Mso203b PDF

Partial Differential Equations
T. Muthukumar
tmk@iitk.ac.in
November 13, 2015

ii
Contents
Notations vii
1 Introduction 1
1.1 Multi-Index Notations . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Classification of PDE . . . . . . . . . . . . . . . . . . . . . . . 4
2 Introduction Continued... 7
2.1 Solution of PDE . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Well-posedness of PDE . . . . . . . . . . . . . . . . . . . . . . 9
3 First Order PDE 11

3.1 Linear Transport Equation . . . . . . . . . . . . . . . . . . . . 11
3.2 Method of Characteristics . . . . . . . . . . . . . . . . . . . . 14
4 Method of Characteristics: Continued... 17
5 Classification of Second Order PDE 23

5.1 Semilinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
6 Classification of SOPDE: Continued 29

6.1 Quasilinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6.2 Why Characteristic Curves? . . . . . . . . . . . . . . . . . . . 29
6.3 Cauchy Boundary Condition . . . . . . . . . . . . . . . . . . . 31
7 Classification of SOPDE: Continued 33

7.1 Invariance of Discriminant . . . . . . . . . . . . . . . . . . . . 33
7.2 Standard or Canonical Forms . . . . . . . . . . . . . . . . . . 34
7.3 Reduction to Standard Form . . . . . . . . . . . . . . . . . . . 35
iii
CONTENTS iv
8 The Laplacian 41
8.1 Properties of Laplacian . . . . . . . . . . . . . . . . . . . . . . 41
8.2 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . 43
8.3 Harmonic Functions . . . . . . . . . . . . . . . . . . . . . . . 45
9 Properties of Harmonic Functions 47

9.0.1 Existence and Uniqueness of Solution . . . . . . . . . . 48
10 Sturm-Liouville Problems 51
10.1 Eigen Value Problems . . . . . . . . . . . . . . . . . . . . . . 51
10.2 Sturm-Liouville Problems . . . . . . . . . . . . . . . . . . . . 52
11 Spectral Results 55
12 Singular Sturm-Liouville Problem 59

12.0.1 EVP of Legendre Operator . . . . . . . . . . . . . . . . 59
12.0.2 EVP of Bessel’s Operator . . . . . . . . . . . . . . . . 61
13 Orthogonality of Eigen Functions 63

13.1 Eigen Function Expansion . . . . . . . . . . . . . . . . . . . . 67
14 Fourier Series 69
14.1 Periodic Functions . . . . . . . . . . . . . . . . . . . . . . . . 69
14.2 Fourier Coefficients and Fourier Series . . . . . . . . . . . . . 70
15 Fourier Series: Continued... 77

15.1 Piecewise Smooth Functions . . . . . . . . . . . . . . . . . . . 77
15.2 Complex Fourier Coefficients . . . . . . . . . . . . . . . . . . . 78
15.3 Orthogonality . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
15.3.1 Odd and Even functions . . . . . . . . . . . . . . . . . 82
15.4 Fourier Sine-Cosine Series . . . . . . . . . . . . . . . . . . . . 82
15.5 Fourier Transform and Integral . . . . . . . . . . . . . . . . . 84
16 Standing Waves: Separation of Variable 87

16.1 Elliptic Equations . . . . . . . . . . . . . . . . . . . . . . . . . 91
17 Parabolic: Heat Equation 99

17.1 Inhomogeneous Equation . . . . . . . . . . . . . . . . . . . . . 102
CONTENTS v
18 Travelling Waves 105

18.1 Domain of Dependence and Influence . . . . . . . . . . . . . . 108
Appendices 109
A Divergence Theorem 111
B Normal Vector of a Surface 113
C Duhamel’s Principle 115
Bibliography 117
Index 119
CONTENTS vi
Notations
Symbols
N denotes the set of natural numbers
Ω denotes an open subset of Rn , not necessarily bounded
∂Ω denotes the boundary of Ω
R denotes the set of real numbers
Rn denotes the n-dimensional Euclidean space

