Chapter 2: Variational Principles & Lagrange’s Equations * If U depends only on qi (i.e.

, not on q˙i and t), the system is

conservative.
2.1 Hamilton’s Principle
* Hamilton’s Principle (for monogenic systems):
Hamilton Principle (integral or variational principle) is a very The motion of the system ! (in configuration space) from time
general, elegant mathematics technique; it has many practical t1 to time t2 is such that the line integral (the action or
applications beside mechanics and can also be used to derive action integral)
Lagrange’s Equations of Motion from a totally different t
I = " t 2 Ldt
approach. 1
has a stationary value for the actual path of motion.
Meaning of “motion of system between time t1 and t2”: L ! T - U = Lagrangian of the system
• A system is characterized by n generalized coordinates !
q1,q2,q3,..qn. •“Stationary value” means I is an extremum (either a
• At time t1: q1(t1),q2(t1),..,qn(t1) represent a point in the n- maximum or a minimum, almost always a minimum).
dimensional configuration space. •In other words: Out of all possible paths q t2
• As time goes on, the system point moves in configuration by which the system point could travel in 2

space tracing out a curve, called the t2 configuration space from t1 to t2, it will
path of motion of the system. q2 t1
ACTUALLY travel along path for which I
• At time t2: q1(t2),q2(t2),.. ,qn(t2) is an extremum (usually a minimum).
represent another point in the n- t1 q1
dimensional configuration space.
q1 •Hamilton Principle states that the motion is such that the
variation of I from the actual path with fixed t1 and t2 is zero:
* Monogenic Systems ! All generalized forces (except t
"I = " # t 2 Ldt = 0
constraint forces) are derivable from a generalized scalar 1
potential U(qi, q˙ i,t) that may be a function of generalized
coordinates, generalized velocities, and time * Hamilton Principle suggests “Nature always minimizes certain
Qi ! - "U/"qi + d("U/" q˙ i) /dt quantities when a physical process takes place”.
!
!
By Prof. Juyang Huang, Texas Tech University. All rights reserved. By Prof. Juyang Huang, Texas Tech University. All rights reserved.
!
*Examples of other “minimal principles”:
(a) Hero, 200 BC. Hero’s Principle of Least Distance: A
light ray traveling from one point to another by reflection * Vary the path y(x) until an extremum (minimum) of J is found.
from a plane mirror, always takes shortest path: " the Law Name this correct path y(x, 0).
of Reflection. * Every “neighboring function”, call it y(x, #), no matter how
(b) Fermat, 1657. Fermat’s Principle of Least Time: A light close to y(x,0), must make J increase.
ray travels in a medium from one point to another by a path * Thus, variation of J from y(x, 0) is zero or
that takes the least time. "J = " # xx2 f (y, y˙, x)dx = 0
" Law of Reflection 1
* The textbook (pages 37-38) shows that the functional f must
" Law of Refraction: “Snell’s Law”
satisfy the Euler-Lagrange differential Equation:
(c) Maupertuis, 1747. Maupertuis’s Principle of Least Action: d "f "f
Dynamical motion takes place with minimum action: ! # =0
dx "y˙ "y
Action ! (Distance) # (Momentum) = (Energy)# (Time)
* The textbook gives several examples of application of
Hamilton Principle in math. Here is one of them:
2.2: Techniques of the Calculus of Variations !
Example: Minimum surface of revolution. A surface of
* Consider the following calculus of variations problem in x-y
plane: A function f(y, y˙ , x) defined on a path y = y(x) between revolution is formed by taking
some curve passing between two
two fixed values x1 and x2, where y˙ = dy/dx. Find a particular
fixed end points (x1, y1) and (x2, y2),
path y(x) such that the line integral J of the function f between x1 and revolving it about the y-axis.
and x2 Find the curve, y=y(x), for which
! x2
y (x2,y2)
J = " x f (y, y˙, x)dx the surface area is minimum.
1 !
has a stationary value (i.e., J is an
extremum). Solution: The total surface area is
(x1,y1) A = 2" #12 x 1+ y˙ 2 dx , and the
! x
integral function f is identified as f = x 1+ y˙ 2

By Prof. Juyang Huang, Texas Tech University. All rights reserved.

! By Prof. Juyang Huang, Texas Tech University. All rights reserved.
!
 Use Euler-Lagrange equation:
$ '
d "f "f d & xy˙ )
# = & =0
dx "y˙ "y dx % 1+ y˙ 2 )(
xy˙
Therefore, = a = constant
1+ y˙ 2
! dy a
or =
dx x 2 + a2
!
The general solution is
y(x) = a arc cosh(x/a) + b, where
! constant a and b are determined
two
by the two fixed end points.

By Prof. Juyang Huang, Texas Tech University. All rights reserved. By Prof. Juyang Huang, Texas Tech University. All rights reserved.
 2.3 Derivation of Lagrange Equations from Hamilton’s From Hamilton Principle to Lagrange’s Equation
Principle
The integral in Hamilton’s principle
Here we extend the calculus of variations results to I = "11L(qi , q˙i ,t)dt
functions with n independent variables. The problem becomes can be obtained with following transformations:
finding a particular path (y1(x), y2(x), …) in n-dimensional x $ t;
configuration space such that the integral J of the function f
y i $ q i;
between x1 and x2 !
f(yi, y˙ i,x) $ L(qi, q˙ i, t)
J = "12 f (y1 (x), y2 (x),..., y˙1 (x), y˙2 (x),..., x)dx
has a stationary value (i.e., J is an extremum). So Lagrange equations of motion can be derived from the
Hamilton Principle:
Equivalently, a variation of the integral J ! !
!
"J = " #12 f (y1 (x), y2 (x),..., y˙1 (x), y˙2 (x),..., x)dx d "L "L
# =0 i = 1, 2, …, n.
must be zero for any “neighboring path” around the correct path. dt "q˙i "qi

Since n variables are independent, similar to the one for monogenic system with holonomic constraints.
! variable case, n independent Euler-Lagrange differential
equations must be satisfied: !

d "f "f
# =0 i = 1, 2, …, n.
dx "y˙i "yi

By Prof. Juyang Huang, Texas Tech University. All rights reserved. By Prof. Juyang Huang, Texas Tech University. All rights reserved.