 𝑥𝑥𝑥, 𝑥𝑥𝑦, 𝑥𝑥𝑧, 𝑥𝑦𝑥, 𝑥𝑦𝑦, 𝑥𝑦𝑧, 𝑥𝑧𝑥, 𝑥𝑧𝑦, 𝑥𝑧𝑧. Any time we
have at least two equal letters, 𝜖𝑖𝑗𝑘 becomes 0. This reduces the nine terms to two terms, namely
the ones with the indices 𝑥𝑦𝑧 and 𝑥𝑧𝑦.
⃗⃗ × 𝐴⃗)
𝑉𝑖 𝐴𝑗 − 𝑉𝑗 𝐴𝑖 = ∑ 𝜖𝑖𝑗𝑘 (𝑉 𝑘
𝑘
Here we have to fix 𝑖 and 𝑗 to 𝑥, 𝑦 or 𝑧 and let 𝑘 cycle through 𝑥, 𝑦 and 𝑧. So we will have one term
on the left and three terms on the right side.
It’s clear that if we take 𝑖 = 𝑗, then both the left and right side becomes 0.
So let’s take 𝑖 = 𝑥, 𝑗 = 𝑦. The indices of 𝜖 are then 𝑥𝑦𝑥, 𝑥𝑦𝑦, 𝑥𝑦𝑧. The three terms reduce to one,
with the index of 𝜖 being 𝑥𝑦𝑧.
⃗⃗ × 𝐴⃗) = (𝑉
𝑉𝑥 𝐴𝑦 − 𝑉𝑦 𝐴𝑥 = 𝜖𝑥𝑦𝑧 (𝑉 ⃗⃗ × 𝐴⃗)
𝑧 𝑧
⃗⃗ × 𝐴⃗) = (𝑉
𝑉𝑦 𝐴𝑧 − 𝑉𝑧 𝐴𝑦 = 𝜖𝑦𝑧𝑥 (𝑉 ⃗⃗ × 𝐴⃗)
𝑥 𝑥
⃗⃗ × 𝐴⃗) = (𝑉
𝑉𝑧 𝐴𝑥 − 𝑉𝑥 𝐴𝑧 = 𝜖𝑧𝑥𝑦 (𝑉 ⃗⃗ × 𝐴⃗)
𝑦 𝑦
The same analogy is true for the reversed orders 𝑖 = 𝑦, 𝑗 = 𝑥 and 𝑖 = 𝑧, 𝑗 = 𝑦 and 𝑖 = 𝑥, 𝑗 = 𝑧.
⃗⃗ × 𝐴⃗) = −(𝑉
𝑉𝑦 𝐴𝑥 − 𝑉𝑥 𝐴𝑦 = 𝜖𝑦𝑥𝑧 (𝑉 ⃗⃗ × 𝐴⃗)
𝑧 𝑧
⃗⃗ × 𝐴⃗) = −(𝑉
𝑉𝑧 𝐴𝑦 − 𝑉𝑦 𝐴𝑧 = 𝜖𝑧𝑦𝑥 (𝑉 ⃗⃗ × 𝐴⃗)
𝑥 𝑥
⃗⃗ × 𝐴⃗) = −(𝑉
𝑉𝑥 𝐴𝑧 − 𝑉𝑧 𝐴𝑥 = 𝜖𝑥𝑧𝑦 (𝑉 ⃗⃗ × 𝐴⃗)
𝑦 𝑦
We start with a vector field ∇⃗⃗ 𝑉(𝑥) that is defined as the gradient of the scalar field 𝑉(𝑥).
The scalar field is a function of space, depending on 𝑥, 𝑦, 𝑧. So the vector field that is derived from it
has three components:
𝜕𝑉
∇𝑥 𝑉 =
𝜕𝑥
𝜕𝑉
∇𝑦 𝑉 =
𝜕𝑦
𝜕𝑉
∇𝑧 𝑉 =
𝜕𝑧
Now we take the curl ⃗∇⃗ × [∇
⃗⃗ 𝑉(𝑥)] of the vector field, of which the components are second order
partial derivatives. We use our knowledge here that the order in which the derivatives are taken
doesn’t matter. This leads to the conclusion that all the components of the vector ⃗∇⃗ × [∇
⃗⃗ 𝑉(𝑥)] are
0.
𝜕 𝜕𝑉 𝜕 𝜕𝑉 𝜕 𝜕𝑉 𝜕 𝜕𝑉
⃗⃗ × ⃗∇⃗ 𝑉) =
(∇ − = − =0
𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑧 𝜕𝑦 𝜕𝑦 𝜕𝑧 𝜕𝑦 𝜕𝑧
𝜕 𝜕𝑉 𝜕 𝜕𝑉 𝜕 𝜕𝑉 𝜕 𝜕𝑉
⃗⃗ × ⃗∇⃗ 𝑉) =
(∇ − = − =0
𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑧 𝜕𝑥
𝜕 𝜕𝑉 𝜕 𝜕𝑉 𝜕 𝜕𝑉 𝜕 𝜕𝑉
⃗⃗ × ⃗∇⃗ 𝑉) =
(∇ − = − =0
𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦
Using the above Hamiltonian, work out Hamilton’s equations of motion and show that you just get
back to the Newton-Lorentz equation of motion.