Pn ∂ 2
∆ i=1 ∂x2i
∂ α1 αn
Dα . . . ∂x∂n αn and α = (α1 , . . . , αn ). In particular, for α = (1, 1, . . . , 1),
∂x1 α1
D = ∇ = D(1,1,,...,1) = ∂x∂ 1 , ∂x∂ 2 , . . . , ∂x∂n
Function Spaces
C(X) is the class of all continuous functions on X
C k (X) is the class of C k functions which admit a continuous extension to

the boundary of X
C k (X) is the class of all k-times (k ≥ 1) continuously differentiable functions

on X
C ∞ (X) is the class of all infinitely differentiable functions on X
C j,k (X × Y ) is the class of all j-times (j ≥ 0) continuously differentiable

functions on X and k-times (k ≥ 0) continuously differentiable func-
tions on Y
vii
NOTATIONS viii
Cc∞ (X) is the class of all infinitely differentiable functions on X with com-
pact support
General Conventions
∇x , ∆x or Dx2 When a PDE involves both the space variable x and time vari-
able t, the quantities like ∇, ∆, D2 , etc. are always taken with respect
to the space variable x only. This is a standard convention. Some-
times the suffix, like ∇x or ∆x , is used to indicate that the operation
is taken w.r.t x.
BVP Boundary value problem
IVP Initial value problem
w.r.t with respect to

Lecture 1
Introduction
A partial differential equation (PDE) is an equation involving an unknown

function u of two or more variables and some or all of its partial derivatives.
The partial differential equation is a tool to analyse the models of nature. The
process of understanding natural system can be divided in to three stages:
(i) Modelling the problem or deriving the mathematical equation (formu-

lating a PDE) describing the natural system. The derivation process is
a result of physical laws such as Newton’s law, momentum, conservation
laws, balancing forces etc.
(ii) Solving the equation (PDE). What constitutes as a solution to a PDE?
(iii) Studying the properties of a solution. Most often the solution of a PDE
may not have a nice formula or representation. How much information
about the solution can one extract without any representation of a
solution? In this text, one encounters similar situation while studying
harmonic functions.
1.1 Multi-Index Notations

Let Ω be an open subset of R. Recall that the derivative of a function
u : Ω → R, at x ∈ Ω, is defined as
u(x + h) − u(x)
u0 (x) := lim
h→0 h
1
LECTURE 1. INTRODUCTION 2
provided the limit exists. Now, let Ω be an open subset of Rn . The directional
derivative of u : Ω → R, at x ∈ Ω and in the direction of a given vector
ξ ∈ Rn , is defined as
∂u u(x + hξ) − u(x)

(x) := lim
∂ξ h→0 h
provided the limit exists. Let the n-tuple ei := (0, 0, . . . , 1, 0, . . . , 0), where 1
is in the i-th place, denote the standard basis vectors of Rn . The i-th partial
derivative of u at x is the directional derivative of u, at x ∈ Ω and along the
direction ei , and is denoted as
∂u u(x + hei ) − u(x)

uxi (x) = (x) = lim .
∂xi h→0 h
A multi-index α = (α1 , . . . , αn ) is a n-tuple where αi , for each 1 ≤ i ≤

n, is a non-negative integer. Let |α| := α1 + . . . + αn . If α and β are
two multi-indices, then α ≤ β means αi ≤ βi , for all 1 ≤ i ≤ n, and
α ± β = (α1 ± β1 , . . . , αn ± βn ). Also, α! = α1 ! . . . αn ! and, for any x ∈ Rn ,
xα = xα1 1 . . . xαnn . The multi-index notation, introduced by L. Schwartz, is
quite handy in representing multi-variable equations in a concise form. For
instance, a k-degree polynomial in n-variables can be written as
X
aα xα .
|α|≤k
The partial differential operator of order α is denoted as
∂ α1 ∂ αn ∂ |α|
Dα = . . . = .
∂x1 α1 ∂xn αn ∂x1 α1 . . . ∂xn αn
One adopts the convention that, among the similar components of α, the
order in which differentiation is performed is irrelevant. This is not a re-
strictive convention because the independence of order of differentiation is
valid for smooth1 functions. For instance, if α = (1, 1, 2) then one adopts the
convention that
∂4 ∂4
= .
∂x1 ∂x2 ∂x3 2 ∂x2 ∂x1 ∂x3 2
1
smooth, usually, refers to as much differentiability as required.
For each k ∈ N, Dk u(x) := {Dα u(x) | |α| = k}. The k = 1 case is
D1 u(x) = D(1,0,...,0) u(x), D(0,1,0,...,0) u(x), . . . , D(0,0,...,0,1) u(x)

∂u(x) ∂u(x) ∂u(x)
= , ,..., .
∂x1 ∂x2 ∂xn
The operator D1 is called the gradient operator and is denoted as D or ∇.

