Theory of Collisional Mechanics

CHAPTER 1
THEORY OF EQUABILITY
(1) POSTULATE – 1
Work done can be defined as the product of difference
of rest mass from moving mass of a body and the
difference of square of velocity of that moving body
from the square of velocity of light in vacuum (c=3*10 8
m/s).
Work done(w) = (m-m0)(c2-v2)
Where,
m=moving mass of a body
m0=rest mass of a body
v=velocity of that body
c=velocity of light in vacuum.
Proof: According to Einstein Special Theory of Relativity
the relativistic moving mass of a body is given by
m=m0/[1-(v2/c2)]1/2 --------------------------(1)
Differentiating both sides w.r.t to ‘t’
dm/dt = d(m0/[1-(v2/c2)]1/2)/dt
dm/dt = d(m0c/[c2-v2]1/2)/dt
dm/dt = {[c2-v2]1/2 d(m0c)/dt – m0c d([c2-
v2]1/2)/dt}/(c2-v2)
dm/dt = {0-m0c*(1/{2[c2-v2]1/2})(-2va)}/(c2-v2)
[since,m0,c are constants]
(c2-v2) dm/dt=(vacm0)/([c2-v2]1/2)
Also, [c2-v2]1/2 = (m0c)/m
(c2-v2) dm/dt = (m0cavm)/(m0c)
Multiplying both sides by dt we get
(c2-v2) dm = ma*vdt
Integrating both sides we get,
(c2-v2) dm=ma*v dt
(c2-v2)(m-m0)=mavt ----------------------------------------(2)
Now, we know f=ma and s(distance) = vt
(c2-v2) (m-m0) =f*s
We know, w=f*s
Hence,
w=(c2-v2) (m-m0)--------------------------------------------(3)
(1.1) RELATION BETWEEN RELATIVISTIC KINETIC

ENERGY WITH RELATIVISTIC WORK DONE.
The formula for relativistic Kinetic Energy of a
particle of rest mass m0 is given by
k=c2[m0/[1-(v2/c2)]1/2-m0]---------------------(4)
where k is the kinetic energy of a particle.
Now,
m=m0/[1-(v2/c2)]1/2
Where,
m is the moving mass of the particle moving with
velocity v.
k=c2(m-m0)
 (m-m0)=k/c2 --------------------------------------(5)
we know that,
relativistic work done (w)
=(m-m0)(c2-v2)
 Putting the value of (m-m0) from equation (5) we
get
w=(k)*[1-(v2/c2)]
w=k[1-(v2/c2)]----------------------------------------(6)
(1.2) DEDUCING THE CLASSICAL EXPRESSION FOR
KINETIC ENERGY.
The formula for relativistic work done of a particle
of rest mass m0 is given by
w=(c2-v2) (m-m0)
But, m=m0/[1-(v2/c2)]1/2
Where,
velocity v.
w=[m0(1-{v2/c2})-1/2-m0][c2-v2]
 w=m0[(1-{v2/c2})-1/2-1][c2-v2] ----------------------(7)
Expanding (1-{v2/c2})-1/2 by binomial theorem we get,
(1-{v2/c2})-1/2 = 1 + (1/2)(v2/c2)+(3/8)(v4/c4)+……
For small velocities v<<c, the terms beyond (1/2)
(v2/c2) are negligible.
Then,
(1-{v2/c2})-1/2 1 + (1/2)(v2/c2)
Making this substitution in equation (7) we obtain,
w m0 [1 + (1/2)(v2/c2)-1] [c2-v2]
w (1/2)(m0v2) [(c2/c2)-(v2/c2)]
w (1/2)(m0v2)[1-( v2/c2)]
 Comparing the above expression with the equation
(6) we get,
k (1/2)(m0v2)--------------------------------(8)
which is the classical expression for the kinetic
energy.
(1.3) DEDUCING WORK-ENERGY THEOREM FROM
EQUATION (6).
In relativity, the momentum p of a particle at any
velocity v can be defined in defined as
p=mv
But, m=m0/[1-(v2/c2)]1/2
Where,
velocity v.
p= m0v/[1-(v2/c2)]1/2
now,
force(f) = dp/dt = d(m0v/[1-(v2/c2)]1/2)/dt -----(9)
the work done by force F in a displacement ds of the
body is equal to the change in the Kinetic Energy k of the
body,
dk=Fds
 putting the value of F from equation (9) we
get,
 dk= {d(m0v/[1-(v2/c2)]1/2)/dt}ds
 (c2-v2)dk=[(m0c*c2)/[c2-v2]1/2]*(dv/dt)*ds
 But, (c2-v2)1/2=(m0c)/m
 (c2-v2)dk = [(m0c*c2m)/m0c]*(ds/dt)*dv
 (c2-v2)dk = mc2*vdv
 Integrating both sides we get,
 (c2-v2) ∫dk=mc2 vdv
 (c2-v2)k=mc2[(v2-u2)/2]
 [(c2-v2)k]/c2 = (1/2)m[v2-u2]---------------(10)
 Since, the particle starts from rest so u=0 and
w=(k)*[1-(v2/c2)]
w=(1/2)mv2---------------------------------------------(11)
(2) POSTULATE – 2
The Gravitational Force between two particles at
rest separated by a distance from each other is
equal to the Gravitational Force between that two
particles moving with equal and uniform velocity
keeping the distance same between them.
(Gm0m’0)/r2 = [(Gmm’)/r2] [1-(v2/c2)]
Where
m0=Rest mass of first particle.
m’0=Rest mass of second particle.
m=Moving mass of first particle.
m’=Moving mass of second particle.
r=Distance between the two particles.
G=Gravitational Constant.
v=Velocity of both the particles.
c=Velocity if light in vacuum.
Proof:- According to Newton’s Law Of Gravitation
F=(Gm0m’0)/r2----------------------------------(12)
Now,
m0=m[1-(v2/c2)]1/2
m’0=m’[1-(v2/c2)]1/2
putting the value of m0 and m’0 in equation (12)
we get,
F=[(Gmm’)/r2] {[1-(v2/c2)]1/2}2
F=[(Gmm’)/r2] [1-(v2/c2)]-----------------------(13)
Hence,
(Gm0m’0)/r2 = [(Gmm’)/r2] [1-(v2/c2)] [proved]
Equation (13) can also be said as the formula to find
relativistic Gravitational Forces.
For example:-
Two particles A and B has rest mass of 20 kg and 30
kg separated by a distance 10m. Suddenly, both of
them started to move with a velocity 2*108m/s.
Verify the theory of equability in this case.
m0 of A=20kg m’0 of B=30kg
r=10m
m of A=26.8 kg m’ of B=40.2kg
F=6G F’=5.9G=6G
Hence,
The theory of equability is Verified.
CHAPTER 2
MOTION: MECHANICS AND
THEORY OF SEPARATIONAL
FORCE
According to Newton’s first law of motion, if a
body is in a state of rest, it will remain in the
state of rest and if it is in the state of motion, it
will remain moving in the same direction with
the same speed unless an external force is
applied on it.
The above law deals with inertial reference
frame. But have we ever imagined that a body
which we call at rest does it actually at rest. The
apparent view is yes but if we imagine that the
molecules which is constituted inside the body is
at motion. Then how can a body is said to be at
rest? Suppose in an electric circuit, an observer
says that the circuit is at rest but if there is a flow
of electricity inside then, does the whole thing at
rest?
According to surrounding the body is at rest
but according to itself its outer visible surface is
at rest but its inner constituents molecules or sub-
atomic particles are in motion. Hence, the body is
in a state of partial rest-motion state.
A body is said to be in partial rest motion state
when the body is apparently viewed appears to be
at rest but its constituent molecules appear to be
at motion.
This partial rest-motion occurs in a body till the
body does not absolute zero temperature. As,
absolute zero temperature, the molecules stop
vibrating at their positions and the body is said to
be at total rest state(or normal rest state).
A body is said to be in total rest state when the
apparent view of the body and its constituent
molecules appears to be at rest. This state is
possible when the temperature is absolute zero.
A body undergoes total motion state when the
body along with its constituent molecules is
appeared to be in motion.
Imagine a clock with a second’s hand. If we ask
a observer that whether the second’s hand is
moving or not, the observer will say the second’s
hand is in motion and if he thoughts about my
concept then he will answer that the body
undergoes total motion state. But if we observe it
carefully then we can see the whole body of the
second’s hand is not in motion the part of body
which is fixed to clock that small part is in rest.
Here, the constituent molecules are in motion, the
second’s hand is in motion but the fixed point of
the second’s hand though it is a part of the
moving hand is not moving or is hot changing its
position. This type of motion in which outer
visible part is related is known as Surface rest-
motion state. See the below mentioned figure:-
CONCEPT OF UNIVERSE AS A SINGLE
BODY:-
Now, if we imagine Universe as a body and
planets, asteroids, and other heavenly bodies as
its constituent particles then two cases evolves:-
Case 1:- If universe only expands and it does not
move neither on its axis nor it revolves then it
undergoes a state of partial rest-motion state.
Case 2:- If Universe moves on its axis or it
revolves on its axis then it undergoes a state of
total motion state.
Now, if case 2, is true then another concept
comes into existence that is Universe has its own
Gravitational Pull. That is if the Universe rotates
on its axis then the galaxies and other heavenly
bodies would fly away from the universe. So,
there must be equal and opposite force acting on
the bodies. This force can be centripetal or
gravitational force or both but it’s value is very
small. That’s why does not cause any effect on
the bodies.
GRAVITATIONAL FORCE AND
COLLISION
Let there be two bodies of mass m1 and m2
separated from each other with a distance r.
The gravitational force acting between them is

F=Gm1m2/r2
In the above relation G, m1, m2 are constants.
Hence,
f∝ 1/r 2
This means, the shorter the distance between

the greater the gravitational force will act
between two bodies. Now let the bodies are
pushed towards each other then a point will
reach when they will meet each other.
In this case r=0 and gravitational force is ∞ but

after collision they move away from each.
This is because of a hidden force is present in

between them and it is Separational Force (f’)
whose value is greater than gravitational force.
F’>f(G.F.)
F’>∞
Hence, the value of f’ is
∞+n , where n>0, n belong to real numbers
F’=(∞+n)
If f’=f
∞+n=∞
 n=∞-∞
 n=0
RELATION BETWEEN GRAVITATIONAL
FORCE AND SEPARATIONAL FORCE:-
Gravitational force is the force that attracts
two bodies towards each other. But
separational force is the force that provides
a repulsion between two bodies.
According to Newton’s third law of motion,
to every action there is always an equal and
opposite reaction. Hence, Gravitational
force and Separational force is an action-
reaction pair.
Mathematical relation:-
Let gravitational force be f and separational
force be f’ .
According to Newton’s third law of motion,
F=f’
 f’=f
 f’=-Gm1m2/r2
... The formula for separational force is:-
F=-Gm1m2/r2
Negative sign denotes the force is opposite in
nature.
Qualitative analysis:-
Put two of mass m1 and m2 beside each other at a
distance r from each (where there is no friction
and no air resistance)
Now, due to gravitational force m1 and m2 should
come together but they does move towards each
other due to the Separational force is present in
between them.
Thus,
Separational force=-Gravitational Force
SEPARATIONAL FORCE AND
CONCLUSION :-
When bodies say m1 and m2 come towards
velocity v1 and v2 they collide with each other
and gets separated due to separational force.
But we know separational force and
gravitational force are equal and opposite but
after collision separational force get an extra
supporting which is the molecules exact.
In the above case, when m1 and m2 strikes
against each other they but after collision the
force obtained from the collision of m1 and m2
is passed through the molecules until the last
molecules provides an extra force.
This force is exacted outwards by the

molecules. We know Separational force act on
two bodies m1 and m2. Hence, the extra force
which is added to separational force=f1+f2
As,f1=f2=f
Force=2f.
Now, separational force
f’=Gm1m2/r2 + 2f----------------(i)
Now,
Let find the force exacted by the molecules.
Let the mass of an atom present in that
molecule be m and velocity v.
p=mv
. momentum of molecule
. .
p=nmv
f=dp/dt
putting the value of f in equation (i) we get,
f’=Gm1m2/r2+2[dp/dt]
multiplying both sides by dt
f’dt=Gm1m2/r2dt+2dp
integrating both sides
f’∫dt=Gm1m2r2∫dt+2∫dp
f’t=Gm1m2/r2*t+2p
f’t=Gm1m2/r2*t+2nmv---------------------------------------(ii)
Hence,
Impulse of separational force=Separational
force+2n momentum of 1 atom.
VERIFICATION OF EQUATION (II)
DIMENSIONALLY:-
Dimension of impulse of force= [MLT-1]
Gravitational force of impulse= [MLT-1]
Momentum of molecule= [MLT-1]
After, writing the dimension on the equation
(ii) we get
[MLT-1]= [MLT-1]+ [MLT-1]
According to principal of homogeneity,
[MLT-1]= [MLT-1]
Hence, verified.
APPLICATION OF SEPARATIONAL
FORCE:-
Some few applications of separational force
are:-
 Due to the presence of separational
force the planets does not collide with
each other. Otherwise, as they both
attract each other, they would come
closer and closer and finally they would
collide with each other.
 In General Theory of Relativity, it is
said that the gravitational force is
formed when a heavy mass is put on the
space-time it provides a depression as it
forms on fibre when a heavy mass is
put on it. But the lighter mass that is
moon does not fall towards other it is
because of Separational force is present
between them.
 In our daily life when we put two
bodies beside each other they do not
collide each other due to the presence
of Separational force is present between
them.
 In atoms thought there is force of
attraction between protons and
electrons but the electrons never fall
into nucleus due to the force of
separation.
CHAPTER 3
MASS CHANGES DURING
REFRACTION
REFRACTION:-
Refraction is the bending of light towards or
away from the light when light travels from
one optical medium to another optical
medium of different optical densities.
i=angle of incidence
r=angle of refraction
Moreover, during Refraction velocity and
wavelength of light also changes.
GENERAL & SPECIAL THEORY OF
RELATIVITY AND REFRACTION:-
According to General theory of Relativity,
light has its own gravitational pull. That’s
why it bends towards the Sun. Then a
substance which has its own gravitational
force should have a mass how small it can be.
According to Special theory of Relativity, a
body has two types of mass rest mass and a
mass during its motion. We all know that
mass increases with increase in velocity and
vice-versa. Then, the mass of light will also
change will travelling from one medium to
other of different optical densities.
Mathematically:-
Let the mass of light in medium1 be m1 and in
medium2 be m2.
Velocity of light in medium1 be v1 and
velocity of light in medium2 be v2.
The light passes from medium1 (rarer) to
medium2 (denser). Like the below figue:-
Change in mass = ∇m=m -m

2 1 [since v1>v2]
Change in velocity=∇v=v -v 1 2
Momentum of light in medium1=p1=m1v1

