You are on page 1of 75

Principle of Least Action

Manoj K. Harbola
Department of physics

Acknowledgement: Varun
LEAST ACTION HERO
DOES A RAY OF LIGHT KNOW WHERE IT'S GOING?
(Jim Holt, Lingua Franca vol. 9, No. 7,October 99)

Suppose you are standing on the beach, at some distance from the water.
You hear cries of distress. Looking to your left, you see someone
drowning. You decide to rescue this person. Taking advantage of your
ability to move faster on land than in water, you run to a point at the edge
of the surf close to the drowning person, and from there you swim directly
toward him. Your path is the quickest one to the swimmer − but it is not a
straight line. Instead, it consists of two straight-line segments, with an
angle between them at the point where you enter the water.
Which path does a particle of total energy E traveling from A to B take?

P.E.=0 h1

h2
P. E.= V>0
B
d

1. path of least distance 2. path of least action

3. path of least time


What does the path of least distance give us?

Least distance means motion in a


straight line which implies P.E.=0
h1
θ1

θ1=θ2
θ2
h2
P. E.= V>0

B
d
What does the path of least action give us?

( )
B

∫ = + + − − +
2 2 2 2
Action = vds 2 mE ( x h1 ) 2 m ( E V ) ( d x ) h2
A

A
minimization of action with respect to x gives

x P.E.=0
x +h
2 2
(E − V ) h1
θ1
1
=
(d − x) E
(d − x) 2 + h22 θ2
x
h2
which is equivalent to P. E.= V>0

B
d
sin θ1 (E − V ) v2
= = <1
sin θ 2 E v1
What does the path of least time give us?

ds x 2 + h12 (d − x) 2 + h22
Total time =
∫v= 2mE
+
2m( E − V )

A
minimization of time with respect to x gives

x P.E.=0
x +h
2 2
E h1
1
= θ1
(d − x) (E − V )
(d − x) 2 + h22
θ2
x h2
which is equivalent to P. E.= V>0

B
d
sin θ1 E v
= = 1 >1
sin θ 2 (E − V ) v2
When particle strikes the surface, the component of velocity along
the surface remains unchanged
A
v1sinθ1

v1
v1 sin θ1 = v 2 sin θ 2
θ1

sin θ1 v 2
v2sinθ2 = <1
θ2 sin θ 2 v1

v2
B
A

θ1 Trajectory of the particle is the


path of least action

θ2

B
Principle of least action

When a particle of fixed energy travels from point A to point B, its


trajectory is such that the corresponding action has the minimum
possible value.

For motion in a straight line


Action=area

x
A B
Test case 1: Can a particle traveling in a straight line from A to B suddenly
reverse its direction of motion, go back for some distance, reverse its motion
again and reach point B?

Principle of least action prevents that from happening

v v

A x B A B x
Test case 2: A cricket ball hit so that it reaches a fielder

y ∆y

Let actual path be y(x)

Let a nearby path be y(x)


x
2
 dy 
B B
2
Action for the actual path y(x) A = ∫ vds = ∫ ( E − mgy ) 1 +   dx
A A
m  dx 

Change in action for a nearby path y(x)

2 
B B
δA = δ ∫ vds = δ ∫  ( E − mgy ) 1 + y′2  dx
A 
A
m

must be zero if action for the actual path is minimum


Change in action arises from:

d 
(i) Change in the speed δ ( E − mgy ) =  ( E − mgy ) δ y ( x)
 dy 

(ii) Change in the length of trajectory

 d 2 
δy(x+∆x)
δ 1 + y′ dx = 
2
1 + y′ δy′dx
 dy ′ 
δy(x)
where
δ y ( x + ∆x) − δ y ( x) d
δ y′ = = δ y ( x)
x x+∆x ∆x dx

Change in action therefore is

B
 d d d 
δA = ∫  1 + y′2 ( E − mgy )δ y ( x) + ( E − mgy ) 1 + y′2 δ y ( x) dx
A  dy dy′ dx 
Integration by parts leads to