Thus, ∇u(x) = (ux1 (x), ux2 (x), . . . , uxn (x)). The directional derivative along
a vector ξ ∈ Rn satisfies the identity
∂u
(x) = ∇u(x) · ξ.
∂ξ
The normal derivative is the directional derivative along the normal direction
ν(x), at x, with respect to the surface in which x lies. The divergence of a
vector function u = (u1 , . . . , un ), denoted as div(u), is defined as div(u) :=
∇ · u. The k = 2 case is
 ∂ 2 u(x) ∂ 2 u(x)

∂x1 2 . . . ∂x1 ∂xn
 ∂ 2 u(x) ∂ 2 u(x) 
2
 ∂x2 ∂x1
. . . ∂x 
2 ∂xn 
D u(x) =  
.. .. .. .
. . .

 
∂ 2 u(x) ∂ 2 u(x)
∂xn ∂x1
... ∂x2n n×n
The matrix D2 u is called the Hessian matrix. Observe that the Hessian
matrix is symmetric due to the independence hypothesis of the order in
which partial derivatives are taken. The Laplace operator,
P denoted as ∆, is
∂2
defined as the trace of the Hessian operator, i.e., ∆ := ni=1 ∂x2 . Note that
i
∆ = ∇ · ∇. Further, for a k-times differentiable function u, the nk -tensor
k
Dk u(x) := {Dα u(x) | |α| = k} may be viewed as a map from Rn to Rn .
Thus, the magnitude of Dk u(x) is
  12
X
|Dk u(x)| :=  |Dα u(x)|2  .
|α|=k
1
In particular, |∇u(x)| = ( ni=1 u2xi (x)) 2 or |∇u(x)|2 = ∇u(x) · ∇u(x) and
P
1
|D2 u(x)| = ( ni,j=1 u2xi xj (x)) 2 .
P
Example 1.1. Let u(x, y) : R2 → R be defined as u(x, y) = ax2 + by 2 . Then

∇u = (ux , uy ) = (2ax, 2by)
and
2 uxx uyx 2a 0
D u= = .
uxy uyy 0 2b
2
Observe that ∇u : R2 → R2 and D2 u : R2 → R4 = R2 .
1.2 Classification of PDE

Definition 1.2.1. Let Ω be an open subset of Rn . A k-th order partial
differential equation of an unknown function u : Ω → R is of the form
F Dk u(x), Dk−1 u(x), . . . Du(x), u(x), x = 0,

(1.2.1)
k k−1
for each x ∈ Ω, where F : Rn × Rn × . . . × Rn × R × Ω → R is a given
map such that F depends, at least, on one k-th partial derivative u and is
independent of (k + j)-th partial derivatives of u for all j ∈ N.
In short, the order of a PDE is the highest partial derivative order that
occurs in the PDE. A first order PDE with two unknown variable (x, y) is
represented as F (ux , uy , u, x, y) = 0 with F depending, at least, on one of ux
and uy . Similarly, a first order PDE with three variable unknown function
u(x, y, z) is written as F (ux , uy , uz , u, x, y, z) = 0 with F depending, at least,
on one of ux , uy and uz . Note the abuse in the usage of x. In the n-variable
case, x ∈ Rn is a vector. In the two and three variable case x is the first
component of the vector. The usage should be clear from the context.
A PDE is a mathematical description of a physical process and solving
a PDE for an unknown u helps in predicting the behaviour of the physical
process. The level of difficulty in solving a PDE may depend on its order k
and linearity of F . We begin by classifying PDEs in a scale of linearity.
Definition 1.2.2. (i) A k-th order PDE is linear if F in (1.2.1) has the
form
G(u(x)) = f (x)
where G(u(x)) := |α|≤k aα (x)Dα u(x) for given functions f and aα ’s.
P
It is called linear because G is linear in u for all derivatives , i.e.,
G(λu1 + µu2 ) = λG(u1 ) + µG(u2 ) for λ, µ ∈ R. In addition, if f ≡ 0
then the PDE is linear and homogeneous.
(ii) A k-th order PDE is semilinear if F is linear only in the highest (k-th)
order, i.e., F has the form
X
aα (x)Dα u(x) + f (Dk−1 u(x), . . . , Du(x), u(x), x) = 0.
|α|=k
(iii) A k-th order PDE is quasilinear if F has the form