Momentum of light in medium2=p2=m2v2
Since, the mass is very small, so momentum
depends on velocity since v1>v2 .
. Change in momentum
. .
 ∇p=p2-p1=m2v2-m1v1
 Dividing v2 on both sides,
 ∇p/v2=m2v2/v2-m1v1/v2
 ∇p/v2=m2-m1*1n2 [ since, 1n2=v1/v2]
 We know, ∇p=∇(mv)= ∇m*∇v
 (∇m*∇v)/v2-m2=-m1*1n2
 ((m2-m1)(v1-v2)-m2v2))/v2=-m1*1n2
 (m2v1-m2v2-m1v1+m1v2-m2v2)/m1v2=-1n2
 (m2/m1)*1n2-(m2/m1)-1n2+1-(m2/m1)=-1n2
 (m2/m1)*1n2-2*(m2/m1)=-1
 (m2/m1)*(1n2-2)=-1
 (1n2-2)=- (m2/m1)
 -(m2/m1)=1n2-2
 (m2/m1)=2-1n2
 (m1/m2)=1/(2-1n2)
m2=m1/(2-1n2)
CONCLUSION:-
The above formula states that mass of light
changes during refraction. The change of
mass is also a cause for refraction. The
unknown mass of light either is second or first
medium can be found out by this. It also
agrees that mass changes with change in
velocity. Thus, the Special Theory of
Relativity is satisfied.
NOTE:-Mass of photon=1*10-18eV/c2 in
vacuum
CHAPTER 4
WAVE COLLISION AND
LAWS OF COLLISION
(1)Introduction
Earlier, we have only known about matter collision and
real life visible collision and had also learnt how to deal
with them. But in the below topic we will learn about
matter wave collision.
The theory of Electromagnetic mass also provides us

details about mass and energy of Electromagnetic waves
and how they are changing with respect to mediums.
The laws of collision will help us to learn the basics of

collision and it will also help us to solve the problems and
with the help of this law we can arrive at the formulae of
Elastic Collision in 2-dimension which are yet to be
discovered
(2)Collision between two matter waves in one dimension.
Let us consider two particles A and B of masses m 1 and
m2 moving in same direction with velocities u1 and u2
collide each other.
The particle A is called the incident particle particle and

the particle B is called the target particle.
Let v1 and v2 be the final velocities of A and B .
According to De-Broglie’s wavelength,
λ 1=h/m1u1 and λ2=h/m2u2
where λ1 and λ2 are the initial wavelength of particles
A and B. λ1’ and λ2’ be the final wavelength after
collision of the particle A and B.
According to the law of Conservation of momentum,
m1u1+m2u2 = m1v1+m2v2
as the particles are matter waves .
So, we can write that
h/ λ1+h/ λ2 = h/ λ1’+h/ λ2’
hence,
1/ λ1 + 1/ λ2 = 1/ λ1’ + 1/ λ2’ ------------------------(1)
Now, we know that relative velocity of approach is equal
to relative velocity of Separation,
(u1-u2)=v2-v1
h/m1 λ1 – h/m2 λ2 = h/m2 λ2’ – h/m1 λ1’
1/m1 λ1 + 1/m1 λ1’ = 1/m2 λ2’ +1/m2 λ2
1/m1(1/ λ1 + 1/ λ1’) = 1/m2(1/ λ2 + 1/ λ2’)
(m1/m2)* (1/ λ2 + 1/ λ2’)= (1/ λ1 + 1/ λ1’)
(m1/m2)= (1/ λ1 + 1/ λ1’)/ (1/ λ2 + 1/ λ2’)
(m1/m2)=[ (λ1’+ λ1)*( λ2’ λ2)]/[ (λ2’+ λ2)* λ1’ λ1]---(2)
From equation (1) we get
(m1/m2)=[( λ1’+ λ1)* (λ2- λ2’)]/[ (λ1’- λ1)* (λ2’+ λ2)]
Again from equation 1 we solve for λ1’ and we get
( λ1’+ λ1)/ (λ1’- λ1)=[2 λ2 λ2’+ λ1 λ2’- λ1 λ2]/[ (λ2-
λ2’)]--------------------------------------------------(3)
putting the above value in equation (2) , we
obtained the result
λ2’= [λ1 λ2*(m1+m2)]/[2m2 λ2+m2 λ1-m1 λ1]---(4)
Similarly, solving for λ2’ from equation (1) and
putting them in equation (2) we get
λ 1’=[λ1 λ2*(m1+m2)]/[2m1 λ1+m1 λ2-m2 λ2]---(5)
(3)THEORY OF ELECTROMAGNETIC MASS

 The frequency of an electromagnetic Wave is directly
proportional to its moving mass.
f ∝m
where,
f=frequency
m=moving mass
 Moving mass of an Electromagnetic wave is
different in different media.
Proof: The speed of electromagnetic wave in any medium
is
C= (1/uε)1/2 -----------------------------(6)
Where,
C=velocity of electromagnetic wave
u=permeability of the medium
ε=permittivity of the same medium
We know that,
C=f λ where λ=wavelength
Putting the value of c from equation (6) we get
(1/uε)1/2 = f λ
1/ λ=f(uε)1/2
But, λ=h/mc where h = planck’s constant
mc/h = f(uε)1/2
m=hf(uε)1/2/c
putting the value of c from equation 6 we get,
m=hf[(uε)1/2]2
m=hfuε----------------------------------------(7)
since, h, u, ε are constants for a particular medium
hence,
 f∝m[proved]
The equation (7) shows that an Electromagnetic wave
which exerts pressure and have momentum should
have mass.
(3.1) ENERGY OF AN ELECTROMAGNETIC
WAVE
Energy of an electromagnetic wave is different in
different medium
We know,
Momentum(p) = Energy(e)/velocity(c)-----------(8)
C=e/p
Now,
m=hfuε
m= hf(uε)1/2/c
putting the value of c from equation (8) we get
m=hfp(uε)1/2/e
e=hfp(uε)1/2/m-------------------------------------(9)
According to de-Broglie’s wave equation
λ =h/p
p=h/ λ-------------------------------------------------(10)
putting the value of equation (10) in equation (9)
we get
E=(h2f)* (uε)1/2/(m λ)--------------------------------(11)
(4) LAWS OF COLLISION

The following are the basic laws of collision:-
 When a incident particle collides with a targeted
particle, the targeted particle starts moving with a
uniform velocity in the line which joins the centre of
masses of both the particle during collision.(The
mass of targeted particle should be very large in
comparision to the incident particle)
 If the mass and velocity of the inclined particle is

very large than the mass and velocity of the target
particle then, the incident particle will continue to
move in the same direction but with different
velocity.
 If the collision between incident particle and Target
particle is not head-on and both the masses of
incident and target particle are same then the two
identical particles move at right angles after elastic
collision in 2-dimension.
 If the mass of a target particle is very large in
comparison to the incident particle then the particle
bounces back with almost same speed.
(4.1) ELASTIC COLLISION IN TWO DIMENSIONS

Let A and B be two particles which have an elastic
(non-head on) collision with each other, the particle
B being at rest. The particle A is called the incident
particle and B is the target particle.
Let m1 , m2 = masses of the particles A and B
u1 = velocity of the particle A along x-axis before
collision.
v1 , v2 = velocity of the particle A and B after collision.
Since, the collision is non-head . Let v1 and v2 make
angles and respectively along axis.
According to the figure below:-
Let,
 OP=b
 The line joining two centre of masses=(r1+r2)=h
 The line joining centre of mass of the target body to the
initial direction of incident body along the x axis
=(r2+a)=p
From the above figure we get
cos =b/(r1+r2) = b/h---------------------------------------(12)
sin =p/(r1+r2)=p/h----------------------------------------(13)
and
m1v1sin =m2v2sin ---------------------------------------(14)
putting the value of sin from equation (13) we
get,
m1v1sin =m2v2(p/h)
m2v2= m1v1sin (h/p)
According to law of conservation of momentum,
m1u1=m1v1cos +m2v2cos -------------------------(15)
m1u1 = m1v1 cos + m1v1 sin *(h/p)*(b/h)
u1 = v1cos + v1sin (b/p)
v1pcos +v1sin =u1p
v1 = (u1p)/(pcos +bsin )
v1 = (u1[r2+a])/([r2+a]cos +bsin )---------------(16)
now, again from equation (14) we get,
m1v1 = [(m2v2)/sin ]*(p/h)------------------------(17)
putting the value of equation (17) in equation (15)
we get,
m1u1 = m2v2cos /sin *(p/h) + m2v2(b/h)
m1u1h = m2v2cot p+m2v2b
v2(m2cot p+m2b)=m1u1h
v2=( m1u1h)/ (m2cot p+m2b)
v2=( m1u1h)/[m2 (pcot +b)]
v2=( m1u1hsin )/[m2 (pcos +bsin )]
v2=( m1u1(r1+r2)sin )/[m2 ({r2+a}cos +bsin )]----
------------------------ -----------------------(18)
According to conservation of kinetic energy
(1/2)m1u12=(1/2)m1v12 +(1/2)m2v22----------(19)
Now squaring equations (18) and (17) and then
putting it to the equation (19) and calculating we
get
 = cot-1([2m2pb]/[m2p2-m2b2+m1h2])
 =cot-1([2m2(r2+a)b]/[m2(r2+a)2-m2b2+m1(r1+r2)2])
-------------------(20)
(5)CONCLUSION
Equations (4) and (5) are used to find the final
wavelengths of two waves after head- on collision
Equations (7) and (11) are used to find the mass and
energy of electromagnetic waves in different
mediums.
Equations (20), (18) and (16) are used to find the
angles, and the final velocities of the bodies after
elastic collision in two dimensions or non-head-on
collisions.
CHAPTER 5
WAVE BOUNDARY LAYERS
AND FRICTION FACTORS
ABSTRACT
In the last two decades the problem of wave height
damping due to bottom friction has received
increasing attention
among near-shore oceanographers. This fact is
reflected
in the wealth of papers on the subject; the list
of references given herein presents a minor
selection only.
This paper is an attempt to re-evaluate and
systematize
the many observations and the rather few detailed
measurements
of the phenomenon. In nature the wave boundary
layer will always be rough turbulent. This is not
necessarily
the case in a hydraulic model. The aim is therefore
to
make it possible to determine the proper flow
regime for a
pure short-period wave motion over a given bed.
Values for
the wave friction factor and the wave boundary
layer thickness
are also proposed.
The main results of the study are presented in
three
diagrams giving flow regimes, friction factors and
boundary
layer thicknesses. Flow parameters are a-jjj/k and
RE =
U-|j_ ai~/v, a-]m and U-|m being maximum bottom
amplitude and
velocity according to first order potential wave
theory,
k is the Nikuradse roughness parameter.
1. INTRODUCTION
It seems to be generally recognized to-day that the
boundary layers developing at the sea bottom
under gravity
waves for all practical purposes can be regarded
as turbulent,
see for instance [6] and [14]. For many years it has
been extensively discussed, however, whether
turbulence could
appear in laboratory studies of wave phenomena,
see [6] and
the long discussions in La Houille Blanche, [2], [4],
[29]
and [30], The experiments by Miche [29], [30],
Vincent [36],
Lhermitte [23], Zhukovets [37] and Collins [6]
demonstrate,
on the other hand, that turbulent oscillatory
boundary layers
can be generated under laboratory conditions also.
The importance of a sound estimate of the wave
friction
factor for shallow water wave forecasting is
obvious.
In this context reference can be made to the
pioneer works
by Bagnold [1] and Johnson and Putnam [13].
Since measurements
in a prototype scale are scarce, and difficult to
perform,
however, it is imperative to know to what extent
model
127
128 COASTAL ENGINEERING
results in this field are applicable in nature, and
vice
versa. This calls for a detailed analysis of the
behaviour
of the wave boundary layer.
As long as the flow is entirely laminar, the
problem
is open for an analytical treatment. This is not the
case
for turbulent flow. Ho consistent theory dealing
with turbulent
oscillatory boundary layers exists. In this paper an
approach by Lundgren, adjusted according to the
experimental
results of the present author, has been adopted for
the rough
turbulent case, see [14] and [15]. The experimental
results
of Bagnold a. o. will be shown to agree quite well
with the
proposed friction factors. Measurements of
turbulent flow
near a smooth wall seem to be missing entirely, so
an analogy
with rough flow has been introduced.
A preliminary report is given in [17]. The more
complex
problem of bottom friction and energy dissipation
in a
wave motion when superimposed by a current has
been studied
in [18].
NOTATION
D (m)
De (m)
Ew (kgf/m s)
H (m)
L (m)
RE (dim.less)
Ee (dim.less)
T (s)
U (m/s)
U1 (m/s)
uc (m/s)
Uf (m/s)
a1 (m)
c (m/s)
d (m)
fe (dim.less)
fw (dim.less)
h (m)
k (m)
P (m)
2.
Water depth
"Equivalent depth"
Specific energy loss per s
Wave height
Wave length
Amplitude Reynolds number
Reynolds number
Wave period
Wave particle velocity
U at the bottom
Current velocity
Friction velocity
Wave particle amplitude at bottom
Wave celerity
Diameter of cylindrical roughness
(Kalkanis)
Wave energy loss factor
Friction factor for %v
Ripple height (Bagnold)
Nikuradse roughness parameter
Ripple pitch (crest to crest)
(Bagnold)
Eq.No.
(3.1)
(3.3)
(3.2)
(3.6) & (5.2)
(3.10)
(3.10)
(5.2)
(3.9)
(3.4)
(3.10)
t (B)
u (m/s)
w (m/s)
X (m)
z (a)
6 (m)
vise (m)
V (m2/s)
Q (kgf s2/
T (kgf/m2)
TW (kgf/m2)
^0 (°)
U) (1/s)
WAVE BOUNDARY LAYERS 129

Eq.No.
Time
Velocity fluctuation in x-direction
Velocity fluctuation in z-direction
Coordinate in direction of wave travel
Coordinate at right angles to bottom
Wave boundary layer thickness (Pig. 2)
Thickness of viscous sublayer (5.21)
Kinematic viscosity
Density
Instantaneous shear stress for a
pure wave motion
x at bottom (3.4)
Phase shift between T^ and U1m
Angular frequency
log loS-lO
Mean value sign
Suffix m denotes maximum.
3. DEFINITIONS
First order potential wave theory is applied
outside
the boundary layer, see Fig. 1. The wave boundary
layer
thickness 6 is conveniently defined from the
velocity profile
shown in Fig. 2. As the thickness of the boundary
layer
for short-period waves is of the order of
magnitude 1/100
of the water depth, it will not affect the motion of
the
body of water, and U-| in Fig. 2 can be taken equal
to the
theoretical bed velocity for a frictionless fluid.
At z = 28, rm is approximately 0.05 x-wm, where
TW is
the bottom shear stress, and "m" denotes
maximum. 26 can
therefore be said to be analogues to the depth of a
steady
flow in an open channel, and could be denoted
"the equivalent
depth", D„, i.e.
De = 26 (3.1)
This analogy will be made use of later. It will be
shown
to yield remarkably reliable results.
At z = 6, Tm equals 0.21 Twm for laminar motion,
see
(5.8), and was measured to be 0.35 ,vvm
in Test No. 1 in the
oscillating water tunnel (fully developed rough
turbulence,
see [14]). Thus it appears, that the boundary layer
thickness
here defined is only similar to the boundary layer
thickness employed in steady flow, in the sense
that it gives
a measure of the thickness of the layer adjacent to
the wall
over which the velocities deviate significantly from
the

free-stream velocity. If the boundary layer is
thought of
as that part of the flow, where shear stresses play a
r61e,
(3.1) gives a more consistent measure.
Two Reynolds numbers are introduced, one with
the
boundary layer thickness, the other with the
maximum amplitude
a-|m (half stroke length) in the free stream as
length
scale, i.e.
TJ. 6
(3.2)
(3.3)
Re
P1mV
RE im im
V
ction factor fw„ :
Twm = fw i Q U1m (3.4)
although T and TJ.. are not simultaneous.
lection it can be shown,
l factor f in the equatic
Tw = f^Q|U1)2 (3.5)
In this connection that if we assume
a constant friction equation
with U1 given by
U1 = U1m sin a»t (3.6)
and the specific energy loss Eyj per s simply by
Ew - *w *1 (3-7>
then the mean specific energy loss per s is
\ - h 1 f U1m (3'8)
It is often implied, that f in (3.8) is identical with
fw. This is obviously not true, for the following
reasons.
Firstly a phase shift should be introduced in (3.5)»
and
secondly the constancy of f during a wave cycle
can be questioned.
Finally, (3.7) is only a good guess. So it can be
stated, that as a matter of principle, fw cannot be
determined
correctly by a wave attenuation test. The side-wall
and surface
corrections, which are difficult to control, and the
reflection, are other sources of error.
f in (3.8) will be denoted fe, so that we obtain the
following equation of definition for the "wave
energy loss
factor":
\ = If « fe U1m <3-9)
It should be mentioned here though, that while it
will be shown, that fw f fe in the laminar case, it
was
found in Test No. 1 (see [H]) for a rough turbulent
boundWAVE
BOUNDARY LAYERS 131

ary layer, that fw was practically equal to fe. For
this
reason no distinction will he made in the turbulent
ease hetween
fw and fe fox the very few neasurements available.
Introducing the right phase shift (~25°) in (5.5) it
was also
found, that f was practically constant, using the
bottom
shear stresses determined from velocity profiles.
The roughness parameter (k) introduced for the
(fixed)
bed is the Nikuradse sand roughness, as defined
from the expression
for the turbulent velooity profile near a rough
bottom: -g
TjS- 5.75 log 2|-2 (3.10)
4. METHODS OP MEASURING THE WAVE
FRICTION FACTOR
The wave friction factor can be found in a variety
of
ways. The "classical" procedure is to measure the
wave
height attenuation in a flume. In the preceding
chapter
certain disadvantages of this method were
outlined. Hence,
a short descriptive review of existing methods
might be of
interest here.
In general one can distinguish between three main
principles: Measurement of energy loss, force or
velocity.
These can again be subdivided as shown below.
Quantitative
information will not be given. This can be found in
chapters
5 and 6 and in the references cited.
MEASUREMENT OF ENERGY LOSS