B
 2 d d  d 2 

δA = ∫  1 + y′ ( E − mgy ) −  ( E − mgy ) 1 + y′  δ y ( x)dx
A
dy dx  dy′ 

Since δy(x) is arbitrary, δA=0 implies

d d  d 2 
1 + y′
2
( E − mgy ) −  ( E − mgy ) 1 + y′  = 0
dy dx  dy′ 

This simplifies to

2 y′′( E − mgy ) + mg (1 + y′2 ) = 0


Equation of the trajectory

2 y′′( E − mgy ) + mg (1 + y′2 ) = 0


Integration of the equation leads to

y′2 = C1 ( E − mgy ) − 1

This gives
(C1 E − 1) C1mg
y= − ( x + C2 ) 2
C1mg 4

Conditions y(0)=0 and y(a)=0 leads to

(C1 E − 1) C1mg 
2
a
y= −  x − 
C1mg 4  2
C1 E − 1
Put y’=0 in y′2 = C1 ( E − mgy ) − 1 to get ymax = and
C1mg
2
mg  a
y = ymax −  x − 
4( E − mgymax )  2

2E ± 4E 2 − m2 g 2a 2
Again the condition y(0)=0 gives ymax =
4mg

Thus there are two parabolic trajectories that the ball can take

ymax2

ymax1

a x
Comparison with Newtonian approach:

Given initial position and velocity of a particle, Newtonian method


builds up its trajectory in an incremental manner by updating the
velocity and position. Energy of the particle may or may not be fixed.

Principle of least action says if a particle of fixed energy has to go from


point A to point B, the path it takes is that which minimizes the action.

But this can't be right, can it? Our explanation for the route taken by the
light beam (particle in our case) − first formulated by Pierre de Fermat
in the seventeenth century as the principle of least time (principle of
least action in the present case) − assumes that the light (particle)
somehow knows where it is going in advance and that it acts
purposefully in getting there. This is what's called a teleological
explanation. (Jim Holt)
The idea that things in nature behave in goal-directed ways goes back
to Aristotle. A final cause, in Aristotle's physics, is the end or telos
toward which a thing undergoing change is aiming. To explain a change
by its final cause is to explain it in terms of the result it achieves. An
efficient cause, by contrast, is that which initiates the process of
change. To explain a change by its efficient cause is to explain it in
terms of prior conditions.

One view of scientific progress is that it consists in replacing


teleological (final cause) explanations with mechanistic (efficient
cause) explanations. The Darwinian revolution, for instance, can be
seen in this way: Traits that seemed to have been purposefully
designed, like the giraffe's long neck, were re-explained as the
outcome of a blind process of chance variation and natural selection.

(Least Action Hero, Jim Holt, Lingua Franca vol. 9, No. 7, October 99)
Plan of the talk:
Aristotle and the motion of planets

Reflection of light and Hero of Alexandria

Fermat’s principle of least time for light propagation; Descartes


versus Fermat

Wave theory and Fermat’s principle

Maupertuis’ principle of least action

Euler-Lagrange formulation

Hamilton’s investigations

Quantum connections
ARISTOTLE
(384-322 BC)
Aristotle on the motion of planets:
If the motion of the heaven is the measure of all movements
whatever in the virtue of being alone continuous and regular
and eternal, and if, in each kind, the measure is the
minimum, and the minimum movement is the swiftest,
then clearly, the movement of the heaven must be the
swiftest of all movements. Now of lines which return upon
themselves the line which bounds the circle is the shortest;
and that movement is the swiftest which follows the
shortest line. Therefore, if the heaven moves in a circle
and moves more swiftly than anything else, it must
necessarily be spherical.
REFLECTION OF LIGHT
&
HERO OF ALEXANDRIA
(125 BC)
Whatever moves with unchanging velocity moves in a
straight line…. For because of the impelling force the object
in motion strives to move over the shortest possible
distance, since it has not the time for slower motion, that is,
for motion over a longer trajectory. The impelling force does
not permit such retardation. And so, by reason of its speed,
the object tends to move over the shortest path. But the
shortest of all lines having the same end points is a straight
line……Now by the same reasoning, that is, by a
consideration of the speed of the incidence and the
reflection, we shall prove that these rays are reflected at
equal angles in the case of plane and spherical mirrors. For
our proof must again make use of minimum lines.
Proof:
A
Let a light ray start from point A
B
and reach point B after reflection.
The true path is AOB such that
rays AO and BO make equal O1 O C
angles from the mirror.

