2022/3/1

Motivation
 x1 , y1 
 x2 , y 2 

Calculus of variations Many problems require minimizing functions.

Ex. Minimum distance between two points.

Toward Euler–Lagrange equation Ex. The shape of a chain hanging between the two points with length constrain.
Ex. Which trajectory of a ball rolling between two points take least time.

And many many more.

Minimizing the travel time of a

Minimum distance
particle
 x1 , y1 
 x2 , y 2 
x2
dS Velocity of a particle can be determined by the
I 
x1
v energy conservation.
1
mgy  mv 2
The shortest distance is by definition a minimum path integration. dS  1  y 2 dx 2
Path integration
v  2 g  y1  y 
 dS    dx    dy 
2 2 2
x2 dS
I   dS dy
1  y 2 dx 1  y 2
x2 x2
  dy  2 
x1 dx
 dS 
2
 dx 2 1    
I  v
  2 g  y1  y 
dx
  dx   x1 x1

dS  1  y 2 dx
x2

I 
x1
1  y 2 dx
 2022/3/1

Minimize the energy –

Minimize the Lagrangian
Hanging chain
This is the most useful form for physical problem.
Hanging chain has minimum potential energy.
Linear mass density
E p  mgh
x2
Lagrangian Lagrangian is defined as
I    AgydS
x2 L  T V
x1

dS  1  y 2 dx
I  L  x, y, y dx
x1 Kinetic energy Potential energy
x2

I   Ag  y 1  y 2 dx
x1
In this problem, there is a constrain
x2
which is the chain length.
l  
x1
1  y2 dx constrain

The math The math (cont’d)

y  x     x  By multiplying , a family of functions
can be considered.
x2
y  x   y  x     x 
Minimum I   F  x, y  x  , y   x   dx
y  x   x x1
y   x   y   x      x 
A function bounded on the two points
 x1 , y1    x1     x2   0
 x2 , y 2 
x2
dI d
Need to be differentialable! d  0
0
d  F  x, y  x  , y  x  dx
 0 x1
y  x Move the differentiation into the
integral, since constant boundaries.
Assume this is the solution (minimized function) x2
d
Let
 d  F  x, y  x  , y  x   0
dx  0 Leibniz rule
y  x   y  x     x  x1

0 
Just like x  dx y  x    x
 F x F y F y  
x2
Take derivative 
y   x   y   x      x  x  x   y   y   dx  0 
1  0
y  x      x 

Consider
x 2

I   F  x, y  x  , y   x   dx The goal is to minimize the x2

 F F 
x1
A function of functions
integration.   y   y    dx  0
x1  0
 2022/3/1

The math (cont’d) Euler-Lagrange equation

F d  F 
x2
 F F   0
  y   y    dx  0 y dx  y  
x1  0
yy
Note that at  , x, y can be generalized coordinates.
y   y

 F F 
x2

  y   y    dx  0 1st variation

x1  0
Weak form

Since   in the eq.

Integration by parts
0 b b

a udv  vu a  a vdu
x2 b
F  F d  F  
x2

        dx  0
y  x1 x1  y dx  y   
Strong form
  x1     x2   0 Since  is arbitrary function, F d  F  Euler-Lagrange
  0 equation
the only solution is y dx  y  

Example: Minimum distance

Example: Minimum distance
(cont’d) y  x 
 const.  C
 x1 , y1   x1 , y1  1   y   x  
2

 x2 , y 2   x2 , y 2 
 y   x  
2

 C2
1   y   x  
2
F d  F 
 0
x2
y dx  y  
 y   x    C 2  C 2  y   x  
2 2
I  1  y  dx
2

x1 F
0
1  C   y  x 
2
y
2
 C2
F  x, y , y  1  y2 2
F y y  x  
C
A

y  1  y 2 1 C2
y  x   Ax  B Straight line!
d y  x 
0
dx 1   y   x   2
  Bounded at  x1 , y1  and  x2 , y2 
y  x  A and B can be solved accordingly.
 const.  C
1   y   x  
2
 2022/3/1

Generalizations Generalizations (cont’d)

Single function of single variable with higher derivatives Single function of several variables with single derivative
dk y x2 Multi-dimensions
y   k I  y   L  x , x ,, x , y , y ,, y  dx
k

x2 dx 1 2 n 1 2 n

I  y    L  x, y , y , y ,, y  k   dx dy
x1
yj  j  1, 2,, n
dx j
L n d  L 
x1

  0
L d  L  d 2  L  k d
k
 L  y j 1 dx j  y j 
      1  0
y dx  y   dx 2  y  dx k  y  k  

Several functions of single variable with single derivative More generalization can be found in wiki.
x2

I  y1 , y2 ,, ym    L  x , y , y ,, y
1 2 m , y1, y2 ,, ym  dx
x1

L d  L 
  0 i  1, 2,, m
yi dx  yi 

https://en.wikipedia.org/wiki/Euler%E2%80%93Lagrange_equation