The quantity measured hereby is really the wave
energy
loss factor, see (3.9) and the appurtenant
discussion.
Direct measurement - Bagnold [1] used a
technique
which was simple and ingenious. A celluloid plate,
to which
fixed imitation ripples were attached, was hung
vertically
in a large tank of water. The plate was oscillated
by a
mechanism driven by a weight in a wire. The
energy dissipation
was simply found from the falling velocity of the
weight,
corrected for mechanical friction.
Measurement of wave height attenuation - This
method
is based upon the principle, that the reduction in
wave power
between two stations equals the energy loss per s
over the
same distance. The procedure has been adopted by
Miche [50],
Imman and Bowen [9], Iwagaki et al. [10], [12],
Zhukovets
[37], and many others.
In this context it must be mentioned that wave
height
attenuation in the presence of a laminar boundary
layer always
seems to exceed the theoretical value. Much
discussion
has been devoted to this problem. In the author's
opinion one
or more of the following three phenomena are
mainly respon132
COASTAL ENGINEERING
sible: The side-wall correction for the zone around
MWL is
underestimated by standard methods. The flow
regime is not
fully laminar (see Fig. 3). Due to (invisible)
contamination,
a boundary layer is present at the surface, see van
Dorn [35].
MEASUREMENT OP FORGE
Direct measurement - Eagleson [7], and Iwagaki et
al.
[12] have measured directly the force exerted on a
smooth
plate by progressive shallow water waves.
Measurement of the slope of mean water level -
Using
the concept of the wave thrust, introduced by
Lundgren [28],
it was shown in [8] and [18] how the wave energy
loss factor
can be found by measuring the slope of the mean
water level.
The method is based upon elimination of dH/dx
from the energy
equation mentioned above, and the equilibrium
condition,
stating that the reduction in wave thrust between
two stations
equals the difference in pressure force from the
rise
of the mean water level over the same distance.
(The wave
thrust is identical to the "radiation stress"
obtained independently
of lundgren by Longuet-Higgins and Stewart, [25]
and [26]).
MEASUREMENT OF VELOCITY
The velocity field can be measured either over an
oscillating plate, Kalkanis [20] and [21], or in an
oscillating
fluid, Jonsson [14].
Equation of motion - A knowledge of the complete
velocity
field makes possible a determination of the bed
shear
stress through integration of the equation of
motion, see
[H].
The law of the wall - Very near the wall, the
turbulent
velocity, relative to the wall, will be logarithmic.
Thus, the friction velocity and from that the
friction factor
can be calculated, see [14].
Velocity measurement at a fixed level - This
method
is analogous to the Preston tube technique, see
Jonsson [16].
If the drag cofficient corresponding to a fixed level
near
the bed is found in a steady flow experiment, the
maximum
shear stress can be found directly from measuring
the maximum
velocity at this fixed level.
5. CHARACTERISTICS OF THE WAVE
BOUNDARY LAYER
DIMENSIONAL CONSIDERATIONS
For a given "form" of the outer (potential) velocity
(here sinusoidal), dimonsional analysis yields
directly the
following relationships for the wave boundary
layer thickness
and the wave friction factor:
6/a1m fw
laminar case f(U1ma1n/v) f(ulma1m/v)
Rough turbulent case f(a-|m/k) f(a-jjj/k)
Smooth turbulent case f (Uima-|„/v) f (U1ma-|jj/v)
"f" denoting "function of". Quantitative
expressions will
be given in the following,
EQUATION OF MOTION
The linearized equation of motion in the boundary
layer reads (for a fixed bed)
with U-j given by (3.6), u and w being the velocity
fluctuations
in the x- and z-directions, respectively. U-|m is
given by ^ &
U1m = ¥ siph
1 &TD (= -TJ5) (5'2)
A solution to (5.1) in the case of turbulent flow has
not been found yet,* It is known, however, that a
logarithmic
velocity distribution is found in the vicinity of the
boundary, see [14], The shear stress gradient at
the boundary
is found from (5.1):
5T SU1
** z=o = " Q **" (5'3)
It is interesting to note, that this gradient is
determined
exclusively by the outer (potential) flow.
LAMINAR CASE
The solution to (5.1) with u~w = 0 reads ([ 22] p.
622):
U = U1m [sin out - exp (- \ f) sin (ait - \ §)] (5.4)
with
6 =yj.y7~f (5.5)
From (3.3) and (5.5) we find
6 = TT
a1m J2~W
where the important "amplitude Reynolds
number" RE formally
makes its first appearance. It can be interpreted as
a measure
of the square of the ratio between amplitude and
theoretical
laminar boundary layer thickness. (Note that the
relationship
6 Re ,R „•>
im
»} See note after refs.
(5.6)

is always valid, see (3.2) and (3.3)). (5.6) gives a
straight
line in Pig. 5.
The shear stress distribution is
Q = 7?^exp (~ * f} cos <•* - H - i> <5-8>
i.e. at the bottom
wm _ TT im /j- 0\ T 7!"T (5'9)
Prom (3.3), (3.4), (5.6) and (5.9) the wave friction
factor
is found
shown in Pig. 6.
On the other hand it oan be shown, that
so from (3.9)
f 3 A/?TT 1 1.67 /r 12%
i.e. different from fw. The variation of f is shown
in
Pig. 6. W e
The "small" Reynolds number is here found to be:
Re = 75-/SI (5.13)
Direct measurements of the shear stress exerted on
a
smooth horizontal bottom by progressive, shallow
water waves
were made by Iwagaki et al. [12]; the results agree
well
with (5.10). This also applies to the energy
dissipation
measurements by Lukasik and Groseh [27], The
rather high
values found by Eagleson [7] are presumably due
to some instrumentation
error. Iwagaki also found that the wave
attenuation
coefficients (= - (dH/dx)/(H/L)) were about 1,4
times
the values as predicted by theory. The deficiency
may be
due to the development of a boundary layer at the
free surface.
(This effect has been studied both experimentally
and
theoretically by van Dorn [35]. Good agreement
was found
between theory and measurement. Although the
present author
does not agree entirely with the analytical
treatment given
in the above mentioned reference, there can be
little doubt
of the importance of the phenomenon). Capillary
effects at
the side walls may play a r61e, also.
According to [23] the roughness can be "felt" for
k/8 > 0.25, so the "start" of the laminar - rough
turbulent
regime is given by
^-^.ya (5.14)
corresponding to line "LR" in Fig. 4. (Note that in
the
classical experiments by Niknrad.se [32] a direct
transition
from laminar to rough turbulent flow was also
found, for the
larger ratios between roughness and pipe radius.
Prom [34]
p. 483 it appears, that the limiting ratio was close
to r/k =
15, r being pipe radius, r is twice the "hydraulic
radius"
i.e. corresponds to 2 De or 4 6 according to (3.1).
The
criterion r/k = 15 is therefore transformed to k/5 =
4/15
which comes very close to the value obtained by
Lhermitte).
In the open channel experiments of Eeinius [33],
the
transition_between laminar and turbulent flow
was found to
occur for TJ0D/v between 500 and 1000, with the
most abrupt
change of f for TJcD/v about 575. It is therefore
proposed -
using (3.1) and (5.2) - that smooth turbulence
starts for
Re = 250, corresponding to
RE = 1.26 • 104 (5.15)
using (5.13) (line "LS" in Fig. 4), This guess was
confirmed
surprisingly well by Collins [6], who found the
value
RE = 1,28* 104 from measurements of mass
transport velocities
at the edge of the boundary layer. (Li [24] found
1.60* 105, and Tincent [36] 6.2 • 10?. These limits
are without
any doubt very subjective because of the method
employed
(visual observation of the stability of dye streaks)).
ROUGH TURBULENT CASE

The only velocity measurements known to the
author
are those reported in [14], [20] and [21], However,
the
measurements of Kalkanis [20], [21] are made too
far from
the (oscillating) wall to allow a determination of
the friction
factor.
In [14] and [15] the following expressions for the
boundary layer thickness and the wave friction
factor were
found
(30 |). log (30 |) = 1.2 -jiS (5.16)
vr+ loe vr=' °'oa + los "^ (5,17)
corresponding to the horizontal lines in Figs, 5 and
6.
(log is log10).
(5.16) can also be written
l1m a1m
a log (^ j^JS) = 0.04 (5.18)

Typical values of 6/a-jm and fw are given in the
table
below.
a1n
k
1 2 5 10 20 50 100 200 500 1000 2000
^-.102
a1a
9.15 6.65 4.72 3.80 3.14 2.54 2,20 1,94 1.67
1.51 1.38
2
f • 10
u
47.8 23.8 11.2 7,00 4.65 2.93 2.19 1.67 1.22
0.985 0.810
It should be mentioned, that (5*16) and (5.17) are
based upon a very simple theory, assuming that
the logarithmic
velocity profile in the turbulent wave boundary
layer
extends uninterrupted to the potential velocity.
This gives
the general "shape" of the formulae, see [27].
Furthermore
the numerical coefficients or terms were
determined from one
test only in the oscillating water tunnel [14].
Consequently,
checks on the validity of the two above expressions
are
naturally called for.
Prom the measurements of Kalkanis [21] values of
6
can be deduced. For the two-dimensional
roughnesses (10 <
a-|_/k < 64, see Fig. 4), values were found being
approximately
20<fo smaller than obtained by (5.18). Considering
the
many sources of error in the measurements this
discrepancy
can be accepted.
More information can be found from the literature
on
fw. Bagnold's measurements have been re-
analysed, and it
was found that his friction factor "k" equals fe/3 ~
fw/3.
The pitch/height ratio p/h of the ripples was 6.7/1
ana the
ripple trough sections consisted of circular arcs
meeting to
form sharp crests at an angle of 114°. Similar
ripples have
been analysed by Motzfeld [31], who found k = 4 h.
The results
are plotted in Fig. 3. (In the three test series not
shown, 2 a-j-j is smaller than p, and they are
therefore without
interest). It appears from Fig. 4 that the tests are
all in the fully developed rough turbulent regime.
Eliasson et al. have measured the slope of the
mean
water level due to the reduction in wave thrust, see
chapter
4 and [8]. Two test series are shown in Fig. 3. After
the
completion of the tests with k = 2.3 cm, the
measuring system
was highly improved, and the last series (with k =
1 cm)
shows reasonable agreement with the theoretical
curve. The
measurements are very near the limit "EL", see
Fig. 4. (Because
of the very low a-|jj/k ratio, no side-wall correction
was introduced. It will be of the order of
magnitude of
20/o), A wave flume measurement of wave height
attenuation
(corrected for side-wall effects) by Inman and
Bowen [9]
(Test No. 1 A) is also shown.

It can toe concluded that the number of reliable
measurements
is very scarce. In the author's opinion most
importance
should probably be attached to Bagnold's
experiments
and to the measurements in the oscillating water
tunnel.
The interpretation of the former is a little difficult,
however. There is information, which indicates,
that Motzfeld
has overestimated k a little. If k was 3 (instead of
4) times h, then Bagnold's results would in fact
coincide
with the curve in Pig. 3.
So the expressions for 6 and fw given by (5.16) and
(5.17) are preserved for the present. It could be
mentioned
here, that the tendency in Pig. 3 - fw decreasing
with increasing
amplitude - was also found by Iwagaki and
Kakinuma
[11] from analysis of prototype observations.
In the preceding discussion we have anticipated
the
existence of limits for the turbulent regime. These
are
found as follows.
It is assumed that complete turbulence is
developed
for U1mDe/v = 1000. So we find from (3.1), (3.2)
and (3.3)
Ee = 500 (5.19)
or a1 1
EE = 500 --jr'jjx (5.20)
corresponding to the lines "El" in Pigs. 4, 5 and 6.
Zhukovets'
observation [37], that "the quadratic region exists
for Eeynolds numbers from 1.5* 104 to 3.3' 104"
agrees well
with these values. The transition curves in Pigs. 5
and 6
are estimated. It is improbable that they will not
be
"smooth", since the main motion is unsteady.
Colorimetric investigations by Miche [29], [30] are
plotted in Pig, 4. In all 7 tests turbulence was
(visually)
present.
The limit between the rough turbulent and the
smooth
turbulent- rough turbulent transition regime is
determined
by the ratio between roughness and thickness of
the viscous
sublayer. This quantity is here defined by
s _ 11»6 v _ 11.6 v _ 18.2 v ,R 91s
vVi1sSeC —ff 2 W (5.21) Uf £. n u
x fm TT ±m
supposing TW to vary as sin2(a>t + cp0).
Assuming
T-^— = 3 (5.22)
vise
by analogy with steady flow conditions, (5.22) can
also be
written as

RE = 77.2 -jL 4S (5,23)
This corresponds to lines "RS" in Pigs. 4, 5 and 6.
It is worth while to look a little closer at the
friction factor curve in Pig. 3. Firstly it is seen, that
in the neighbourhood of a^j/lc equal to one we find
fw ~
(aijj/k)"1, which can be shown to yield an
exponential wave
height variation. It is sometimes stated, that this
variation
is a "sign of laminar damping"; it is interesting to
note, that this well may be a false interpretation.
Secondly it can be shown, that the relationship
between
fw and 6/k, as given by (5.16) and (5.17), is very
closely fitted by
f* _ 0.0604 /c 9yn w B . 2 22 & (5*24)
log —<g—
which is identical to the friction factor in a steady,
uniform
flow, if D is put equal to 2 5 (ef. (3.1)).
From certain compatibility relations ([19]) the
phase
shift qp0 (between T^• and U-jm) is found to
decrease with
increasing a^jj/k. The following values are
proposed!
a1m/k = 100 * qp0 = 29°, aij/k • 1000 * cp0 = 110.
SMOOTH TURBULENT CASE

Apparantly only Kalkanis [20], [21] has measured
velocities
at a smooth wall. The measurements do not allow
a determination of fw, however. And the boundary
layer
thicknesses, which can be deduced from the
measurements are
strangely enough equal to or smaller than the
laminar thicknesses
corresponding to the same values of RE. We are
therefore
compelled to make a reasonable guess, which will
be to
use the formulae for the rough turbulent case, with
a formal
roughness parameter, defined by
k = 031?; = °*287 6visc (5.25)
(This relation corresponds to a von Karman
number $ equal to
one).
Using (5.25) together with (5.16) and (5.17) we
obtain
6 _ 0.0465 (5.26)
a1m 1VSBT
and
+ 2 log —L- = log RE - 1.55 (5.27)
wfw
shown in Figs. 5 and 6, ((5.26) is a very close
approximation
to a complicated expression). Some typical values
are listed
below.