Draw an alternate path AO1B


B1

Drop a perpendicular BC on the mirror and extend it to B1 so that BC=B1C.


Join O and B1 and O1 and B1.

From congruency of ∆BOC and ∆B1OC and the fact that AO and BO
make equal angles from the mirror, it follows that AOB1is a straight line.

In ∆AO1B1: AO1+O1B1 > AB1( = AO+OB1=AO+OB)

Path AOB is the shortest


REFRACTION OF LIGHT
&
FERMAT’S PRINCIPLE OF
LEAST TIME
Newton (1642-1727)

Refraction of light and Snell’s law (1621):

sin θ1
θ1 = constant
medium 1
sin θ 2

θ2 For medium 2 denser than medium 1


medium 2
constant > 1

Historical note: There is evidence that Thomas Hariot in England had


also discovered the same law around 1600.
Descartes (1596-1650)
Newton (1642-1727)
Descartes’ explanation (1637) of Snell’s law:

Argument given by Descartes is a mechanical one, based on the fact that the
component of velocity along the surface remains unchanged

v1sinθ1
v1 v1 sin θ1 = v 2 sin θ 2
θ1

v2sinθ2 sin θ1 v2
= constant =
θ2 sin θ 2 v1
v2

Descartes explained the constancy of the ration of sine of angles in terms of


the ratio of the speed of light in the two media

By Descartes’ explanation, light had to be moving faster in the denser


medium
Fermat (1601-1665)
Newton (1642-1727)
Fermat’s principle of least time (1658):

Generalization of Hero’s explanation of reflection to include refraction of light.

During refraction a light ray does not take the path of least distance; that
would be a straight line.
Between two points, a light ray travels in such a manner that it take the
least time.
For reflection this leads to equal angles of incidence and reflection

A For refraction this implies


θ1

sin θ1 v
= constant = 1
θ2 sin θ 2 v2
B

By Fermat’s principle of least time, light moves slower in the denser


medium.
Descartes (1596-1650)
Fermat (1601-1665)
Descartes versus Fermat: Newton (1642-1727)

Descartes believed that light traveled infinitely fast.

Fermat on Descartes:
1. “of all the infinite ways to analyze the motion of light the author has
taken only that one which serves him for his conclusion; he has
therefore accommodated his means to his end, and we know as little
about the subject as we did before.”
2. rejects Descartes’ assertion of infinite speed of light and therefore his
illogical conclusion that light travels faster in water than in air .

According to Fermat light traveled at finite speed in air and slowed down
in water.

Experimental verification of finiteness of speed of light – 1675 by Roemer


Measurement of speed of light in water – 1850 by Fizeau and Foucault
These observations CONFIRM Fermat’s principle of least time
Objections of Cartesians to Fermat’s theory

From Clerselier’s letter to Fermat:


“That path, which you reckon the shortest because it is the quickest, is
only a path of error and bewilderment, which Nature in no way follows and
cannot intend to follow. For, as Nature is determinate in everything she
does, she will only and always tend to conduct her works in a straight line”

Clerselier on the velocity of light:


“M. Descartes − in 23rd page of his Dioptrique − proves and does not
simply suppose, that light moves more easily through dense bodies than
through rare one”
Fermat’s reply

“I have often said to M. de la Chambre and yourself that I do not claim


and that I have never claimed, to be in the private confidence of
Nature.

She has obscure and hidden ways that I have never had the initiative
to penetrate; I have merely offered her a small geometrical assistance
in the matter of refraction, supposing that she has need of it.

