Classical Mech Talk

Mechanics and Symplectic Geometry
Ayan Chatterjee
Department of Physics and Astronomical Science,
Central University of Himachal Pradesh
December 14, 2014
Outline
Introduction
Newtons formulation of classical mechanics
Kinematics
Dynamics
Solution and their stability
Configuration manifold and Euler- Lagrange equations
Configuration manifolds
Examples of configuration manifolds
Variational principle
Introduction
I
I shall try to introduce you to the following things:

I
I
While doing so, we shall learn about the geometrical

formulation of classical mechanics, specially, we shall try to
understand the
I
Newtons formulation of classical mechanics.

Lagranges and Hamiltons formulation of classical mechanics.
Configuration manifold Q, tangent bundle T Q, cotangent

bundle T Q.
Symplectic formulation of classical mechanics..
This, I hope, shall help clarify the role and importance of

geometry in mechanics.
Kinematics
I
The goal of classical mechanics is to describe the motion of a

body when acted upon by external forces.
A macroscopic object is described by a collection of points in

space (we take the space to be E3 ).
Thus , one must find a way to specify the location of the

points that make up the object in E3 .
Introduce: reference frame: These are coordinate systems

where each point is given a coordinate.
By which we mean that each point is associated with a set of

three real numbers (called the coordinate of that point) and
the association is unique, different points in space have
different coordinates.
Examples
1. Cartesian coordinate system
2. Cylindrical coordinate system
3. Spherical polar coordinate system.
Given a reference frame, the position of a point is specified by

giving the radius vector X from the origin of the frame to that
point.
X = X1 e1 + X2 e2 + X3 e3
3
X
=
Xi ei := Xi ei
i=1
Y
X2
X1
X3
With motion the coordinates of the particle changes.

X1
t
X3
X2
t2
Define a parameter t which levels the different points of the

particle moves X X(t). The parameter t is taken to be
monotonically increasing. We call this parameter time.
So as time lapses, that is , the parameter increases, we use it

to order events. We say X1 (t1 ) is the position at earlier time
t1 and X2 (t2 ) is the position at later timet2 if t1 < t2 .
the motion of a body is thus given through an equation of a

trajectory written as Xi = Xi (t) where the functions Xi (t) are
dependent on the particular coordinate chosen. Of course one
is free to chose his/ her own comfortable choice of
coordinates.
A few more quantities are also defined for the study of kinematics
I
Distance between two points (X, Y) in E3 with coordinates

(Xi , Yi ) is given through the Pythagorean theorem.
v
u 3
uX
D = t (Xi Yi )2
i=1
Velocity
vi (t) =
Acceleration
dXi
dt
d 2 Xi
dt 2
Kinematics deals with these quantities.
ai (t) =
Dynamics
In whatever we have said before, we have been dealing with

kinematics of points, that is how points move in a space when
specified by a trajectory .
However real objects are not point particles, but we shall
continue to use point kinematics for two reasons.
1. small objects approximate points,
2. results about motion of extended objects can be derived form
axioms (of dynamics) for idealised point particles.
The axioms of dynamics are however not valid everywhere.

1. The size of the objects are not of atomic dimensions 108 m
(quantum mechanics sets it).
2. The velocity of the objects are not high v c (where
relativistic effects set in).
The axioms of classical dynamics are the Newtons laws.
1. There are reference frames called inertial frames where, every

isolated body (on which no force acts) moves in a straight line.
2. F = ma = m ddtX2 .
3. F21 = F12 .
The general problem of mechanics is to solve for X(t) when

t).
F(X, X,
Solution and stability
Solution to the Newtons equation of motion (second order

differential equation) requires a set of initial conditions. Eg.,
at time t0 .
X, X
The solution is unique except in cases where the force may

have a singularity in its derivatives and the initial conditions
are specified at that point. Eg. F (x) = x 1/3 . For initial
conditions x = x = 0 at t = 0, there are three solutions x = 0
and x = (1/6)3/2 t 3 .
Even though the system is deterministic, predicting the final

state requires the knowledge of initial states with infinite
precision. Many systems turn out to be chaotic.
A
B
A
b
a
To understand stability, consider two solutions with initial

conditions X1 (t0 ) and X2 (t0 ) = X1 (t0 ) + X(t0 ).
X1
How does X behave as t increases ? If |X| 0 or a

constant as t , the solution is stable, otherwise unstable.
Instability is not rare, but is common.
Configuration manifolds
In the last class, we reviewed Newtons laws and the basic

formulation of classical mechanics.
The rules were applied to dynamical systems whose motion

takes place in E3 , where it is natural to use the Cartesian
coordinate system.
However, for most dynamical systems, it is convenient to