RE
4
3 -10 105 105 107
^•102 1.70 1.45 1.13 0.92
2
f • 10 1.26 0,916 0.5W 0.354
If we again use (5.19) as a criterion for fully
developed
turbulenoe, (5.26) yields
HE = 3.00 • 104 (5.28)
corresponding to line "SL" in Pig. 4. Pigs. 5 and 6
suggest
a transition from the laminar regime which is
much more
"gentle" than in steady flow, as would be expected.
The limit between the smooth turbulent and the
smooth
turbulent - rough turbulent transition regime is
supposed to
be determined from
T-£— = 0.287 (5.29)
vise
which seems to be verified by Lhermitte [23], This
condition
is identical with (5.25), so line "SR" in Pig. 4 can
be
found simply from the intersection of the smooth
turbulent
curve in Pig. 5 (and 6) with the horizontal lines
from the
rough regime.
The values of the friction factor in the region
between
the smooth turbulent and the rough turbulent
regimes
may be affected by the type of roughness. Note the
difference
in steady flow between uniform roughness
elements
(Nikuradse [32]) and non-uniform roughness
elements (Colebrook
[5]). The real transition lines can presumably only
be determined by means of experiment.
A good approximation to (5.27) is
w 0.09 "HE,-0.2
(5.30)
The exponent is seen to be the same as in the
expression for
the friction factor for a smooth plate boundary
layer (with
zero pressure gradient), see [34] p. 500.
6. NUMERICAL EXAMPLES
COMPARISON BETWEEN PROTOTYPE AND
MODEL
Prototype - D = 12 m, H = 2.3 m, T = 8 s, k = 0.1 m
=> L = 76 m, a^ = 1.0 m, a1?/k =10, RE = 7.9* 105.
Pig.
4 shows that we are well within the rough
turbulent regime,
and Pig. 6 yields fw = 7.0 • 10~d.

Model (scale 1:100, Froude) - D = 12 cm, H = 2.3
cm,
T = 0.8 s, k - 0.1 cm » L = 76 cm, a-|m = 1.0 cm, a-
im/k =
10, EE = 7.9 * 10 . Fig. 4 shows that we are in the
laminar
- rough turbulent transition regime, and Fig, 6
yields fw »
10 • 10-2. (Incidentally, had the model bottom been
perfectly
smooth, the flow would have been laminar, and the
same friction
factor as in nature would have been obtained).
The example shows, that by Froude scaling the
shear
stresses may easily become 40$ too high in the
model.
SMOOTHNESS OF A M0DEI BED
A necessary condition for the bed to act
hydraulica.1-
ly smooth (in the turbulent regime) is that HE >
3.00* 104,
see Fig. 4. Even for the rather high value H/D =0.3,
it is
found that waves of steepness 1.5$» 3$ and 6$
require depths
larger than 27 cm, 47 cm and 102 cm, respectively,
to reach
this limit.
The bottom amplitudes corresponding to the
above wave
data are approximately 10 cm. Since from Fig. 4,
a-|jj/k must
at least exceed 475 for the bed to be smooth, it is
required,
that k should be smaller than 0.2 mm. This
corresponds to
the hydraulic roughness of a smooth plaster finish.
So it
will be realized, that pure smooth turbulent flow
will hardly
ever be met.
7. CONCLUSIONS
(a) For a simple harmonic motion over a fixed bed,
the
proper flow regime can be found from Fig. 4, when
the ratio
between maximum amplitude and bottom
roughness (a-jn/k) and
the Reynolds number (as given by (3.3)) are
known. Figs. 5
and 6 then yield the boundary layer thickness
(defined by
Fig. 2), and the friction factor (defined by (3.4)).
(b) In the laminar case experimental results of
Iwagaki
et al. [12] for the wave friction factor agree well
with linear
theory. In the rough turbulent case, the proposed
friction
factors are probably not wrong by more than 20$,
see
Fig. 3, In the case of smooth turbulent flow no
measurements
are available, so future revisions in the diagrams
are not
excluded. More measurements of the wave friction
factor are
earnestly needed, especially for the laminar- rough
turbulent
transition regime.
(c) In the laboratory the laminar - rough turbulent
transition
regime will often be found. Pure smooth flow is
exceptional.
(d) In nature the boundary layer is always rough
turbulent.
The friction factor here will often exceed the value
of 2 • 10-2 adopted by Bretschneider [3]. This is
also confirmed
by observations of Iwagaki and Kakinuma [11].
(e) It may be difficult to attain the same friction
factor
in a wave model study as in nature.
8. REFERENCES
[1] Bagnold, R. A. (1946). Motion of Waves in
Shallow Water.
Interaction between Waves and Sand Bottoms.
Proc.Royal
Soe. (A), J87, 1-15.
[2] Biesel, F. and Carry, C. (1956). A propos de
l»amortissement
des houles dans le domaine de 1'eau peu profonde.
La Houille Blanche, JM, 843-853.
[3] Bretschneider, C. L, (1965). Generation of
Waves by
Wind. State of the Art. Hat. Eng. So. Co.,
Wash.,D.C.
[4] Carry, C. (1956). Calcul de 1'amortissement
d'une houle
dans un liquide visqueux en profondeur finie.
la Houille Blanche, JH, 75-79.
[5] Colebrook, C. F. (1939). Turbulent Flow in
Pipes, with
Particular Reference to the Transitive Region
between
the Smooth and Rough Pipe Laws. J. Inst. Civil
Engrs.,
133-156.
[6] Collins, J, I. (1963). Inception of Turbulence at
the
Bed under Periodic Gravity Waves. J. Geophys.
Res.,
.68, 6007-6014.
[7] Eagleson, P. S. (1962). Laminar Damping of
Oscillatory
Waves. Am. Soc. Civ. Engrs., 88, HY 3, 155-181.
[8] Eliasson, J., Jonsson, I. G., and Skougaard, C.
(1964).
A Hew Way of Measuring the Wave Friction
Factor in a
Wave Flume. Basic Res. - Prog. Rep. No. 5, 5-8.
Coastal Engrg. Lab., Tech. Univ. of Denmark.
[9] Inman, D. L. and Bowen, A. J. (1962). Flume
Experiments
on Sand Transport by Waves and Currents. Proc.
8th
Conf. Coastal Engrg., 137-150, Mexico City.
[10] Iwagaki, Y. and Kakinuma, T. (1963). On the
Bottom
Friction Factor of the Akita Coast, Coastal Engrg.
in
Japan, 6, 83-91.
[11] Iwagaki, Y. and Kakinuma, T. (1965). Some
Examples of
the Transformation of Ocean Wave Spectra in
Shallow
Water. Bull. Disaster Prevention Res. Inst., Kyoto
Univ., 14, 43-44.
[12] Iwagaki, Y., Tsuchiya, Y., and Sakai, M.
(1965). Basic
Studies on the Wave Damping due to Bottom
Friction (2).
On the Measurement of Bottom Shearing Stress.
Bull,
Disaster Prevention Res. Inst., Kyoto Univ., 1_4_,
45-46.
[13] Johnson, J. ¥. and Putnam, J. A. (1949). The
Dissipation
of Wave Energy by Bottom Friction. Trans. Am.
Geoph. Union, 30, 67-74.
[14] Jonsson, I. G. (1963). Measurements in the
Turbulent
Wave Boundary Layer. Intern. Assoc. Hydr. Res.,
Proc.
10th Congress, 1, 85-92, London.
[15] Jonsson, I. G. and Lundgren, H. (1965).
Derivation of
Formulae for Phenomena in the Turbulent Wave
Boundary
Layer. Basic Res. - Prog. Rep. No. 9, 8-14. Coastal
Engrg. Lab,, Tech. Univ. of Denmark.
[16] Jonsson, I. G. (1965). Determination of the
Maximum Bed
Shear Stress in Oscillatory, Turbulent Flow, Basic
Res.
- Prog. Rep. No. 9, 14-20. Coastal Engrg. Lab.,
Tech.
Univ. of Denmark.
[17] Jonsson, I. G. (1965). Friction Factor
Diagrams for
Oscillatory Boundary Layers. Basic Res. - Prog,
Rep.
No. 10, 10-21. Coastal Engrg. Lab., Tech. Univ. of
Denmark.
[18] Jonsson, I. G. (1966). The Friction Factor for
a Current
Superimposed by Waves. Basic Res. - Prog. Rep.
No.
11, 2-12. Coastal Engrg. Lab., Tech. Univ. of
Denmark.
[19] Jonsson, I. G. (1966). On the Existence of
Universal
Velocity Distributions in an Oscillatory, Turbulent
Boundary Layer. Basic Res. - Prog. Rep, No, 12, 2-
10.
Coastal Engrg. Lab., Tech. Univ. of Denmark.
[20] Kalkanis, G. (1957). Turbulent Flow near an
Oscillating
Wall. Univ. of California, Inst. of Eng. Res., Ser.
No, 72, Issue No. 3.
[21] Kalkanis, G. (1964). Transportation of Bed
Material due
to Wave Action, U, S, Army, Coastal Engrg. Res.
Center,
Tech. Memo No. 2,
[22] Lamb, H. (1963). Hydrodynamics. Cambridge
at the
University Press.
[23] Lhermitte, P. (1958). Contribution a I'e'tude
de la
couche limite des houles monochromatiques. La
Houille
Blanche, T3_, A, 366-376.
[24] Li, H. (1954). Stability of Oscillatory Laminar
Flow
along a Wall. U. S. Army, Beach Erosion Board,
Tech.
Mem. No. 47.
[25] Longuet-Higgins, M. S. and Stewart, R. W.
(1960).
Changes in the Form of Short Gravity Waves on
Long
Waves and Tidal Currents. J. Fluid Mech., 8, 565-
583.
[26] Longuet-Higgins, M. S. and Stewart, R. W.
(1962).
Radiation Stress and Mass Transport in Gravity
Waves,
with Application to "Surf Beats". J. Fluid Mech.,
13,
481-504. ~~
[27] Lukasik, S. J. and Grosch, C. E. (1963).
Laminar Damping
of Oscillatory Waves, Am. Soc. Civ. Engrs., 89,
HY 1, 231-239. —
[28] lundgren, H. (1963). Wave Thrust and Wave
Energy Level.
Intern. Assoc. Hydr. Res., Proc. 10th Congress,
J_»
147-151, London.
[29] Miche, R. (1956). Amortissement des houles
dans le
domaine de 1'eau peu profonde. La Houille
Blanche, 11,
726-745. ~~
[30] Miche, R. (1958). Sur quelques re"sultats d1
amortissement
de houles de laboratoire et leur interpretation.
La Houille Blanche, r3, 40-74.
[31] Motzfeld, H. (1937). Die turbulente StrBmung
an welligen
Wanden. ZAMM, 17.. 193-212.
[32] Uikuradse, J. (1933). StrSmungsgesetze in
rauhen Rohren.
Forschungs-Arb. Ing.-Wesen, Heft 361.
[33] Reinius, E. (1961). Steady Uniform Plow in
Open Channels,
Div. of Hydraulics, Royal Inst. of Tech.,
Stockholm, Bull. No. 60.
[34] Sehlichting, H. (1958). Grenzschicht - Theorie.
G. Braun, Karlsruhe.
[35] Van Dora, W. G. (1966). Boundary
Dissipation of Oscillatory
Waves, J. Fluid Mech., 24, 769-779.
[36] Vincent, G. E. (1957). Contribution to the
Study of
Sediment Transport on a Horizontal Bed due to
Wave
Action. Proc. 6th Conf. Coast. Eng., 326-355,
Florida.
[37] Zhukovets, A. M. (1963). The Influence of
Bottom Roughness
on Wave Motion in a Shallow Body of Water. Bull.
(Izv.) Acad. Sci. USSR, Geophys. Ser., Ho. 10, 943-
948.
Transl. from Geophys. Ser,, Wo, 10, 1561-1570,
Added in proof: In a private communication ("On
the Bottom
Ji'rictional Stress in a Turbulent Oscillatory
Flow" March 8,
1965, received Aug. 1966), Professor Kinjiro
Kajiura, Earthq..
Res. Inst., Univ. of Tokyo, has made an important
contribution
to the analytical treatment of oscillatory turbulent
boundary layers. Due to the time limit for the
completionof this paper, the author unfortunately
was unable to incorporate
ADDITIONAL
TOPICS
CHAPTER 6
ON THE STAR PUZZLE
ABSTRACT
In the star puzzle, there are four pegs, the usual three pegs, S, P
and D, and a fourth one at 0.
Starting with a tower of n discs on the peg P, the objective is to
transfer it to the peg D, in minimum
number of moves, under the conditions of the classical Tower of
Hanoi problem and the additional
condition that all disc movements are either to or from the fourth
peg. Denoting by MS(n) the
minimum number of moves required to solve this variant, MS(n)
satisfies the recurrence relation
( ) 1 1{2 ( ) 3 1}, 2; (0) 0
min
MS n k n MS n k n MS k . This paper studies
rigorously
and extensively the above recurrence relation, and gives a
solution of it.
Keywords : Star puzzle, Three-in-a-row puzzle, Recurrence
relation
1. Introduction
The star puzzle, due to Stockmeyer [6], is as follows : There are
four pegs, the usual three pegs S,
P and D, and a fourth one at 0. Initially, the n discs, d1, d2, …,
dn, rest on the source peg, S, in a
tower (with the smallest disc at the top, the second smallest
above it, and so on, with largest disc at
the bottom), as shown in the figure below. The objective is to
transfer the tower from S to the
destination peg, D, in minimum number of moves, (using the
auxiliary peg P), under the
conditions that (1) each move can shift only the topmost disc
from one peg to another, (2) no disc
can be placed on top of a smaller one, and (3) each disc
movement is either to or from 0.
Let MS(n) be the minimum number of moves required to solve
the star puzzle with n ( 1) discs.
To find the recurrence relation satisfied by MS(n), we consider
the following scheme to transfer
the tower of n discs from the source peg, S, to the destination
peg, D.
Step 1 : move the topmost tower of n – k discs from the peg S to
the peg P, using the available four
pegs, in (minimum) MS(n – k) moves.
Step 2 : shift the k discs (resting on the peg S) to the peg D.
Clearly, only three pegs are available.
In this step, the three pegs S, 0 and D (in this order) may be
treated as being arranged in a row,
forming the three-in-a-row puzzle with k discs. The minimum
number of moves involved is 3k – 1.
Step 3 : transfer the tower of n – k discs from the peg P to the
peg D, completing the tower on the
peg D, in (minimum) MS(n – k) moves.
Thus, the total number of moves required is FS(n, k) 2 MS(n –
k) + 3k – 1, where k (1 ≤ k ≤ n – 1) is
such that FS(n, k) is minimized. Therefore, MS(n) satisfies the
following dynamic programming
equation
MS n k n MS n k n MS k (1.1)
with
MS(0) = 0, MS(1) = 2, (1.2)
with the convention that MS(1) is attained at the point k = 1.
Step 2 of the scheme involves a three-in-a-row puzzle, and we
refer to Scorer, Grundy and Smith
[5] and Majumdar [1] for details on the three-in-a-row puzzle.
Since 3k – 1 is even for all k 1, we see that the term inside the
curly brackets on the right-hand
side of (1.1) is even. Thus, MS(n) is even for all n 1. Let
MSM(n) =
MS(n)
2
(1.3)
Then, we have the following result.
Lemma 1.1 : MS(n) is attained at k = K if and only if MSM(n) is
attained at k = K.
Proof : is evident from the defining equation (1.3).
It may be noted here that, the recurrence relation satisfied by
MSM(n) is
with
MSM(0) = 0, MSM(1) = 1. (1.5)
Let
an = MSM(n) – MSM(n –1) for all n 1. (1.6)
Let the sequence 
n n1 b be defined as follows :
bn = 2i 3m, i ≥ 0, m ≥ 0, (1.7)
arranged in increasing order. In this paper, we give a rigorous
proof of the result that
an = bn for all n 1,
by showing that an and bn satisfy the same recurrence relation.
This is done in Section 3. In Section
2, we give some preliminary results related to MSM(n).
2. Some Preliminary Results
Some local-value relationships satisfied by MS(n) have been
derived in Majumdar [3]. By Lemma
1.1, they may be adapted to MSM(n), and are given below.
Lemma 2.1 : For any integer n ≥ 1,
(a) 0 < MSM(n + 1) – MSM(n) < MSM(n + 2) – MSM(n + 1)
≤ 2{MSM(n + 1) – MSM(n)},
(b) MSM(n) is attained at a unique value of k.
Corollary 2.1 : For any integers m 1, n 1,
MSM(n + 1) – MSM(n) = MSM(m + 1) – MSM(m) (1)
if and only if m = n.
Proof : The proof of the “if” part is trivial. To prove the “only if”
part, let the equality (1) hold true
for some integers m and n. Now, if m n, then either m > n or n
> m. In either case, part (a) of
Lemma 2.1 is violated.
Lemma 2.2 : Let MSM(n) be attained at the point k = k1 and
MSM(n + 1) be attained at the point k
= k2. Then, k1 ≤ k2 ≤ k1 + 1.
Lemma 2.2 above states that, if MSM(n) is attained at the point k
= K, then MSM(n + 1) is attained
either at
k = K or at k = K + 1.
Corollary 2.2 : Let MSM(n) be attained at k = K and MSM(n +
1) be attained at k = K + 1. Then,
MSM(n + 2) is attained at k = K + 1.
Lemma 2.3 : Let MSM(n) – MSM(n – 1) be of the form 3m for
some integers n 1, m ≥ 0. Let MS(n)
be attained at k = K. Then, MS(n – 1) is attained at k = K – 1, and
MSM(n + 1) is attained at k = K.
Lemma 2.4 : Let, for some N ≥ 1, MSM(N – 1) be attained at k = K and MSM(N) be attained at k = K
+ 1, so that
MSM(N) – MSM(N – 1) = 3K. (2.1)
Then, there is an integer M ≥ 1 such that
MSM(N + M + 1) – MSM(N + M) = 3K + 1, (2.2)
with M = 1 if and only if N = 1 (so that K = 0).
Moreover, M > K.
Proof : The first part of the lemma has been established in