But since you, Sir, assure me that she can conduct her affairs without
this, and that she is satisfied with the order that M. Descartes has
prescribed for her, I willingly relinquish my pretended conquest of
physics and shall be content if you will leave me with a geometrial
problem, quite pure and in abstracto, by means of which there can be
found the path of a particle which travels through two different media
and seeks to accomplish its motion as quickly as it can.”
HUYGENS’ WAVE THEORY
&
FERMAT’S PRINCIPLE
Newton (1642-1727)

Huygens’ wave theory (1629-1695) and the principle of least time

According to Huygens’ theory, light travels as a wave with path of light ray
being in the direction perpendicular to the wavefront

Consider the true ray path from A to B


B and also an alternate path. Because
true path is perpendicular to the
wavefronts,
A
AB (true path) < AB (alternate path)
Actual principle is the principle of stationary (minimum, maximum
or saddle point) time

Consider a light beam that starts in a


B diverging manner from point A and then
converges to point B.

All nearby paths of light should not have


A different time of arrivals implying
stationarity of time of travel.

Example 1: In an elliptical mirror, light starting from one focus of the


ellipse and reaches the other focus after reflection from the mirror. There
are many possible paths for this and all of them are equal.

F1
F2
Example 2: An elliptical mirror

minimum time maximum time

stationary time
minimum time

Example 3: A circular mirror 45°


A B
60°

stationary time
How does light know which path to take?

Wave Theory: Light does not know which path to take. It takes all
possible paths with certain probability amplitude and phase and these
probability amplitudes interfere. The phase depends on the time of
passage.

Path of least time is where the interference is constructive to the


largest extent possible. This is because phase for other nearby paths
does not vary much.

Light source Image


Does light really go through all possible paths?

Light source No light

Experiment : Take a point


different from the image.
No light from the source
reaches the that point.

Light source Image


Now blacken parallel strips
on the mirror to remove light
of opposite phase. It forms
a grating and light reaches
many different points.
PRINCIPLE OF LEAST ACTION
MAUPERTUIS, LAGRANGE & EULER
Newton
(1642-1727)
Principle of Least Action

Nature acts in a way so that it renders a quantity called action a


minimum
Maupertuis in “The agreement between different laws of Nature that had,
until now, seemed incompatible” read on April 15, 1744 to Académie des
sciences.

Comment: Maupertuis’ attempts was to explain the propagation of light


and movement of a particle by a single principle.

Action is defined as the product of the mass, the velocity and the
distance.

Action = m × v × s
Example 1: To find the final velocity of two masses involved in a perfect
inelastic collision.

u1 u2 v
m1 m2

Treat distance as that covered in one second, that is as velocity.

change in action = m1 (u1 − v) + m2 (u 2 − v)


2 2

Minimization of change in action with respect to v leads to

m1u1 + m2u2
v=
m1 + m2
Example 2: To find the relationship between final velocities of two
masses involved in a perfect elastic collision.

m1 u1 u2 m1 v1 v2
m2 m2

change in action = m1 (u1 − v1 ) + m2 (u 2 − v 2 )


2 2

In a perfect elastic collision: v1 − v 2 = u2 − u1

Minimization of change in action with respect to v1 leads to

m1u1 + m2u2 = m1v1 + m2 v 2


Example 3: Refraction of light

A Action = v1 × AO + v 2 × OB
θ1
O
Minimization of action leads to
θ2
sin θ1 v 2
B = = constant
sin θ 2 v1

Snell’s law is verified through the principle of least action, and agrees
with Descartes’ conclusions.
Lagrange (1736-1813) on Maupertuis’ principle in Mécanique Analytique, 1788

“This principle, looked at analytically, consists in that, in the


motion of bodies which act upon each other, the sum of the product of
the masses with the velocities and with the distances travelled is a
minimum. The author deduced from it the laws of reflection and
refraction of light, as well as those of the impact of bodies.
But these applications are too particular to be used for
establishing the truth of a general principle. Besides, they have
somewhat vague and arbitrary character, which can only render the
conclusions that might have been deduced from the true correctness of
the principle unsure………………
But there is another way in which it may be regarded, more
general, more rigorous, and which itself merits the attention of the
geometers. Euler gave the first hint of this at the end of his Traité des
isopérimètres, printed at Lausanne in 1744.”
Principle of Least Action in Mechanics