describe the motion in a space called Q rather than E3 . This
is for the following reason. Most systems (even the one
dimensional ones) are not free to move in E3 , but in a
restricted space. The motion is constrained and its arena of
motion is neither E3 or any simple space (they are not even
Euclidean).
In the space Q, one writes down the Newtons laws but

appropriately modified. This was done by Lagrange and the
resulting formulation is called the Lagrangian formulation of
dynamics.
Lagrangian formulation also applies to free motion in E3 and

is in most of the cases, better suited than Newtons
formulation to obtain the solution.
Let us take an example: Consider a bead which is allowed to

move on a wire palced in E3 . The motion of the bead is not
free. We can use Newtons laws to describe the motion of the
bead if we know the constraint force that forces the bead to
stay on the wire. However, the problem is, most of the time,
we donot exactly know the constraint force and even if we do,
the solution of Newtons laws are very complicated.
Lagrange: Even if we do not know the forces of constraint, we

know the geometrical effects of such forces: Th constraint
force will keep the motion of the system along a particular
geometrical shape. Use this trick.
1. Construct the space which is appropriate /natural space for the
dynamical system to move when constraint forces acts on it.
2. Introduce appropriate coordinates, called generalised
coordinates (denoted by q ), which lie on Q. The number of
coordinates (value of ) is the number of freedom. The
dimension of Q is the number of q s.
3. As the system moves, the trajectory is described by q (t) in Q.
Let us take some examples.
A finite line: Consider a bead on the wire. A curve in 3dimensional space is given by two equations fI (X) = 0,
I = 1, 2. The number of particles is N = 1. Thus the number
of constraints are (K = 2). The number of degrees of freedom
is n = 3N K = 1.
q=l/2
q=0
q=-l/2
Here, Q is of dimension 1 and the coordinate system system

may be chosen to be the coordinate of the wire, such that
2l q 2l .
A Circle: Consider a pendulum, which contains a single

particle at a fixed distance. There are two constraints, (a)
x 2 + y 2 = l 2 and (b) z = 0. The number of degrees of
freedom is 3 2 = 1.
=, -
=-/2
=-/2
=0, 2
Here, Q is of dimension 1, the motion takes place in S 1 . The

generalised coordinate is taken to be the angle .
A plane: Consider a particle which is constrained to move on

a table. The constraint is the equation of a plane given by an
equation in the form f (x, y ) = 0. The number of freedom is
3 1 = 2. Thus Q is two dimensional and can be
conveniently taken to be the plane.
Y
X
Here, Q is of dimension 2, the generalised coordinate is taken

to be the cartesian X , Y or polar r , .
S2 : Consider a spherical pendulum, which contains a single

particle attached to a rod of length l. The rod is free to move
about the center. Constraint is K = 1 and the number of
particles is N = 1. The number of degrees of freedom is 2, Q
Y
is 2 dimensional.
I
Here, Q is S2 , with /2 /2 and .
S2 S2 : Consider a double spherical pendulum, which

contains a two particles. One particle forming a spherical
pendulum and another rod connected to the end of the first
particle, forming another spherical pendulum. The system is
two particles attached back to back. Constraints are K = 2
and the number of particles is N = 2. The number of degrees
of freedom is 4, Q is 4 dimensional.
Here, Q is S2 S2 .
There is a way to rewrite the Newtons laws in terms of these

generalised coordinates. This leads, through the principle of
virtual work, to the Euler- Lagrange equations, which describe
the motion of the system in Q.
However, we shall derive the E- L equations using the

variational principle for two two reasons (a) the principle is
general to be applicable to various other fields like
electrodynamics, quantum mechanics, general , etc. and (b) it
is mathematically elegant.
Variational principle
I
Dynamical systems move so as to extremise the quantity

called the action S:
Z t1
S(q, t0 , t1 ) :=
L(q, q,
t) dt,
(1)
t0
where L(q, q,
t) is called the Lagrangian of the system. Of all
possible trajectory q (t) in the space Q, the the actual
trajectory is the one which extremises the action.
I
S is called a functional of q: when q is put in the L, the

integrand is a function of t, which when integrated over dt
gives a number. The action functional depends on q for all
t0 t t1 .
Consider two trajectories given by q(t; a) and q(t; b) such that

q(t0 ; a) = q(t0 ; b) q(t0 )
q(t1 ; a) = q(t1 ; b) q(t1 )
q(t 1 )
t1
q(t;a)
q(t; 1)
q(t;2)
q(t;b)
q (t 0)
t0
Many such trajectories can be thought of with the end points

remaining the same. Each trajectory gives a different value for
S. The question is: out of all these possible trajectories, what
is the trajectory that particle takes ?
Hamiltons principle: Critical functions of S leads to the EulerLagrange equation, whose solution gives the true trajectory.
Consider the previous figure and a family of many trajectories

q(t; ), all starting and ending at q(t0 ) and q(t1 ), and is an
index labelling each particular trajectory of the family.
Each choice of gives a different trajectory and hence a

different action.
Require that takes its values in the real numbers and

parametrises the family of trajectories continuously and
differentiably. This means that the family of curves, labelled
by forms a connected surface and the partial derivative
q(t;)
exists for all t [t0 , t1 ].