Majumdar [3]. It then remains to show
that M > K.
Now, for any integer L with N < L < M, MSM(N + L + 1) and
MSM(N + L) both are attained at the
point k = K + 1, so that
MSM(N + L + 1) – MSM(N + L)
=2[MSM(N + L – K) – MSM(N + L – K – 1)].
Choosing L = K, we get
MSM(N + K + 1) – MSM(N + K)
= 2[MSM(N) – MSM(N – 1)] = 2.3K, (2.3)
where, in the last equality, we have made use of (2.1). The above
equality, in view of (2.2), shows
that M > K.
To complete the proof of the lemma, we have to show that, M =
1 in (2.2) if and only if N = 1
(and K = 0). Now, (2.2) with N = 1 (and K = 0) reads as
MSM(M + 2) – MSM(M + 1) = 3.
Since MSM(3) – MSM(2) = 3, it follows, by virtue of Corollary
2.1, that M = 1.
Again, if M = 1 in (2.2), we get
MSM(N + 2) – MSM(N + 1) = 3K + 1. (2)
From (2.1), (2.3) and (2), we must have K = 0.
An implication of Lemma 2.4 is that, there are infinitely many
integers N such that MSM(N) –
MSM(N – 1) is of the form 3K for some integer K 0.
Given any integer N ( 1), by part (b) of Lemma 2.1, there is a
unique integer K ( 1) such that
MSM(N) is attained at k = K. Now, given any integer K ( 1), is
there an integer N ( 1) such that
MSM(N) is attained at k = K? The following proposition answers
the question in the affirmative.
Proposition 2.1 : Given any integer K 1, there is an integer N
1 such that MSM(N) is attained at
the point k = K 1.
Proof : The proof is by induction on K. The result is true for K =
1 with N = 1. So, we assume that
the result is true for some integer K 1, that is, we assume that,
for K ( 1), there is an integer N
such that MSM(N) is attained at k = K, so that
MSM(N) = 2MSM(N – K) + 1
2 (3K – 1).
Now, by Lemma 2.2, MSM(N + 1) is attained either at k = K, or
else, at k = K + 1. If MSM(N + 1) is
attained at k = K + 1, the proof by induction is complete.
Otherwise, MS(N + 1) is attained at k =
K, so that
MSM(N + 1) = 2MSM(N – K + 1) +
1
2
(3K – 1)
< 2MSM(N – K) +
1
2
(3K+1 – 1),
and hence
MSM(N + 1) – MSM(N) < 3K.
Now, if MS(N + 2) is attained at k = K + 1, the proof is
complete; otherwise
MSM(N + 2) = 2MSM(N – K + 2) +
1
2
(3K – 1)
< 2MSM(N – K + 1) +
1
2
(3K+1 – 1),
giving
MSM(N + 2) – MSM(N + 1) < 3K.
Thus, MSM(N + 1), MSM(N + 2), …, MSM(N + m), … are all
attained at k = K, with
MSM(N + i) – MSM(N + i – 1) < 3k, i = 1, 2, … .
But since the sequence 
1 ( ) ( 1 n MSM N i MSM N i is strictly increasing in i (
1) (by part
(a) of Lemma 2.1), there is an integer m ( 1) such that
MSM(N + m) – MSM(N + m – 1) 3K.
For the minimum such m, say, m = M, MSM(N + M – 1) is
attained at the point k = K but MSM(N
+ M) is attained at k = K + 1, with
MSM(N + M) – MSM(N + M – 1) = 3K.
Thus, corresponding to K + 1, there is an integer N + M such
that MSM(N + M) is attained at k = K
+ 1, which we intended to prove.
Proposition 2.1 shows that, given any integer K 1, there is an
integer N ( 1) such that MSM(N)
is attained at k = K. However, note that, such N is not unique.
For example, both MSM(3) and
MSM(4) are attained at k = 2.
Let the sequence 
n n1 b be defined by
an = MSM(n) MSM(n 1), n ≥ 1. (2.4)
Let kj ≥ 1 be defined by
a MSM(k ) MSM(k 1) 3 j; j 0 k j j j
(2.5)
with
k0 = 1. (2.6)
The Corollary below follows from Lemma 2.4, a proof of which
is given in Majumdar [3].
Corollary 2.2 : For all j 0, MSM(kj – 1) is attained at k = j;
moreover, for all n satisfying the
inequality kj ≤ n ≤ kj+1 – 1, MSM(n) is attained at k = j + 1.
Theorem 2.1 : For all j ≥ 0 and 1 s kj+1 – kj – 1,
MSM(kj + s) – MSM(kj + s – 1)= 2[MSM(kj + s – j – 1) –
MSM(kj + s – j – 2)].
Proof : By Corollary 2.2, MS(kj + s – 1) and MS(kj + s) both are
attained at the point k = j + 1, so
that by (1.5),
MSM(kj + s – 1) = 2MSM(kj + s – j – 2) + 1
2 (3j+1 – 1),
MSM(kj + s) = 2MSM(kj + s – j – 1) + 1
2 (3j+1 – 1).
We have the following result.
Lemma 2.5 : For all n ≥ 1, an is of the form 2i 3m for some
integers i ≥ 0, m ≥ 0.
Proof : Cleary, the result is true when n = 1 (with i = 0, m = 0).
To proceed by induction, we
assume that the result is true for all t ≤ n.
To prove the result for n + 1, we have to consider, by Lemma
2.2, the following two possibilities :
Case 1 : MSM(n) and MSM(n + 1) both are attained at k = K. In
this case,
an+1 = 2N – K;
Case 2 : MSM(n) is attained at k = K and MSM(n + 1) is attained
at k = K + 1. Here,
an+1 = 2[MSM(K + 1) – MSM(K)] = 2 aK+1.
Now, by the induction hypothesis, aK+1 is of the form 2i 3m,
and hence, so is an+1.
Lemma 2.5 shows that, for any n ≥ 1, an (defined through the
equation (2.4)) is of the form 2i 3m for some integers i ≥ 0 and
m ≥ 0. In the next section, we prove the converse, namely that,
given any
integer of the form 2i 3m, there is an integer n such that an = 2i
3m.
3. Main Results
Let 
n n1 b be the sequence of integers, arranged in (strictly)
increasing order :
bn = 2i 3m, i ≥ 0, m ≥ 0. (3.1)
The first few terms of n n 1 b


are
1, 2, 3, 4, 6, 8, 9, 12, 16, 18, 24, …,
and are known as 3-smooth numbers.
Given any integer j ≥ 0, we first derive the number of elements
of the sequence 
n n1 b such that 3j
< bn < 3j + 1, that is
3j < 2i 3m < 3j+1. (3)
Clearly, i must satisfy the inequality :
(j m )
ln3
ln2
< i < (j m + 1)
ln3
ln2
. (4)
From (3), we observe that, when j = 0, m = 0.
Let N(n, j) be the number of elements of the sequence 
n n1 b satisfying the inequality (3), that is,
( , ) :3 3 :3 2 3 3 . 1 1 j i m j
n
j
n
j
n N n j b b b (5)
Then, we have the following lemma, due to Majumdar [3],
which gives the recurrence relation
satisfied by N(n, j).
Lemma 3.1 : For any integer j ≥ 1,
( , 1) ( , ) 3 2 3 . 1 2 j i j N n j N n j i :
Corollary 3.1 : For any integer j ≥ 0,
N(n, j) = 
ln 2
ln3
:3 3 max : 2 3 ( 1) 1 1 b b i j j i j
n
j
n
where xis the floor of the real number x > 0.
Proof : The left-side part of the above chain of equalities has
been proved in Majumdar [3]. Now,
since
2i < 3j+1 if and only if i ln2 < (j + 1) ln 3,
the remaining part follows.
Next, we find the number of elements of the sequence 
n n1 b such that 3j < bn < 2.3j for any
integer j ≥ 0, that is,
3j < 2i 3m < 2.3j. (6)
: 3 2.3 : 3 2 3 2.3 . b b b j j j
n
j
n
j
n
i m 
Proof : First note that the inequality (6) is satisfied if and only if
(j m)
ln3
ln2
< i < (j m)
ln3
ln2
+1. (7)
When j = 0, m = 0, and the result is true. So, let j ≥ 1. In this
case, the inequality (7) admits j
number of solutions, corresponding to m = 0, 1, …, j – 1.
Since
|(j + 1)
ln3
ln2
|>j
for all j ≥ 1,
from Corollary 3.1 and Lemma 3.2, we see that
: : j j j j+1
n n n n b 3 b 2.3 b 3 b 3 , if j ≥ 1.
a : 3 a 2.3 b : 3 b 2.3 j. j j j j
n n n n 
Proof : In Lemma 2.4, let
N = kj, K = j.
Then, (2.3) reads as
j
k +j+1
j
a 2.3 .
Now, the number of elements of the sequence 
n n1 b between k
j
a and k +j+1
j
a is j. This,
coupled with Lemma 3.2, gives the result desired.
Let the sequence of numbers {pj}j0 be defined as follows :
p
j
b 3j, j 0. (3.2)
Clearly,
p0 = 1. (3.3)
The following lemma gives a recurrence relation satisfied by
Lemma 3.4 : For any n such that 3j < bn < 3j+1 for some integer
j ≥ 0, bn = 2bnj1.
Proof : When j = 0, 1 < b1 < 3, and
b2 = 2b1 = 2 b2– o–1,
and the result is true. So, we assume that the result is true for
(and up to) some j, that is, we
assume that, for all n such that 3s–1 < bn < 3s, 1 ≤ s ≤ j,
bn = 2 bn–s.
To prove the result for j + 1, we assume that
bp 3j < bn < 3j + 1bp
Clearly, any such bn is even. Also, since 3j < 2.3j
< 3j + 1, it follows that, for any N with 3j < bN <
2.3j, bN = 2 bM, for some integer M with 3j – 1 < bM < 3j.
Then, by the induction hypothesis,
bM = 2 bM–j.
Now,
: M N : .
nnnMnN
bb
22
b b b b b b N M 1
Therefore,
N,
M-j+(N-M-1) N-j-1
b
2
b b
so that
bN = 2 bN–j–1; 3j < bN < 2.3j.
Let L be the maximum such N so that
bL = 2 bL–j–1, bL+1 = 2.3j.
Now, since (by Lemma 3.2), there are j elements of the sequence

n n1 b such that
p
j
b < bn < p ,
j
2b
it follows that
L – pj = j,
so that
bL+1 = 2 bL–j.
Thus, the result is true for all n with 3j < bn ≤ 2.3j.
If j 1, there is at least one bn such that p
j
2b < bn < p
j+1
b , the minimum of which is bL+ 2.
Clearly,
L+2 p L-j + 1.
j+1
b 2b 2b
In general, if bL+ s < p ,
j+1
b 2 s ln3 ,
ln2
(j 1) j 1

then bL+ s = 2 bL+s–j–1.
We now prove the main result of the paper, given in the theorem
below.
Theorem 3.1 : For all n 1, an = bn.
Proof : Let kj n < kj+1 for some integer j 0 (so that 3j an
< 3j + 1, 3j bn< 3j + 1).
Then, by Theorem 2.1,
ksksj1
jj
a 2a for all 1 s kj+1 – kj – 1.
Thus, by virtue of Lemma 3.3, an and bn satisfy the same
recurrence relation, and hence, we get the
desired result.
The following theorem gives the solution of the recurrence
relation (1.2).
Theorem 3.2 : For n ≥ 1,
MSM(n) a b .
n
m
m
n
m
m 
1 1
Proof : Since MSM(n) can be written as
n
m1
MSM(n) MSM(m)-MSM(m-1) ,


we get the desired expression of MSM(n) by Theorem 3.1. Also,
by Corollary 2.2, MSM(n) is
attained at the unique point k = j + 1. Now, since 3j bn <
3j+1, we see that j must satisfy the
inequality j ln3 ≤ ln(bn) < (j + 1) ln3.
We then get the desired expression of k, given in the lemma.
A consequence of Theorem 3.2 is the following
Corollary 3.2 : For all j 0,
kj+1 = kj + ln3
ln2
(j 1) 

+ 1, k0 = 1.
Proof : follows immediately, since (by Corollary 3.1), the
number of elements of the sequence
Since k0 = 1, Corollary 3.2 allows us to find kj recursively in j.

Table 3.1 below gives the values of
kj for some small values of k, calculated on a computer, using
the recurrence relation in Corollary
3.2.
Corollary 3.2 is a new result which, together with Lemma 3.4,

enables us to find bn recursively for
any fixed n. For example, to find b17, we proceed as follows :
Looking at Table 3.1, since 12 < 17
< 19, it follows that j = 3 in Corollary 3.2, and so, by Lemma
3.4, b17 = 2 b13.
Thus, we need to find b13. From Table 3.1, we see that j = 3 in
Corollary 3.2, and so by Lemma
3.4, b13 = 2 b9.
To find b9, from Table 3.1, we see that j = 2 in Corollary 3.2, so
that b9 = 2 b6.
Similarly, b6 = 2 b4 = 8.
Plugging in the values of b7 and b10, we finally get, b17 = 64.
Corollary 3.3 : For all n 1, ( ) 2 .
n
m1
n
m1


MS n am bm
Proof
: follows by virtue of Theorem 3.2, together with (1.3) and
Lemma 1.1.
The expression of MS(n) is given in Corollary 3.3 above. As has
been pointed out by Stockmeyer
[6] (without proof), an interpretation of the expression of MS(n)
is as follows : In the star puzzle
with n discs d1, d2, …, dn in increasing order (so that d1 is the
smallest disc and dn is the largest
one), the disc dn – m + 1 needs exactly 2 bm number of moves
(1 m n) under the optimal strategy.
4. Concluding Remarks
Stockmeyer [6] gives an outline of a proof of the results in
Corollary 3.3, but his argument is
heuristic, is not supported by any theoretical development, and is
incomplete in the sense that it
lacks the proof that
MS(n) – MS(n – 1) = 2bn.
This paper gives a rigorous treatment of the problem, which
unveils many interesting properties.
Lemma 3.4 gives a new recurrence relation satisfied by the
sequence 
n n1 b and Corollary 3.2
gives a recurrence relation satisfied by kj. It is interesting that
MSM(n) satisfies the same
recurrence relation as that of bn.
From Corollary 3.3, we observe that, it in fact offers two
methods of finding MS(n) (for any n 1).
The first method is in terms of the sequence 
n n1 b . When n is small, we can readily find MS(n)
by this method. For example, adding the first 6 terms of the
sequence 
n n1 b , we see that MS(6)
= 48. For large n, we may use Lemma 3.4 to find bn. Thus, for
example, to find MS(17), by
Corollary 3.3, we have
MS(17) = 2[b1
+ b2 + b3 + (b4 + b5
+ b6) + b7 + (b8 + b9 + b10 + b11)
+ b12 + (b13 + b14 + b15 + b16 + b17)]
Now,
b8 = 2b5, b9 = 2b6, b10 = 2b7, b11 = 2b8,
b13 = 2b9, b14 = 2b10, b15 = 2b11, b16 = 2b12, b17 = 2b13,
b7 = 32 = 9, b12 = 33 = 27.
Thus,
MS(17) = 2[1 + 2 + 3 + (4 + 6 + 8) + 9 + (12 + 16 + 18 + 24)
+ 27 + (32 + 36 + 48 + 54 + 64)] = 728.
Alternatively, from Corollary 3.3, since MS(17) is attained at k
= 3 + 1 = 4, we get
MS(17) = 2MS(13) + 34 – 1 = 2MS(13) + 80.
Now, by Corollary 3.3 again, MS(13) is attained at k = 3 + 1 =
4, so that
MS(13) = 2MS(9) + 34 – 1 = 2MS(9) + 80.
Since MS(9) is attained at k = 3, we get
MS(9) = 2MS(6) + 33 – 1 = 2MS(6) + 26.
Now, MS(6) is attained at k = 2, and so
MS(6) = 2MS(4) + 32 – 1 = 40 + 8 = 48.
Finally, we get MS(17) = 728.
The following recurrence relation has been considered by
Matsuura [4] :
min , ,
n k T(n, ) T(k, ) 2 1 n 1
0kn1


T(0, α) = 0 for all 3,
where 3 is an integer. The problem was taken up by
Majumdar [2], who studied some of the
properties of T(n, ). Of particular interest is T(n, 3). It has been
shown by Matsuura [4], by
induction on n, that
T(n, 3) – T(n – 1, 3) = bn for all n 1,
so that MSM(n) = T(n, 3) for all n ≥ 1.
In passing, it may be mentioned here that, a second recurrence
relation satisfied by the sequence

n n1 b has been derived by Matsuura [4], and is given below.
Lemma 4.1 : For any n such that 2i < bn < 2i +1 for some
integer
i 0. bn = 3bn – i – 1.
To make the above result applicable, it is to be supplemented by
Lemma 4.3. To prove Lemma
4.3, we need the result below.
Lemma 4.2 : For any integer i 0,
: i i+1
nn
ln2
ln3
b 2 b 2 (i 1) .