Proper mathematical foundation is provided by Euler (1744)

Before paying attention to this problem, Euler had already developed


Calculus of Variations and given the Euler condition for making the
variation of an integral of the form
( x2 , y 2 )

∫ f ( y, y′, x)dx
( x1 , y1 )
between two fixed points (x1,y1) and (x2,y2) vanish with respect to arbitrary
variations in y(x)

Euler condition

d  ∂f  ∂f
  − =0
dx  ∂y′  ∂y
Consider a particle moving in xy plane under the influence of a force with
x-component Fx and y-component Fy

Y Fy
Centripetal force

Fx mv Fx y′ − Fy
2
=
r 1+ y ′ 2

Aim is to see if the principle of least action gives the same answer
Action = m ∫ v ds = m ∫ v 1 + y dx
′ 2

Euler condition gives

d  ∂

dx  ∂y′
v (
1 + y ′ )
2 
 −
 ∂y

v 1 (
+ y ′ 2
=0)
Use the relations

∂ 1 2 ∂ 1 2
 mv  = Fx ,  mv  = Fy
∂x  2  ∂y  2 

and
d ∂ ∂
= + y′
dx ∂x ∂y
mv y′′
2
− Fx y′ + Fy
=
(
1 + y′ )
2 32
1+ y ′ 2

Y radius of curvature r=
(
1 + y′ ) 2

Fy
y′′

mv 2 Fx y′ − Fy
Fx

=
r 1 + y′ 2

MAKING ACTION STATIONARY LEADS TO THE


CORRECT FORCE BALANCE EQUATION
What does the minimum action principle imply for one-dimensional motion?

x
δx
A = ∫ v( x) dx

δA = ∫ δv( x) dx + ∫ v( x)δ (dx)


t

 ∂ 2
δ v( x) = (E − U ( x ) ) δ x
 ∂x m
 1 ∂U ( x)
 = − δx
2m(E − U ( x) ) ∂x
If the total energy is constant

  1  ∂U ( x)
 = −  δx
  mv  ∂x
δ (dx) = δ ( x + ∆x) − δx
δ(∆x) ∆(δx)
δ(x+∆x) = {( x + ∆x ) − ( x + ∆x)}− {x − x}
x ∆x
δx
= ∆x − ∆x
∆x
= ∆(x − x )
= d (δ x)
∆t

If the end points of the trajectory are kept fixed

 1  ∂U ( x)
δA = ∫ −   δxdx + ∫ vd (δx)
 mv  ∂x
 1  ∂U ( x)
= ∫−  δxdx + ∫ dvδx
 mv  ∂x
 1  ∂U ( x)
δA = ∫ −   δ xdx + ∫ dvδ x
 mv  ∂x
Since δx(t) is arbitrary, δA=0 implies that

 1  ∂U ( x) ∂U ( x) dv dx
−  dx = dv or − =m using = dt
 mv  ∂x ∂x dt v
Making Action stationary is absolutely equivalent to a particle’s
equation of motion

For many interacting particles also, minimization of ∑ ∫ m v ds


i
i i i
leads to the equation of motion (Lagrange)

∂U ({ri }) dv iα
− = mi
∂riα dt
Making Action stationary is equivalent to Newton’s second law
Does the principle of least action teach us anything new?

In a characteristic way, the principle of least action did not at first


exercise an appreciable effect on the advance of science, even after
Lagrange had completely established it as a part of mechanics. It was
considered more as an interesting mathematical curiosity and an
unnecessary corollary to Newton’s laws of motion. Even in 1837
Poisson could only call it “a useless rule”. (From an essay by
Planck)

This, however, changed when Hamilton (1805-1865) entered the scene


in 1830’s
HAMILTON’S PRINCIPLE OF
VARYING ACTION
PARALLEL BETWEEN GEOMETRIC OPTICS AND MECHANICS
Consider the action integral

∆s A( x) = ∫ v(x1 , x 2 , x 3 ) ds
as a function of the end points of the true path.
The integral is obviously taken along the true path.