X
aα (Dk−1 u(x), . . . , u(x), x)Dα u + f (Dk−1 u(x), . . . , u(x), x) = 0,
|α|=k
i.e., the coefficient of its highest (k-th) order derivative depends on u

and its derivative only upto the previous (k − 1)-th orders.
(iv) A k-th order PDE is fully nonlinear if it depends nonlinearly on the

highest (k-th) order derivatives.
Observe that, for a semilinear PDE, f is never linear in u and its deriva-
tives, otherwise it reduces to being linear. For a quasilinear PDE, aα (with
|α| = k), cannot be independent of u or its derivatives, otherwise it reduces
to being semilinear or linear.
Example 1.2. (i) xuy − yux = u is linear.
(ii) xux + yuy = x2 + y 2 is linear.
(iii) utt − c2 uxx = f (x, t) is linear.
(iv) y 2 uxx + xuyy = 0 is linear.
(v) ux + uy − u2 = 0 is semilinear.
(vi) ut + uux + uxxx = 0 is semilinear.
(vii) u2tt + uxxxx = 0 is semilinear.
(viii) ux + uuy − u2 = 0 is quasilinear.
(ix) uux + uy = 2 is quasilinear.
(x) ux uy − u = 0 is nonlinear.
Lecture 2
Introduction Continued...
2.1 Solution of PDE

Definition 2.1.1. We say u : Ω → R is a (classical) solution to the k-th
order PDE (1.2.1),
• if Dα u exists for all |α| explicitly present in (1.2.1);
• and u satisfies the equation (1.2.1).
Example 2.1. Consider the first order equation ux (x, y) = 0 in R2 . Freezing

the y-variable, the PDE can be viewed as an ODE in x-variable. On integrat-
ing both sides w.r.t x, u(x, y) = f (y) for any arbitrary function f : R → R.
Therefore, for every choice of f : R → R, there is a solution u of the PDE.
Note that the solution u is not necessarily in C 1 (R2 ) to solve the first order
PDE as is the case in solving an ODE. In fact, choosing a discontinuous
function f , we obtain a solution which is discontinuous in the y-direction.
Similarly, the solution of uy (x, y) = 0 is u(x, y) = f (x) for any choice of
f : R → R. We say u is a classical solution if u satisfies the equation and
u ∈ C 1,0 (R × R).
Example 2.2. Consider the first order equation ut (x, t) = u(x, t) in R×(0, ∞)
such that u(x, t) 6= 0, for all (x, t). Freezing the x-variable, the PDE can be
viewed as an ODE in t-variable. Integrating both sides w.r.t t we obtain
u(x, t) = f (x)et , for some arbitrary function f : R → R.
Example 2.3. Consider the second order PDE uxy (x, y) = 0 in R2 . In contrast
to the previous two examples, the PDE involves derivatives in both variables.
7
LECTURE 2. INTRODUCTION CONTINUED... 8
On integrating both sides w.r.t x we obtain uy (x, y) = F (y), for any arbitrary
integrable function F : R → R. Now, integrating both sides w.r.t y, u(x, y) =
f (y)+g(x) for an arbitrary g : R → R and a f ∈ C 1 (R)1 . But the u obtained
above is not a solution to uyx (x, y) = 0 if g is not differentiable. Since we
assume mixed derivatives to be same we need to assume f, g ∈ C 1 (R) for the
solution to exist.
Example 2.4. Consider the first order equation ux (x, y) = uy (x, y) in R2 .
On first glance, the PDE does not seem simple to solve. But, by change
of coordinates, the PDE can be rewritten in a simpler form. Choose the
coordinates w = x + y and z = x − y and, by chain rule, ux = uw + uz
and uy = uw − uz . In the new coordinates, the PDE becomes uz (w, z) = 0
which is in the form considered in Example 2.1. Therefore, its solution is
u(w, z) = f (w) for any arbitrary f : R → R and, hence, u(x, y) = f (x + y).
But now an arbitrary f cannot be a solution. We impose that f ∈ C 1 (R).
The family of solutions, obtained in the above examples, may not be
the only family that solves the given PDE. Following example illustrates a
situation where three different family of solutions exist (more may exist too)
for the same PDE.
Example 2.5. Consider the second order PDE ut (x, t) = uxx (x, t).
(i) Note that u(x, t) = c is a solution of the PDE, for any constant c ∈ R.
This is a family of solutions indexed by c ∈ R.
2
(ii) The function u : R2 → R defined as u(x, t) = x2 +t+c, for any constant
c ∈ R, is also a family of solutions of the PDE. Because ut = 1, ux = x
and uxx = 1. This family is not covered in the first case.
(iii) The function u(x, t) = ec(x+ct) is also a family of solutions to the PDE,
for each c ∈ R. Because ut = c2 u, ux = cu and uxx = c2 u. This family
is not covered in the previous two cases.
Recall that the family of solutions of an ODE is indexed by constants. In