Proof : is similar to that of Corollary 3.1, and is left as an
exercise.
Lemma 4.3 : Let the sequence of integers {qi}i0 be defined as
follows :
qi b = 2i, i 0; q0 = 1.
Then,
qi+1 = qi + ( )ln2
ln3
i + 1 

+ 1.
Proof : is similar to Corollary 3.2, and is omitted here.
Using Lemma 4.3, we can form the table below, which would
help us in finding bn for any n 1
fixed.
CHAPTER 7
ANALYTIC FUNCTIONS
2.1 Introduction
The main goal of this topic is to define and give some of
the important properties of complex analytic functions.
A function f(z) is analytic if it has a complex derivative f 0
(z). In general, the rules for computing derivatives will be
familiar to you from single variable calculus. However, a
much richer set of conclusions can be drawn about a
complex analytic function than is generally true about
real differentiable functions.
2.2 The derivative: preliminaries
In calculus we defined the derivative as a limit. In
complex analysis we will do the same. f 0 (z) = lim ∆z→0
∆f ∆z = lim ∆z→0 f(z + ∆z) − f(z) ∆z . Before giving the
derivative our full attention we are going to have to
spend some time exploring and understanding limits. To
motivate this we’ll first look at two simple examples –
one positive and one negative.
Example 2.1. Find the derivative of f(z) = z 2 . Solution:
We compute using the definition of the derivative as a
limit. lim ∆z→0 (z + ∆z) 2 − z 2 ∆z = lim ∆z→0 z 2 + 2z∆z +
(∆z) 2 − z 2 ∆z = lim ∆z→0 2z + ∆z = 2z. That was a
positive example. Here’s a negative one which shows
that we need a careful understanding of limits.
Example 2.2. Let f(z) = z. Show that the limit for f 0 (0)
does not converge. Solution: Let’s try to compute f 0 (0)
using a limit: f 0 (0) = lim ∆z→0 f(∆z) − f(0) ∆z = lim ∆z→0
∆z ∆z = ∆x − i∆y ∆x + i∆y . Here we used ∆z = ∆x + i∆y.
Now, ∆z → 0 means both ∆x and ∆y have to go to 0.
There are lots of ways to do this. For example, if we let
∆z go to 0 along the x-axis then, ∆y = 0 while ∆x goes to
0. In this case, we would have f 0 (0) = lim ∆x→0 ∆x ∆x =
1. On the other hand, if we let ∆z go to 0 along the
positive y-axis then f 0 (0) = lim ∆y→0 −i∆y i∆y = −1. 1 2
ANALYTIC FUNCTIONS 2
The limits don’t agree! The problem is that the limit
depends on how ∆z approaches 0. If we came from other
directions we’d get other values. There’s nothing to do,
but agree that the limit does not exist. Well, there is
something we can do: explore and understand limits.
Let’s do that now. 2.3 Open disks, open deleted disks,
open regions Definition. The open disk of radius r around
z0 is the set of points z with |z − z0| < r, i.e. all points
within distance r of z0. The open deleted disk of radius r
around z0 is the set of points z with 0 < |z−z0| < r. That
is, we remove the center z0 from the open disk. A
deleted disk is also called a punctured disk. z0 r z0 r Left:
an open disk around z0; right: a deleted open disk
around z0 Definition. An open region in the complex
plane is a set A with the property that every point in A
can be be surrounded by an open disk that lies entirely
in A. We will often drop the word open and simply call A
a region. In the figure below, the set A on the left is an
open region because for every point in A we can draw a
little circle around the point that is completely in A. (The
dashed boundary line indicates that the boundary of A is
not part of A.) In contrast, the set B is not an open
region. Notice the point z shown is on the boundary, so
every disk around z contains points outside B. Left: an
open region A; right: B is not an open region 2.4 Limits
and continuous functions Definition. If f(z) is defined on
a punctured disk around z0 then we say limz→z0 f(z) =
w0 if f(z) goes to w0 no matter what direction z
approaches z0. The figure below shows several
sequences of points that approach z0. If limz→z0 f(z) =
w0 then f(z) must go to w0 along each of these
sequences. ANALYTIC FUNCTIONS 3 Sequences going to
z0 are mapped to sequences going to w0.
Example 2.3. Many functions have obvious limits. For
example: lim z→2 z 2 = 4 and lim z→2 (z 2 + 2)/(z 3 + 1) =
6/9. Here is an example where the limit doesn’t exist
because different sequences give different limits.
Example 2.4. (No limit) Show that lim z→0 z z = lim z→0
x + iy x − iy does not exist. Solution: On the real axis we
have z z = x x = 1, so the limit as z → 0 along the real axis
is 1. By contrast, on the imaginary axis we have z z = iy
−iy = −1, so the limit as z → 0 along the imaginary axis is
-1. Since the two limits do not agree the limit as z → 0
does not exist! 2.4.1 Properties of limits We have the
usual properties of limits. Suppose limz→z0 f(z) = w1 and
limz→z0 g(z) = w2 then • limz→z0 f(z) + g(z) = w1 + w2. •
limz→z0 f(z)g(z) = w1 · w2. 2 ANALYTIC FUNCTIONS 4 • If
w2 6= 0 then limz→z0 f(z)/g(z) = w1/w2 • If h(z) is
continuous and defined on a neighborhood of w1 then
limz→z0 h(f(z)) = h(w1) (Note: we will give the official
definition of continuity in the next section.) We won’t
give a proof of these properties. As a challenge, you can
try to supply it using the formal definition of limits given
in the appendix. We can restate the definition of limit in
terms of functions of (x, y). To this end, let’s write f(z) =
f(x + iy) = u(x, y) + iv(x, y) and abbreviate P = (x, y), P0 =
(x0, y0), w0 = u0 + iv0. Then limz→z0 f(z) = w0 iff
( limP→P0 u(x, y) = u0 limP→P0 v(x, y) = v0. Note. The
term ‘iff’ stands for ‘if and only if’ which is another way
of saying ‘is equivalent to’. 2.4.2 Continuous functions A
function is continuous if it doesn’t have any sudden
jumps. This is the gist of the following definition.
Definition.
If the function f(z) is defined on an open disk around z0
and limz→z0 f(z) = f(z0) then we say f is continuous at z0.
If f is defined on an open region A then the phrase ‘f is
continuous on A’ means that f is continuous at every
point in A. As usual, we can rephrase this in terms of
functions of (x, y): Fact. f(z) = u(x, y) + iv(x, y) is
continuous iff u(x, y) and v(x, y) are continuous as
functions of two variables. Example 2.5. (Some
continuous functions) (i) A polynomial P(z) = a0 + a1z +
a2z 2 + . . . + anz n is continuous on the entire plane.
Reason: it is clear that each power (x+iy) k is continuous
as a function of (x, y). (ii) The exponential function is
continuous on the entire plane. Reason: e z = ex+iy = ex
cos(y) + ie x sin(y). So the both the real and imaginary
parts are clearly continuous as a function of (x, y). (iii)
The principal branch Arg(z) is continuous on the plane
minus the non-positive real axis. Reason: this is clear and
is the reason we defined branch cuts for arg. We have to
remove the negative real axis because Arg(z) jumps by
2π when you cross it. We also have to remove z = 0
because Arg(z) is not even defined at 0.
(iv) The principal branch of the function log(z) is
continuous on the plane minus the nonpositive real axis.
Reason: the principal branch of log has log(z) = log(r) + i
Arg(z). So the continuity of log(z) follows from the
continuity of Arg(z). 2.4.3 Properties of continuous
functions Since continuity is defined in terms of limits,
we have the following properties of continuous
functions. Suppose f(z) and g(z) are continuous on a
region A. Then • f(z) + g(z) is continuous on A. • f(z)g(z) is
continuous on A. • f(z)/g(z) is continuous on A except
(possibly) at points where g(z) = 0. • If h is continuous on
f(A) then h(f(z)) is continuous on A. Using these
properties we can claim continuity for each of the
following functions: • e z 2 • cos(z) = (eiz + e−iz)/2 • If
P(z) and Q(z) are polynomials then P(z)/Q(z) is
continuous except at roots of Q(z). 2.5 The point at
infinity By definition the extended complex plane =
C∪{∞}. That is, we have one point at infinity to be
thought of in a limiting sense described as follows. A
sequence of points {zn} goes to infinity if |zn| goes to
infinity. This “point at infinity” is approached in any
direction we go. All of the sequences shown in the figure
below are growing, so they all go to the (same) “point at
infinity”. Various sequences all going to infinity.
If we draw a large circle around 0 in the plane, then we
call the region outside this circle a neighborhood of
infinity. R Re(z) Im(z) The shaded region outside the
circle of radius R is a neighborhood of infinity. 2.5.1
Limits involving infinity The key idea is 1/∞ = 0. By this
we mean limz→∞ 1 z = 0 We then have the following
facts: • limz→z0 f(z) = ∞ ⇔ limz→z0 1/f(z) = 0 • limz→∞
f(z) = w0 ⇔ lim z→0 f(1/z) = w0 • limz→∞ f(z) = ∞ ⇔ lim
z→0 1 f(1/z) = 0 Example 2.6. limz→∞ e z is not defined
because it has different values if we go to infinity in
different directions, e.g. we have ez = ex e iy and lim
x→−∞ e x e iy = 0 lim x→+∞ e x e iy = ∞ lim y→+∞ e x e
iy is not defined, since x is constant, so ex e iy loops in a
circle indefinitely. Example 2.7. Show limz→∞ z n = ∞
(for n a positive integer). Solution: We need to show that
|z n | gets large as |z| gets large. Write z = Reiθ, then |z
n | = |R n e inθ| = R n = |z| n 2.5.2 Stereographic
projection from the Riemann sphere This is a lovely
section and we suggest you read it. However it will be a
while before we use it in 18.04.
One way to visualize the point at ∞ is by using a (unit)
Riemann sphere and the associated stereographic
projection. The figure below shows a sphere whose
equator is the unit circle in the complex plane.
Stereographic projection from the sphere to the plane.
Stereographic projection from the sphere to the plane is
accomplished by drawing the secant line from the north
pole N through a point on the sphere and seeing where
it intersects the plane. This gives a 1-1 correspondence
between a point on the sphere P and a point in the
complex plane z. It is easy to see show that the formula
for stereographic projection is P = (a, b, c) 7→ z = a 1 − c
+ i b 1 − c . The point N = (0, 0, 1) is special, the secant
lines from N through P become tangent lines to the
sphere at N which never intersect the plane. We
consider N the point at infinity. In the figure above, the
region outside the large circle through the point z is a
neighborhood of infinity. It corresponds to the small
circular cap around N on the sphere. That is, the small
cap around N is a neighborhood of the point at infinity
on the sphere! The figure below shows another common
version of stereographic projection. In this figure the
sphere sits with its south pole at the origin. We still
project using secant lines from the north pole. 2.6
Derivatives The definition of the complex derivative of a
complex function is similar to that of a real derivative of
a real function: For a function f(z) the derivative f at z0 is
defined as f 0 (z0) = limz→z0 f(z) − f(z0) z − z0 Provided,
of course, that the limit exists. If the limit exists we say f
is analytic at z0 or f is differentiable at z0. 2 ANALYTIC
FUNCTIONS 8 Remember: The limit has to exist and be
the same no matter how you approach z0! If f is analytic
at all the points in an open region A then we say f is
analytic on A. As usual with derivatives there are several
alternative notations. For example, if w = f(z) we can
write f 0 (z0) = dw dz z0 = limz→z0 f(z) − f(z0) z − z0 =
lim ∆z→0 ∆w ∆z Example 2.8. Find the derivative of f(z) =
z 2 . Solution: We did this above in Example 2.1. Take a
look at that now. Of course, f 0 (z) = 2z.
Example 2.9. Show f(z) = z is not differentiable at any
point z. Solution: We did this above in Example 2.2. Take
a look at that now. Challenge. Use polar coordinates to
show the limit in the previous example can be any value
with modulus 1 depending on the angle at which z
approaches z0. 2.6.1 Derivative rules It wouldn’t be
much fun to compute every derivative using limits.
Fortunately, we have the same differentiation formulas
as for real-valued functions. That is, assuming f and g are
differentiable we have: • Sum rule: d dz (f(z) + g(z)) = f 0
+ g 0 • Product rule: d dz (f(z)g(z)) = f 0 g + fg0 • Quotient
rule: d dz (f(z)/g(z)) = f 0 g − fg0 g 2 • Chain rule: d dz
g(f(z)) = g 0 (f(z))f 0 (z) • Inverse rule: df −1 (z) dz = 1 f 0(f
−1(z)) To give you the flavor of these arguments we’ll
prove the product rule. d dz (f(z)g(z)) = limz→z0 f(z)g(z) −
f(z0)g(z0) z − z0 = limz→z0 (f(z) − f(z0))g(z) + f(z0)(g(z) −
g(z0)) z − z0 = limz→z0 f(z) − f(z0) z − z0 g(z) + f(z0) (g(z) −
g(z0)) z − z0 = f 0 (z0)g(z0) + f(z0)g 0 (z0) Here is an
important fact that you would have guessed. We will
prove it in the next section. Theorem. If f(z) is defined
and differentiable on an open disk and f 0 (z) = 0 on the
disk then f(z) is constant.
2.7 Cauchy-Riemann equations The Cauchy-Riemann
equations are our first consequence of the fact that the
limit defining f(z) must be the same no matter which
direction you approach z from. The CauchyRiemann
equations will be one of the most important tools in our
toolbox. 2.7.1 Partial derivatives as limits Before getting
to the Cauchy-Riemann equations we remind you about
partial derivatives. If u(x, y) is a function of two variables
then the partial derivatives of u are defined as ∂u ∂x(x,
y) = lim ∆x→0 u(x + ∆x, y) − u(x, y) ∆x , i.e. the derivative
of u holding y constant. ∂u ∂y (x, y) = lim ∆y→0 u(x, y +
∆y) − u(x, y) ∆y , i.e. the derivative of u holding x
constant. 2.7.2 The Cauchy-Riemann equations The
Cauchy-Riemann equations use the partial derivatives of
u and v to allow us to do two things: first, to check if f
has a complex derivative and second, to compute that
derivative. We start by stating the equations as a
theorem. Theorem 2.10. (Cauchy-Riemann equations) If
f(z) = u(x, y) + iv(x, y) is analytic (complex differentiable)
then f 0 (z) = ∂u ∂x + i ∂v ∂x = ∂v ∂y − i ∂u ∂y In
particular, ∂u ∂x = ∂v ∂y and ∂u ∂y = − ∂v ∂x. This last set
of partial differential equations is what is usually meant
by the Cauchy-Riemann equations. Here is the short
form of the Cauchy-Riemann equations: ux = vy uy = −vx
Proof. Let’s suppose that f(z) is differentiable in some
region A and f(z) = f(x + iy) = u(x, y) + iv(x, y). We’ll
compute f 0 (z) by approaching z first from the horizontal
direction and then from the vertical direction. We’ll use
the formula f 0 (z) = lim ∆z→0 f(z + ∆z) − f(z) ∆z , 2
ANALYTIC FUNCTIONS 10 where ∆z = ∆x + i∆y. Horizontal
direction: ∆y = 0, ∆z = ∆x f 0 (z) = lim ∆z→0 f(z + ∆z) − f(z)
∆z = lim ∆x→0 f(x + ∆x + iy) − f(x + iy) ∆x = lim ∆x→0 (u(x
+ ∆x, y) + iv(x + ∆x, y)) − (u(x, y) + iv(x, y)) ∆x = lim ∆x→0
u(x + ∆x, y) − u(x, y) ∆x + i v(x + ∆x, y) − v(x, y) ∆x = ∂u
∂x(x, y) + i ∂v ∂x(x, y) Vertical direction: ∆x = 0, ∆z = i∆y
(We’ll do this one a little faster.) f 0 (z) = lim ∆z→0 f(z +
∆z) − f(z) ∆z = lim ∆y→0 (u(x, y + ∆y) + iv(x, y + ∆y)) − (u(x,
y) + iv(x, y)) i∆y = lim ∆y→0 u(x, y + ∆y) − u(x, y) i∆y + i
v(x, y + ∆y) − v(x, y) i∆y = 1 i ∂u ∂y (x, y) + ∂v ∂y (x, y) = ∂v
∂y (x, y) − i ∂u ∂y (x, y) We have found two different
representations of f 0 (z) in terms of the partials of u and
v. If put them together we have the Cauchy-Riemann
equations: f 0 (z) = ∂u ∂x + i ∂v ∂x = ∂v ∂y − i ∂u ∂y ⇒ ∂u
∂x = ∂v ∂y , and − ∂u ∂y = ∂v ∂x. It turns out that the
converse is true and will be very useful to us. Theorem.
Consider the function f(z) = u(x, y) + iv(x, y) defined on a
region A. If u and v satisfy the Cauchy-Riemann
equations and have continuous partials then f(z) is
differentiable on A. The proof of this is a tricky exercise
in analysis. It is somewhat beyond the scope of this
class, so we will skip it. If you’re interested, with a little
effort you should be able to grasp it. 2.7.3 Using the
Cauchy-Riemann equations The Cauchy-Riemann
equations provide us with a direct way of checking that a
function is differentiable and computing its derivative.
Example 2.11. Use the Cauchy-Riemann equations to
show that ez is differentiable and its derivative is ez . 2
ANALYTIC FUNCTIONS 11 Solution: We write ez = ex+iy =
ex cos(y) + ie x sin(y). So u(x, y) = ex cos(y) and v(x, y) =
ex sin(y). Computing partial derivatives we have ux = ex
cos(y), uy = −e x sin(y) vx = ex sin(y), vy = ex cos(y) We
see that ux = vy and uy = −vx, so the Cauchy-Riemann
equations are satisfied. Thus, e z is differentiable and d
dz e z = ux + ivx = ex cos(y) + ie x sin(y) = ez . Example
2.12. Use the Cauchy-Riemann equations to show that
f(z) = z is not differentiable. Solution: f(x + iy) = x − iy, so
u(x, y) = x, v(x, y) = −y. Taking partial derivatives ux = 1,
uy = 0, vx = 0, vy = −1 Since ux 6= vy the Cauchy-Riemann
equations are not satisfied and therefore f is not
differentiable. Theorem. If f(z) is differentiable on a disk