As the path is increased by ∆s to the next point, we have

dA
∆A = v( x)∆s OR = v( x )
ds

Can this equation be used instead to find the path taken by light?
Conventional approach (Hamilton):

Function v(x1,x2,x3;α1,α2,α3) is considered to be a function of the directional


cosines {αi} of the ray of light. Making action stationary with respect to
variations in {xi; αi} gives the equation for {αi} .

Recall that in an earlier minimization, the integrand was taken to be function of


y(x) and y′(x). Thus
A = ∫ v( y ) 1 + y′2 dx
y(x) is then found by making the variation of the action vanish with respect to
variations δy(x).

Now the independent variables are taken to be {xi; αi} instead. Thus

A = ∫ v({xi ; α i }) ds
{αi(x)} are found by making the variation of the action vanish with respect to
variations {δxi} and {δαi}. Using α12 + α 22 + α 32 = 1 make v({xi;αi}) homogeneous
Of degree 1 in {αi}
Take the true path and a varied path around it
dx+δ(dx) obtained by shifting the line element by {δxi);
by changing its length by δ(ds) and by
δy
changing its directional cosines by {δαi}

dx
δA = δ ∫ v ds = ∫ δ v ds + ∫ vδ (ds )
δx

 ∂v ∂v 
δA = ∑ ∫  δ xi ds + δα i ds  + ∫ vδ (ds )
i  ∂xi ∂α i 

δα i ds + α i δ (ds ) = δ (α i ds ) = δ (dxi ) = d (δ xi )
⇒ δα i ds = d (δ xi ) − α i δ (ds )
 ∂v ∂v   ∂v 
δA = ∑ ∫  δ xi ds + d (δxi )  + ∫  v − ∑ α i δ (ds )
i  ∂xi ∂α i   i ∂α i 

Now consider the variation of action integral


between the true path and the varied path

Integration by parts leads to


 ∂v 1  ∂v 0  ∂v ∂v 
δA = ∑  δ xi  − ∑  δ xi  + ∫  ds − d δ xi
i  ∂α i  i  ∂α i   ∂xi ∂α i 
 ∂v 
+ ∫  v − ∑ α i δ (ds )
 i ∂α i 
δx1(0)= variation of coordinate at the final (initial) point of the path
 ∂v 1  ∂v 0  ∂v ∂v 
δA = ∑  δ xi  − ∑  δ xi  + ∫  ds − d δ xi
i  ∂α i  i  ∂α i   ∂xi ∂α i 
 ∂v 
+ ∫  v − ∑ α i δ (ds )
 i ∂α i 

Demand that δA vanish for arbitrary variations with the end points of the path
fixed i.e. δx1/0=0. This gives

∂v
v = ∑αi since v is a homogeneous function of {αi} of degree 1
i ∂α i

Thus v( x;α ) = v( x1 , x2 , x3 ) α 12 + α 22 + α 32 Note: α12 + α 22 + α 32 = 1

∂v ∂v
Differential equation for the path of light ray ds = d
∂xi ∂α i
Now consider the integral

 ∂v 1  ∂v 0
δ ∫ v ds = ∑  δ xi  − ∑  δ xi 
i  ∂α i  i  ∂α i 
with the initial point fixed and δx1 non zero and in the direction of the path.

δ xi1 = α iδ s (δα i = 0)
 ∂v 1  ∂v 
Then δ ∫ v ds = ∑  δ xi  = ∑  α iδ s  = vδ s
i  ∂α i  i  ∂α i 
The action
∫ vds is a function of the end points of the true path
CONCLUSION: Stationary action implies existence of a characteristic
function A(x) such that

∂A( x) ∂v
=
∂xi
1
∂α i
How to find the path if A(x) and v(x,α) are given?