𝜕𝐻 2 𝑒 1 𝑒 𝑒
𝑥̇ = = (𝑝𝑥 − 𝐴𝑥 ) = (𝑝𝑥 − 𝐴𝑥 ) ⇔ 𝑚𝑥̇ = (𝑝𝑥 − 𝐴𝑥 )
𝜕𝑝𝑥 2𝑚 𝑐 𝑚 𝑐 𝑐
𝜕𝐻 2 𝑒 1 𝑒 𝑒
𝑦̇ = = (𝑝𝑦 − 𝐴𝑦 ) = (𝑝𝑦 − 𝐴𝑦 ) ⇔ 𝑚𝑦̇ = (𝑝𝑦 − 𝐴𝑦 )
𝜕𝑝𝑦 2𝑚 𝑐 𝑚 𝑐 𝑐
𝜕𝐻 2 𝑒 1 𝑒 𝑒
𝑧̇ = = (𝑝𝑧 − 𝐴𝑧 ) = (𝑝𝑧 − 𝐴𝑧 ) ⇔ 𝑚𝑧̇ = (𝑝𝑧 − 𝐴𝑧 )
𝜕𝑝𝑧 2𝑚 𝑐 𝑚 𝑐 𝑐
−𝜕𝐻
We continue with 𝑝̇ = 𝜕𝑞
for the x component and use the above results to plug in on the way.
With this result we get back to 𝑚𝑥̇ and take its time derivative.
𝑑 𝑑 𝑒 𝑒 𝑑𝐴𝑥 𝑒 𝜕𝐴𝑥 𝜕𝐴𝑥 𝜕𝐴𝑥
𝑚𝑥̇ = (𝑝𝑥 − 𝐴𝑥 ) = 𝑝̇𝑥 − = 𝑝̇𝑥 − ( 𝑥̇ + 𝑦̇ + 𝑧̇ )
𝑑𝑡 𝑑𝑡 𝑐 𝑐 𝑑𝑡 𝑐 𝜕𝑥 𝜕𝑦 𝜕𝑧
𝑒 𝜕𝐴𝑥 𝜕𝐴𝑦 𝜕𝐴𝑧 𝑒 𝜕𝐴𝑥 𝜕𝐴𝑥 𝜕𝐴𝑥
⇔ 𝑚𝑥̈ = (𝑥̇ + 𝑦̇ + 𝑧̇ )− ( 𝑥̇ + 𝑦̇ + 𝑧̇ )
𝑐 𝜕𝑥 𝜕𝑥 𝜕𝑥 𝑐 𝜕𝑥 𝜕𝑦 𝜕𝑧
We know that 𝑎𝑥 = 𝑣̇𝑥 and that 𝑎𝑦 = 𝑣̇𝑦 . So we can rewrite the equations in terms of the velocity
components.
𝑒𝑏
𝑣̇𝑦 = − 𝑣
𝑚𝑐 𝑥
and
𝑒𝑏
𝑣̇𝑥 = 𝑣
𝑚𝑐 𝑦
Now we make another switch from the 𝑣𝑥 , 𝑣𝑦 plane to the 𝑥, 𝑦 plane. So 𝑥 = 𝑣𝑥 and 𝑦 = 𝑣𝑦 .
𝑒𝑏
We can also put all the constants in one expression = 𝜔.
𝑚𝑐
𝑦̇ = −𝜔𝑥 and 𝑥̇ = 𝜔𝑦
This is something we recognise. These are Hamilton’s equations of motion that describe the
trajectories in phase space of the harmonic oscillator. These trajectories are concentric circular orbits
in the 𝑞, 𝑝 plane. Here it is in the 𝑥, 𝑦 plane or in the 𝑣𝑥 , 𝑣𝑦 plane.
The fixed radius of this circle can be expressed as 𝑟. Now we just use the Pythagorean theorem.
𝑟 2 = 𝑞 2 + 𝑝2
⇔ 𝑟 = √𝑞 2 + 𝑝2 = √𝑥 2 + 𝑦 2 = √𝑣𝑥 2 + 𝑣𝑦 2
𝑎𝑥 = −𝑅𝜔2 cos 𝜔𝑡
𝑎𝑦 = −𝑅𝜔2 sin 𝜔𝑡
These were derived as the second order time derivatives of the position coordinates.
𝑥(𝑡) = 𝑅 cos 𝜔𝑡
𝑦(𝑡) = 𝑅 sin 𝜔𝑡
In our circular planetary orbit example we have defined polar coordinates from the Cartesian
coordinates in a similar way.
𝑟 = √𝑥 2 + 𝑦 2
𝑥
cos 𝜃 = ⇔ 𝑥 = 𝑟 cos 𝜃
𝑟
𝑦
sin 𝜃 = ⇔ 𝑦 = 𝑟 sin 𝜃
𝑟
So we can further derive that:
𝑅=𝑟
𝜔𝑡 = 𝜃 and so 𝜃̇ = 𝜔
Now we take the time derivatives to first and second order to get the acceleration.
𝑦̇ = 𝑣𝑦 = 𝑟 𝜃̇ cos 𝜃 = 𝑟𝜔 cos 𝜃
The components 𝑎𝑥 and 𝑎𝑦 are the components of the acceleration vector 𝑎⃗.
We look for the magnitude of this vector.