and f 0 (z) = 0 on the disk then f(z) is constant. Proof.
Since f is differentiable and f 0 (z) ≡ 0, the Cauchy-
Riemann equations show that ux(x, y) = uy(x, y) = vx(x, y)
= vy(x, y) = 0 We know from multivariable calculus that a
function of (x, y) with both partials identically zero is
constant. Thus u and v are constant, and therefore so is
f. 2.7.4 f 0 (z) as a 2 × 2 matrix Recall that we could
represent a complex number a + ib as a 2 × 2 matrix a +
ib ↔ a −b b a . (1) Now if we write f(z) in terms of (x, y)
we have f(z) = f(x + iy) = u(x, y) + iv(x, y) ↔ f(x, y) = (u(x,
y), v(x, y)). We have f 0 (z) = ux + ivx, so we can represent
f 0 (z) as ux −vx vx ux . 2 ANALYTIC FUNCTIONS 12 Using
the Cauchy-Riemann equations we can replace −vx by uy
and ux by vy which gives us the representation f 0 (z) ↔
ux uy vx vy , i.e, f 0 (z) is just the Jacobian of f(x, y). For
me, it is easier to remember the Jacobian than the
Cauchy-Riemann equations. Since f 0 (z) is a complex
number I can use the matrix representation in Equation
1 to remember the Cauchy-Riemann equations! 2.8
Cauchy-Riemann all the way down We’ve defined an
analytic function as one having a complex derivative.
The following theorem shows that if f is analytic then so
is f 0 . Thus, there are derivatives all the way down!
Theorem 2.13. Assume the second order partials of u
and v exist and are continuous. If f(z) = u + iv is analytic,
then so is f 0 (z). Proof. To show this we have to prove
that f 0 (z) satisfies the Cauchy-Riemann equations. If f =
u + iv we know ux = vy, uy = −vx, f0 = ux + ivx. Let’s write
f 0 = U + iV, so, by Cauchy-Riemann, U = ux = vy, V = vx =
−uy. (2) We want to show that Ux = Vy and Uy = −Vx. We
do them one at a time. To prove Ux = Vy, we use
Equation ?? to see that Ux = vyx and Vy = vxy. Since vxy =
vyx, we have Ux = Vy. Similarly, to show Uy = −Vx, we
compute Uy = uxy and Vx = −uyx. So, Uy = −Vx. QED.
Technical point. We’ve assumed as many partials as we
need. So far we can’t guarantee that all the partials
exist. Soon we will have a theorem which says that an
analytic function has derivatives of all order. We’ll just
assume that for now. In any case, in most examples this
will be obvious. 2.9 Gallery of functions In this section
we’ll look at many of the functions you know and love as
functions of z. For each one we’ll have to do three
things. 2 ANALYTIC FUNCTIONS 13 1. Define how to
compute it. 2. Specify a branch (if necessary) giving its
range. 3. Specify a domain (with branch cut if necessary)
where it is analytic. 4. Compute its derivative. Most
often, we can compute the derivatives of a function
using the algebraic rules like the quotient rule. If
necessary we can use the Cauchy-Riemann equations or,
as a last resort, even the definition of the derivative as a
limit. Before we start on the gallery we define the term
“entire function”. Definition. A function that is analytic
at every point in the complex plane is called an entire
function. We will see that ez , z n , sin(z) are all entire
functions. 2.9.1 Gallery of functions, derivatives and
properties The following is a concise list of a number of
functions and their complex derivatives. None of the
derivatives will surprise you. We also give important
properties for some of the functions. The proofs for each
follow below. 1. f(z) = ez = ex cos(y) + ie x sin(y). Domain
= all of C (f is entire). f 0 (z) = ez . 2. f(z) ≡ c (constant)
Domain = all of C (f is entire). f 0 (z) = 0. 3. f(z) = z n (n an
integer ≥ 0) Domain = all of C (f is entire). f 0 (z) = nzn−1 .
4. P(z) (polynomial) A polynomial has the form P(z) = anz
n + an−1z n−1 + . . . + a0. Domain = all of C (P(z) is entire).
P 0 (z) = nanz n−1 + (n − 1)an−1z n−1 + . . . + 2a2z + a1. 5.
f(z) = 1/z Domain = C − {0} (the punctured plane). f 0 (z) =
−1/z2 . 6. f(z) = P(z)/Q(z) (rational function). When P and
Q are polynomials P(z)/Q(z) is called a rational function.
2 ANALYTIC FUNCTIONS 14 If we assume that P and Q
have no common roots, then: Domain = C − {roots of Q} f
0 (z) = P 0Q − P Q0 Q2 . 7. sin(z), cos(z) Definition. cos(z) =
e iz + e−iz 2 , sin(z) = e iz − e −iz 2i (By Euler’s formula we
know this is consistent with cos(x) and sin(x) when z = x
is real.) Domain: these functions are entire. d cos(z) dz =
− sin(z), d sin(z) dz = cos(z). Other key properties of sin
and cos: - cos2 (z) + sin2 (z) = 1 - e z = cos(z) + isin(z) -
Periodic in x with period 2π, e.g. sin(x + 2π + iy) = sin(x +
iy). - They are not bounded! - In the form f(z) = u(x, y) +
iv(x, y) we have cos(z) = cos(x) cosh(y) − isin(x) sinh(y)
sin(z) = sin(x) cosh(y) + i cos(x) sinh(y) (cosh and sinh are
defined below.) - The zeros of sin(z) are z = nπ for n any
integer. The zeros of cos(z) are z = π/2 + nπ for n any
integer. (That is, they have only real zeros that you
learned about in your trig. class.) 8. Other trig functions
cot(z), sec(z) etc. Definition. The same as for the real
versions of these function, e.g. cot(z) = cos(z)/ sin(z),
sec(z) = 1/ cos(z). Domain: The entire plane minus the
zeros of the denominator. Derivative: Compute using the
quotient rule, e.g. d tan(z) dz = d dz sin(z) cos(z) = cos(z)
cos(z) − sin(z)(− sin(z)) cos2(z) = 1 cos2(z) = sec2 z (No
surprises there!) 9. sinh(z), cosh(z) (hyperbolic sine and
cosine) Definition. cosh(z) = e z + e−z 2 , sinh(z) = e z − e
−z 2 2 ANALYTIC FUNCTIONS 15 Domain: these functions
are entire. d cosh(z) dz = sinh(z), d sinh(z) dz = cosh(z)
Other key properties of cosh and sinh: - cosh2 (z) − sinh2
(z) = 1 - For real x, cosh(x) is real and positive, sinh(x) is
real. - cosh(iz) = cos(z), sinh(z) = −isin(iz). 10. log(z) (See
Topic 1.) Definition. log(z) = log(|z|) + i arg(z). Branch:
Any branch of arg(z). Domain: C minus a branch cut
where the chosen branch of arg(z) is discontinuous. d dz
log(z) = 1 z 11. z a (any complex a) Definition. z a = ea
log(z) . Branch: Any branch of log(z). Domain: Generally
the domain is C minus a branch cut of log. If a is an
integer ≥ 0 then z a is entire. If a is a negative integer
then z a is defined and analytic on C − {0}. dza dz =
aza−1 . 12. sin−1 (z) Definition. sin−1 (z) = −ilog(iz + √ 1 −
z 2). The definition is chosen so that sin(sin−1 (z)) = z. The
derivation of the formula is as follows. Let w = sin−1 (z),
so z = sin(w). Then, z = e iw − e −iw 2i ⇒ e 2iw − 2ize iw −
1 = 0 Solving the quadratic in eiw gives e iw = 2iz + √ −4z
2 + 4 2 = iz + p 1 − z 2. Taking the log gives iw = log(iz + p
1 − z 2) ⇔ w = −ilog(iz + p 1 − z 2). From the definition we
can compute the derivative: d dz sin−1 (z) = 1 √ 1 − z 2 .
Choosing a branch is tricky because both the square root
and the log require choices. We will look at this more
carefully in the future. For now, the following discussion
and figure are for your amusement. Sine (likewise
cosine) is not a 1-1 function, so if we want sin−1 (z) to be
single-valued then we have to choose a region where
sin(z) is 1-1. (This will be a branch of sin−1 (z), i.e. a range
for the image,) The figure below shows a domain where
sin(z) is 1-1. The domain consists of the vertical strip z = x
+ iy with −π/2 < x < π/2 together with the two rays on
boundary where y ≥ 0 (shown as red lines). The figure
indicates how the regions making up the domain in the
z-plane are mapped to the quadrants in the w-plane. A
domain where z 7→ w = sin(z) is one-to-one 2.9.2 A few
proofs Here we prove at least some of the facts stated in
the list just above. 1. f(z) = ez . This was done in Example
2.11 using the Cauchy-Riemann equations. 2. f(z) ≡ c
(constant). This case is trivial. 3. f(z) = z n (n an integer ≥
0): show f 0 (z) = nzn−1 It’s probably easiest to use the
definition of derivative directly. Before doing that we
note the factorization z n − z n 0 = (z − z0)(z n−1 + z n−2
z0 + z n−3 z 2 0 + . . . + z 2 z n−3 0 + zzn−2 0 + z n−1 0 )
Now f 0 (z0) = limz→z0 f(z) − f(z0) z − z0 = limz→z0 z n − z
n 0 z − z0 = limz→z0 (z n−1 + z n−2 z0 + z n−3 z 2 0 + . . . +
z 2 z n−3 0 + zzn−2 0 + z n−1 0 ) = nzn−1 0 . Since we
showed directly that the derivative exists for all z, the
function must be entire.
4. P(z) (polynomial). Since a polynomial is a sum of
monomials, the formula for the derivative follows from
the derivative rule for sums and the case f(z) = z n .
Likewise the fact the P(z) is entire. 5. f(z) = 1/z. This
follows from the quotient rule. 6. f(z) = P(z)/Q(z). This
also follows from the quotient rule. 7. sin(z), cos(z). All
the facts about sin(z) and cos(z) follow from their
definition in terms of exponentials. 8. Other trig
functions cot(z), sec(z) etc. Since these are all defined in
terms of cos and sin, all the facts about these functions
follow from the derivative rules. 9. sinh(z), cosh(z). All
the facts about sinh(z) and cosh(z) follow from their
definition in terms of exponentials. 10. log(z). The
derivative of log(z) can be found by differentiating the
relation elog(z) = z using the chain rule. Let w = log(z), so
ew = z and d dz e w = dz dz = 1 ⇒ de w dw dw dz = 1 ⇒ e
w dw dz = 1 ⇒ dw dz = 1 e w Using w = log(z) we get d
log(z) dz = 1 z . 11. z a (any complex a). The derivative for
this follows from the formula z a = ea log(z) ⇒ dza dz =
ea log(z) · a z = aza z = aza−1 2.10 Branch cuts and
function composition We often compose functions, i.e.
f(g(z)). In general in this case we have the chain rule to
compute the derivative. However we need to specify the
domain for z where the function is analytic. And when
branches and branch cuts are involved we need to take
care. Example 2.14. Let f(z) = ez 2 . Since ez and z 2 are
both entire functions, so is f(z) = ez 2 . The chain rule
gives us f 0 (z) = ez 2 (2z). Example 2.15. Let f(z) = ez and
g(z) = 1/z. f(z) is entire and g(z) is analytic everywhere
but 0. So f(g(z)) is analytic except at 0 and df(g(z)) dz = f 0
(g(z))g 0 (z) = e1/z · −1 z 2 . Example 2.16. Let h(z) = 1/(ez
−1). Clearly h is entire except where the denominator is
0. The denominator is 0 when ez − 1 = 0. That is, when z
= 2πni for any integer n. Thus, h(z) is analytic on the set C
− {2πni, where n is any integer} 2 ANALYTIC FUNCTIONS
18 The quotient rule gives h 0 (z) = −e z /(ez − 1)2 . A little
more formally: h(z) = f(g(z)). where f(w) = 1/w and w =
g(z) = ez −1. We know that g(z) is entire and f(w) is
analytic everywhere except w = 0. Therefore, f(g(z)) is
analytic everywhere except where g(z) = 0. Example
2.17. It can happen that the derivative has a larger
domain where it is analytic than the original function.
The main example is f(z) = log(z). This is analytic on C
minus a branch cut. However d dz log(z) = 1 z is analytic
on C − {0}. The converse can’t happen. Example 2.18.
Define a region where √ 1 − z is analytic. Solution:
Choosing the principal branch of argument, we have √ w
is analytic on C − {x ≤ 0, y = 0}, (see figure below.). So √ 1
− z is analytic except where w = 1 − z is on the branch
cut, i.e. where w = 1 − z is real and ≤ 0. It’s easy to see
that w = 1 − z is real and ≤ 0 ⇔ z is real and ≥ 1. So √ 1 − z
is analytic on the region (see figure below) C − {x ≥ 1, y =
0} Note. A different branch choice for √ w would lead to
a different region where √ 1 − z is analytic. The figure
below shows the domains with branch cuts for this
example. Re(w) Im(w) Re(z) Im(z) 1 domain for √ w
domain for √ 1 − z Example 2.19. Define a region where
f(z) = √ 1 + ez is analytic. Solution: Again, let’s take √ w to
be analytic on the region C − {x ≤ 0, y = 0} So, f(z) is
analytic except where 1 + ez is real and ≤ 0. That is,
except where ez is real and ≤ −1. Now, ez = ex e iy is real
only when y is a multiple of π. It is negative only when y
is an odd mutltiple of π. It has magnitude greater than 1
only when x > 0. Therefore f(z) is analytic on the region C
− {x ≥ 0, y = odd multiple of π} 2 ANALYTIC FUNCTIONS
19 The figure below shows the domains with branch cuts
for this example. Re(w) Im(w) Re(z) Im(z) −3πi −πi πi 3πi
domain for √ w domain for √ e z + 1 2.11 Appendix:
Limits The intuitive idea behind limits is relatively
simple. Still, in the 19th century mathematicians were
troubled by the lack of rigor, so they set about putting
limits and analysis on a firm footing with careful
definitions and proofs. In this appendix we give you the
formal definition and connect it to the intuitive idea. In
18.04 we will not need this level of formality. Still, it’s
nice to know the foundations are solid, and some
students may find this interesting. 2.11.1 Limits of
sequences Intuitively, we say a sequence of complex
numbers z1, z2, . . . converges to a if for large n, zn is
really close to a. To be a little more precise, if we put a
small circle of radius around a then eventually the
sequence should stay inside the circle. Let’s refer to this
as the sequence being captured by the circle. This has to
be true for any circle no matter how small, though it may
take longer for the sequence to be ‘captured’ by a
smaller circle. This is illustrated in the figure below. The
sequence is strung along the curve shown heading
towards a. The bigger circle of radius 2 captures the
sequence by the time n = 47, the smaller circle doesn’t
capture it till n = 59. Note that z25 is inside the larger
circle, but since later points are outside the circle we
don’t say the sequence is captured at n = 25 2 ANALYTIC
FUNCTIONS 20 A sequence of points converging to a
Definition. The sequence z1, z2, z3, . . . converges to the
value a if for every > 0 there is a number Nsuch that |zn
− a| < for all n > N. We write this as limn→∞ zn = a.
Again, the definition just says that eventually the
sequence is within of a, no matter how small you choose
. Example 2.20. Show that the sequence zn = (1/n + i) 2
has limit -1. Solution: This is clear because 1/n → 0. For
practice, let’s phrase it in terms of epsilons: given > 0 we
have to choose Nsuch that |zn − (−1)| < for all n > NOne
strategy is to look at |zn + 1| and see what Nshould be.
We have |zn − (−1)| = 1 n + i 2 + 1 = 1 n2 + 2i n
< 1 n2 + 2 n So all we have to do is pick N large enough
that 1 N2 + 2 N < Since this can clearly be done we have
proved that zn → i. This was clearly more work than we
want to do for every limit. Fortunately, most of the time
we can apply general rules to determine a limit without
resorting to epsilons! Remarks. 1. In 18.04 we will be
able to spot the limit of most concrete examples of
sequences. The formal definition is needed when dealing
abstractly with sequences. 2. To mathematicians is one
of the go-to symbols for a small number. The prominent
and rather eccentric mathematician Paul Erdos used to
refer to children as epsilons, as in ‘How are the epsilons
doing?’ 3. The term ‘captured by the circle’ is not in
common usage, but it does capture what is happening. 2
ANALYTIC FUNCTIONS 21 2.11.2 limz→z0 f(z) Sometimes
we need limits of the form limz→z0 f(z) = a. Again, the
intuitive meaning is clear: as z gets close to z0 we should
see f(z) get close to a. Here is the technical definition
Definition. Suppose f(z) is defined on a punctured disk 0
< |z − z0| < r around z0. We say limz→z0 f(z) = a if for
every > 0 there is a δ such that |f(z) − a| < whenever 0 <
|z − z0| < δ This says exactly that as z gets closer (within
δ) to z0 we have f(z) is close (within ) to a. Since can be
made as small as we want, f(z) must go to a. Remarks. 1.
Using the punctured disk (also called a deleted
neighborhood) means that f(z) does not have to be
defined at z0 and, if it is then f(z0) does not necessarily
equal a. If f(z0) = a then we say the f is continuous at z0.
2. Ask any mathematician to complete the phrase “For
every ” and the odds are that they will respond “there is
a δ . . . ” 2.11.3
Connection between limits of sequences and limits of
functions
Here’s an equivalent way to define limits of functions:
the limit limz→z0 f(z) = a if, for every sequence of points
{zn} with limit z0 the sequence {f(zn)} has limit a.
CHAPTER 8
PERMUTATION AND
COMBINATION SERIES
ABSTRACT
The Permutation Series and Combination Series will ensure
the readers to understand proper calculation of Permutation
and Combination and will also help them to grew their
knowledge on the topic.
It is very helpful topic for the students to strengthen their
roots in mathematics. The magic that Mathematics have in it
will be exposed to the readers to understand their
Mathematics thoroughly and will also help them to enlarge
their thinking capabilities.
Hence, this topic is very interesting as it provides a link
between A.P, G.P with Permutation and Combination.
It will help to calculate many selections and others in a very
small time.
PERMUTATION SERIES AND COMBINATION