∂A( x) ∂v
From the equation = solve for (α1 , α 2 , α 3 ) as a function of
∂xi
1
∂α i
(x1,x2,x3)

Differential equation for the characteristic function A(x)

∂A( x) ∂v
= = αi v
∂xi
1
∂α i

α12 + α 22 + α 32 = 1
2 2 2
 ∂A( x)   ∂A( x)   ∂A( x) 
  +   +   = v 2 ( x)
 ∂x1   ∂x2   ∂x3 
v(x) is the refractive index of the medium
Direction of light ray and surfaces of constant Action:

Direction of light ray α1 xˆ1 + α 2 xˆ2 + α 3 xˆ3 = v(α1 xˆ1 + α 2 xˆ2 + α 3 xˆ3 )
∂v ∂v ∂v
= x1 +
ˆ x2 +
ˆ xˆ3
∂α1 ∂α 2 ∂α 3
∂A ∂A ∂A
= xˆ1 + xˆ2 + xˆ3
∂x1 ∂x2 ∂x3

= ∇A( x)
Thus light ray moves in the direction of the gradient of the characteristic
function

EQUIVALENTLY
If the points of equal action for each ray are joined together, light rays
move in the direction perpendicular to surface so formed i.e.
perpendicular to the surfaces of constant action

A( x1 , x2 , x3 ) = constant
Light rays and surfaces of constant action

light rays
ACTION
is

surfaces of constant action EQUIVALENT


to
Huygens theory of light waves
SPACE-PART
of
light rays PHASE

wavefronts
Mechanical systems:

In going from point A to point B, a particle also satisfies the principle of


least action

δ ∫ v ds = 0

v = v 12 + v 22 + v 32 = v ( x1 , x2 , x3 ) α 12 + α 22 + α 32 is the speed of the particle

Thus there exists a characteristic function A(x) for a mechanical system


also such that

∂A({xi })  
vi = OR v = ∇A({xi })
∂xi
And the path of a particle can be determined if we know the characteristic
function
Equation for the characteristic function

2 2 2
 ∂A( x)   ∂A( x)   ∂A( x) 
  +   +   = v 2 ( x)
 ∂x1   ∂x2   ∂x3 

From the energy conservation equation

1 2
E = mv + U ({xi })
2

Thus the equation for the characteristic function is

2 2 2
 ∂A( x)   ∂A( x)   ∂A( x) 
  +   +   + 2mU ({xi }) = 2mE
 ∂x1   ∂x2   ∂x3 
Example: A projectile thrown with initial speed v0 at an angle φ0 in a
gravitational field
U ( y ) = mgy
The equation for the characteristic function

2
 ∂A   ∂A 
2

  +   = m( v 0 − 2 gy )
2

 ∂x   ∂y 
Solve the equation by separation of variables to get

1 2 1 2
A( x, y ) = (v 0 − k ) + k x −
2 32
( v 0 − k 2 − 2 gy ) 3 2
3g 3g
Values of A(x,y) are obtained by substituting for x and y, the coordinates of
a trajectory
Action
1 2 1 2
A( x, y ) = (v 0 − k ) + k x −
2 32
( v 0 − k 2 − 2 gy ) 3 2
3g 3g
Velocity of the projectile
∂A
vx = = k = v 0 cos φ0
∂x
∂A
vy = = v 02 − k 2 − 2 gy = v 02 sin 2 φ0 − 2 gy
∂y

Integration of these equations leads to


1 2
x = v 0 cos φ0 t y = v 0 sin φ0 t − gt
2
These give the trajectory and the action for projectile motion
Surfaces of constant action

Trajectories are lines perpendicular to the surfaces of constant action


Trajectories of the projectile

Mechanical motion of a particle is like the motion of ray of light and


therefore equivalent to Geometric optics.

Question: Can we associate the action of a particle with a phase?


QUANTUM CONNECTIONS
Fast forward to 1920s: When it was discovered that particles have a
wave associated with them, Hamilton’s theory became the natural choice to
account for it and develop the quantum-mechanical wave equation.