We choose = 0 to denote the physical trajectory. Thus, in

the family, the physical trajectory is given by
I
I
dS
d
=0
d
:=
d
t1
t0

L(q, q,
t)
=0
(2)
=0
Note that the variation in L is due to the dependence of

q(t; ).
d
Denote d
:=
=0
R t1
Then, S = t0 L(q, q,
t) = 0 is the formulation of the
problem.
The solution is given by the Euler- Lagrange equations:
d
dt
I
L
q
L
= 0.
q
Eg: Simple harmonic oscillator L = 21 ml 2 2 + mgl cos .
I L
= mgl sin , and
= ml 2 .
Then,the E- L eqn is ml 2 + mgl sin = 0 whose solution

gives the motion of the pendulum, that how the pendulum
moves from one point of the circle to the next point.
(3)

Classical Mech Talk

Uploaded by

Copyright:

Available Formats

Classical Mech Talk

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Classical Mech Talk

Uploaded by

Copyright:

Available Formats

Mechanics and Symplectic Geometry

December 14, 2014

I shall try to introduce you to the following things:

While doing so, we shall learn about the geometrical

Newtons formulation of classical mechanics.

Configuration manifold Q, tangent bundle T Q, cotangent

This, I hope, shall help clarify the role and importance of

The goal of classical mechanics is to describe the motion of a

A macroscopic object is described by a collection of points in

Thus , one must find a way to specify the location of the

Introduce: reference frame: These are coordinate systems

By which we mean that each point is associated with a set of

Given a reference frame, the position of a point is specified by

With motion the coordinates of the particle changes.

Define a parameter t which levels the different points of the

So as time lapses, that is , the parameter increases, we use it

the motion of a body is thus given through an equation of a

Distance between two points (X, Y) in E3 with coordinates

In whatever we have said before, we have been dealing with

The axioms of dynamics are however not valid everywhere.

The axioms of classical dynamics are the Newtons laws.

1. There are reference frames called inertial frames where, every

The general problem of mechanics is to solve for X(t) when

Solution and stability

Solution to the Newtons equation of motion (second order

The solution is unique except in cases where the force may

Even though the system is deterministic, predicting the final

To understand stability, consider two solutions with initial

How does X behave as t increases ? If |X| 0 or a

Instability is not rare, but is common.

In the last class, we reviewed Newtons laws and the basic

The rules were applied to dynamical systems whose motion

However, for most dynamical systems, it is convenient to

In the space Q, one writes down the Newtons laws but

Lagrangian formulation also applies to free motion in E3 and

Let us take an example: Consider a bead which is allowed to

Lagrange: Even if we do not know the forces of constraint, we

Let us take some examples.

Here, Q is of dimension 1 and the coordinate system system

A Circle: Consider a pendulum, which contains a single

Here, Q is of dimension 1, the motion takes place in S 1 . The

A plane: Consider a particle which is constrained to move on

Here, Q is of dimension 2, the generalised coordinate is taken

S2 : Consider a spherical pendulum, which contains a single

Here, Q is S2 , with /2 /2 and .

S2 S2 : Consider a double spherical pendulum, which

There is a way to rewrite the Newtons laws in terms of these

However, we shall derive the E- L equations using the

Dynamical systems move so as to extremise the quantity

S is called a functional of q: when q is put in the L, the

Consider two trajectories given by q(t; a) and q(t; b) such that

Many such trajectories can be thought of with the end points

Consider the previous figure and a family of many trajectories

Each choice of  gives a different trajectory and hence a

Require that  takes its values in the real numbers and

We choose  = 0 to denote the physical trajectory. Thus, in

Note that the  variation in L is due to the  dependence of

The solution is given by the Euler- Lagrange equations:

Eg: Simple harmonic oscillator L = 21 ml 2 2 + mgl cos .

Each choice of gives a different trajectory and hence a

Require that takes its values in the real numbers and

We choose = 0 to denote the physical trajectory. Thus, in

Note that the variation in L is due to the dependence of