SERIES
(1) INTRODUCTION
Till date we have just used permutations and combinations in
arrangements, dearrangements, selection and others. But
here, Permutation and Combination is described in a new
form called Permutation series and Combination Series.
(2) IMPORTANCE OF PERMUTATION SERIES AND
COMBINATION SERIES
Just like A.P, G.P, H.P Permutation series and Combination
Series we also help us to find out the next term, previous
term, sum of Permutations and Combinations upto m terms.
It will also strengthen Permutations and Combinations in the
minds of the peoples
This will make a easy way to find larger terms in a very
short time interval. Hence, the Permutation series and
Combination series is largely important.
(3) PERMUTATION SERIES
The formula to find Permutation of n things when r things is
taken at a time is:-
n
pr= n!/(n-r)!-------------------------------------------------------------(1)
but Permutation series deals with more than 1 permutation
keeping n fixed.
Let us take an example:-
n
pr+ npr1+ npr2+……………………………….+npr
In Permutation and Combination Series the last highest term
is npn or ncn as n cannot be less than r.
(3.1) To find the next term when any term is given(except
the first term)
Let tf=npr and tf+1=npr+1
Tf/tf+1=n!/(n-r)!/n!/(n-r)!
Tf/tf+1=[n!*(n-r-1)!]/[(n-r)!*n!]
Tf/tf+1=(n-r-1)!/[(n-r)(n-r-1)!]
Tf/tf+1=1/(n-r)
Tf+1/tf=n-r
Hence,
Tf+1=(n-r)*tf ---------------------------------------------------------(2)
(3.2) To find the previous term when any term is
given(except the first term).
Tf=npr and tf-1=npr-1
Tf-1/tf=n!/(n-r+1)!/n!/(n-r)!
Tf-1/tf=[n!*(n-r)!]/[(n-r+1)(n-r)!*n!]
Tf-1/tf=1/(n-r+1)
Hence,
Tf-1=tf/(n-r+1) --------------------------------------------------------(3)
(3.3) To find any term when a term is given and common
difference is known(except the first term).
Suppose, ta1=npr1 and ta2=npr2
d=r2-r1
r2=d+r1
ta1/ta2=n!/(n-r1)!/n!/(n-r2)!
ta1/ta2=[n!*(n-r2)!]/[(n-r1)!*n!]
ta1/ta2=[n-(d+r1)]!/(n-r1)!
Hence,
Ta2=[(n-r1)!/(n-d-r1)!]*ta1 -------------------------------------------(4)
Where r1 can be any number less than n and d is the common
difference between r1, r2, r3 and so on.
(3.4) To find mth term in a Permutation Series.
Suppose, a Permutation Series be
n
p1+np2+np3+…………………………………….+npn
find the 3rd term.
Solution: To find the third term.
First let write the details
First term ar=1
Common difference=d=2-1=3-2=1
Hence,
Third term:-
T3=np1+(3-1)1
=np3
Hence,
The formula to find mth term in a Permutation Series
Tm=npa +(m-1)d ------------------------------------------------------------(5)
r
Where, ar is the first term, d is the common difference.

(3.5) To find sum of m terms.
Suppose, a series be
n
p1+np2+…………………………………………………………+npn
find the sum of first 3 terms:
solution: sum of first 3 terms
sm=n![1/(n-1)!+1/(n-2)!+1/(n-3)!]
hence,
In general, to find sum of m terms the formula
Sm=n![1/(n-r)!+1/(n-(r+d))!+……….+1/{n-r-(m-1)d}!] -----------(6)
Where sm means sum of the series.
(4) COMBINATION SERIES
The formula to find combination of n things when r things is
taken at a time is :-
n
cr=n!/[r!*(n-r)!] ------------------------------------------------------ (7)
Just like Permutation series, Combination Series will have
highest terms as ncr as r cannot be more than n keeping n
constant. For example:-
n
c1+nc2+nc3+………………………………………….+ncn
(4.1) To find the next term when any term is given(except the
first term).
Let tf=ncr and tf+1=ncr+1
Tf/tf+1=n!/[r!(n-r)!]/n!/[(r+1)!*(n-r-1)!]
Tf/tf+1=[n!*(r+1)!*(n-r-1)!]/[r!*(n-r)!*n!]
Tf/tf+1=[(r+1)(r!)*(n-r-1)!]/[r!*(n-r)(n-r-1)!]
Tf+1/tf=(n-r)/(r+1)
Hence,
Tf+1=[(n-r)/(r+1)]*tf ------------------------------------------------(8)
(4.2) To find the previous term when any term is given(except

the first term)
Let tf=ncr and tf+1=ncr-1
Tf-1/tf =n!/[(r-1)!*(n-r+1)!]/n!/[r!*(n-r)!]
Tf-1/tf =[n!*r!*(n-r)!]/[(r-1)!*(n-r+1)!*n!]
Tf-1/tf =[r(r-1)!*(n-r)!]/[(r-1)!*(n-r+1)(n-r)!]
Tf-1/tf =r/(n-r+1)
Hence,
Tf-1=[r/(n-r+1)]*tf ----------------------------------------------------(9)
(4.3) To find any term when a term is given and common
difference is known(except the first term).
Let ta=ncr1 and t’a =ncr2
r2=r1+d where d is common difference and r1 is any number less
than n.
t’a/ta =[n!*r!*(n-r1)!]/[r2!*(n-r2)!*n!]
hence,
t’a=[{r1!*(n-r1)!}/{(r1+d)!*(n-r1-d)!}]*ta ------------------------(10)
(4.4) To find mth term.
Let the first term be ncr
a=r
Just like Permutation Series, in Combination Series, to find the
mth term we use the formula:-
Tm=nca+(m-1)d ----------------------------------------------------------------(11)
Where d is common difference.
(4.5) To find sum of m terms.
Just like Permutation Series, if sm be the sum of m terms of a
permutation series and first term is ncr where a=r then the
formula:-
Sm=n![1/(r!(n-r)!)+1/((r+d)!(n-r-d)!)+....................…1/((r+(m-
1)d)*(n-r-(m-1)d)] --------------------------------------------------------
(12)
(4.6) Some special combination series with their formula for
sum:-
 nc0+nc1+nc2+……………..+ncn = 2n
 nc1+nc2+……………..+ncn = 2n-1
 nc0-nc2+nc4-nc6+…………= (2^0.5)n cos(n*180/4)
 nc1-nc3+nc5-nc7+…………= (2^0.5)n sin(n*180/4)
 nc0-nc1+nc2-nc3+……….+(-1)n ncn = 0
 nc1-2nc2+3nc3-……………….. = 0
 nc0+2nc2+3nc3+…………+(n+1)ncn = (n+2)2n-1
 nc0ncr+ nc1ncr+1+………….+ ncn-rncn = 2ncn-r=(2n)!/[(n-r)!*(n+r)!]
 nc02+nc12+nc22+……………..+ncn2 = (2n)!/(n!)2
 nc02-nc12+nc22-……………..+(-1)n ncn2={(-1)^n/2.nc if n is even0 if n is odd}
n/2
 nc12-2nc22+3nc32-……………..+(-1)n n.ncn2 =(-1)n/2-1n/2[(n!)/{(n/2)!

*(n/2)!}] when n is even
 nc0+nc1/2+nc2/3+……………..+ncn/(n+1) = (2n+1-1)/(n+1)
 nc0+nc1/2+nc2/22+……………..+ncn/2n = (3/2)n
 2n+1c0+2n+1c1+2n+1c2+…………..+2n+1cn = 22n
 ncn+n+1cn+n+2cn+……………..+2n-1cn = 2ncn+1
 (-1)r ncr{1/2r+3r/22r+7r/23r+15r/24r+…….+m terms}=(2mn-
1)/(2mn(2n-1))
REFERENCES:-
[1]Amit Rastogi,Mathematics
[2]Saha and Saha, ISC Mathematics xii
[3] Ml.Aggarwal, Understanding ISC Mathematics xii

Theory of Collisional Mechanics

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Theory of Collisional Mechanics

Загружено:

Авторское право:

Доступные форматы

CHAPTER 1

(1.1) RELATION BETWEEN RELATIVISTIC KINETIC

The gravitational force acting between them is

This means, the shorter the distance between

In this case r=0 and gravitational force is ∞ but

This is because of a hidden force is present in

This force is exacted outwards by the

Change in mass = ∇m=m -m

Momentum of light in medium1=p1=m1v1

The theory of Electromagnetic mass also provides us

The laws of collision will help us to learn the basics of

The particle A is called the incident particle particle and

(3)THEORY OF ELECTROMAGNETIC MASS

(4) LAWS OF COLLISION

 If the mass and velocity of the inclined particle is

(4.1) ELASTIC COLLISION IN TWO DIMENSIONS

WAVE BOUNDARY LAYERS 129

130 COASTAL ENGINEERING

BOUNDARY LAYERS 131

MEASUREMENT OF ENERGY LOSS

134 COASTAL ENGINEERING

ROUGH TURBULENT CASE

136 COASTAL ENGINEERING

WAVE BOUNDARY LAYERS 137

138 COASTAL ENGINEERING

SMOOTH TURBULENT CASE

WAVE BOUNDARY LAYERS 139

140 COASTAL ENGINEERING

Proof : The first part of the lemma has been established in

Since k0 = 1, Corollary 3.2 allows us to find kj recursively in j.

Corollary 3.2 is a new result which, together with Lemma 3.4,

PERMUTATION SERIES AND COMBINATION

Where, ar is the first term, d is the common difference.

(4.2) To find the previous term when any term is given(except

 nc12-2nc22+3nc32-……………..+(-1)n n.ncn2 =(-1)n/2-1n/2[(n!)/{(n/2)!

Вам также может понравиться