How Schrödinger obtained the wave equation (Ist paper by Schrödinger)

Start with the equation for the action

2 2 2
 ∂A( x)   ∂A( x)   ∂A( x) 
  +   +   + 2mU ({xi }) = 2mE
 ∂x1   ∂x2   ∂x3 

Treating action like phase, take the wavefunction Ψ as

Ψ = exp( A( x) / K ) OR A( x) = K log Ψ

Substitute this wavefunction in the equation for action to obtain a quadratic


form in Ψ, which is equal to zero
2 2 2
 ∂Ψ ( x)   ∂Ψ ( x)   ∂Ψ ( x) 
  +   +   + 2m(U ({xi }) − E ) Ψ 2 = 0
 ∂x1   ∂x2   ∂x3 
Rather than looking for solutions of this equation, seek a function Ψ such that
the integral of the quadratic form above over all space is stationary for any
arbitrary variations of Ψ.

 ∂Ψ ( x)  2  ∂Ψ ( x)  2  ∂Ψ ( x)  2  
δ ∫   +   +   + 2m(U ({xi }) − E ) Ψ  dr = 0
2

 ∂x1   ∂x2   ∂x3  

For well-behaved Ψ vanishing at infinity, this leads to the Schrödinger equation

K2 2
− ∇ Ψ +U Ψ = E Ψ
2m
Direct connection (IInd paper by Schrödinger):

Space part of the phase of matter waves = A(x)

E ; h = Planck’s constant
Frequency of the waves =
h

Calculate the phase velocity of the wave treating surfaces of constant


action as wavefronts

 mA( x) − Et 
φ = 2π   ; m = mass of the particle
 h 
As the wavefront moves with phase velocity uphase covering distance ∆x in
time ∆t, we have

∆φ = m∆A( x) − E∆t = 0
∂A
∆φ = m∆A( x) − E∆t = 0 ∆A( x) = ∆x = v particle ∆x
∂x
This gives

∆x E E
u phase = = =
∆t mv particle 2m( E − U )

u phase
h h
u phase ≠ v particle AND λ = = =
E h 2m( E − U ) m v particle
Group velocity of the waves

E
∂ 
∂ω ∂E
=  =
h
u group = = v particle
∂k  1  ∂ 2m( E − U )
∂ 
λ
And finally the wave equation

1 ∂ Ψ
2
− ∇ 2
Ψ=0
u 2phase ∂t 2

Substituting

E  E 
u phase = AND Ψ ( x; t ) = Ψ ( x) exp − i 2π t 
2( E − U )  h 

leads to the Schrödinger equation

h2
− ∇ 2 Ψ + U Ψ = EΨ
8π 2 m
A comparison between Classical and Quantum Mechanics (Feynman):

Classical Mechanics: Path of a particle is that of least action and therefore


normal to the surfaces of constant action

Quantum Mechanics: Because of the waves associated with a particle, it


does not know which path to take. It takes all possible paths with certain
probability amplitude and phase and these probability amplitudes interfere.
The phase depends on the action.

Path of least action is where the interference is constructive to the largest


extent possible.

Classically we see only those result when amplitudes interfere


constructively giving a large final amplitude
Planck on the Principle of Least Action :

As long as physical science exists, the highest goal to which it aspires is


the solution of the problems of embracing all natural phenomena,
observed and still to be observed, in one simple principle which will allow
all past and, especially, future occurrences to be calculated.

Among the more or less general laws, the discovery of which characterize
the development of physical science during the last century, the principle
of Least Action is at present certainly one which, by its form and
comprehensiveness, may be said to have approached most closely to the
ideal aim of theoretical inquiry.
Thank
you
According to Planck:
"on this occasion everyone has to decide for himself which point of view he
thinks is the basic one."

You can be a teleologist if you wish. You can be a mechanist if that better
suits your fancy. Or you may be left wondering whether this is yet another
metaphysical distinction that does not make a difference.

You might also like