contrast to ODE, observe that the family of solutions of a PDE is indexed
by either functions or constants.
1
In fact, it is enough to assume f is differentiable a.e. which is beyond the scope of
this text
2.2 Well-posedness of PDE

It has been illustrated via examples that a PDE has a family of solutions.
The choice of one solution from the family of solutions is made by imposing
boundary conditions (boundary value problem) or initial conditions (initial
value problem). If too many initial/boundary conditions are specified, then
the PDE may have no solution. If too few initial/boundary conditions are
specified, then the PDE may have many solutions. Even with right amount
of initial/boundary conditions, but at wrong places, the solution may fail
to be stable, i.e., may not depend continuously on the initial or boundary
data. It is, usually, desirable to solve a well-posed problem, in the sense of
Hadamard. A PDE, along with the boundary condition or initial condition,
is said to be well-posedness if the PDE
(a) has a solution (existence);
(b) the solution is unique (uniqueness);
(c) and the solution depends continuously on the data given (stability).
Any PDE not meeting the above criteria is said to be ill-posed. If the PDE
(with boundary/initial conditions) is viewed as a map then the well-posedness
of the PDE is expressed in terms of the surjectivity, injectivity and continuity
of the “inverse” map. The existence and uniqueness condition depends on the
notion of solution in consideration. There are three notions of solution, viz.,
classical solutions, weak solutions and strong solutions. This textbook, for
the most part, is in the classical situation. Further, the stability condition
means that a small “change” in the data reflects a small “change” in the
solution. The change is measured using a metric or “distance” in the function
space of data and solution, respectively. Though in this text we study only
well-posed problems there are ill-posed problems which are also of interest.
The following example illustrates the idea of continuous dependence of
solution on data in the uniform metric on the space of continuous functions.
Example 2.6. The initial value problem (IVP)

utt (x, t) = uxx (x, t) in R × (0, ∞)
u(x, 0) = ut (x, 0) = 0
has the trivial solution u(x, t) = 0. Consider the IVP with a small change in
data, 
 utt (x, t) = uxx (x, t) in R × (0, ∞)
u(x, 0) = 0
ut (x, 0) = ε sin xε

which has the unique2 solution uε (x, t) = ε2 sin(x/ε) sin(t/ε). The change in
solution of the IVP is measured using the uniform metric as
sup{|uε (x, t) − u(x, t)|} = ε2 sup {|sin(x/ε) sin(t/ε)|} = ε2 .

(x,t) (x,t)
Thus, a small change in data induces a small enough change in solution under
the uniform metric3 .
Example 2.7 (Ill-posed). The IVP

utt (x, t) = −uxx (x, t) in R × (0, ∞)
u(x, 0) = ut (x, 0) =0
has the trivial solution u(x, t) = 0. Consider the IVP with a small change in
data, 
 utt (x, t) = −uxx (x, t) in R × (0, ∞)
u(x, 0) = 0
ut (x, 0) = ε sin xε

which has the unique solution uε (x, t) = ε2 sin(x/ε) sinh(t/ε). The solution
of the IVP is not stable because the data change is small, i.e.,
sup{|uεt (x, 0) − ut (x, 0)|} = ε sup {|sin(x/ε)|} = ε

x x
and the solution change is not at all small, i.e.,
lim sup{|uε (x, t) − u(x, t)|} = lim ε2 |sinh(t/ε)| = +∞.

t→∞ x t→∞
In fact, the solution will not converge in any reasonable metric.
2
This claim will be proved in later chapters.
3
The space R × (0, ∞) is not compact and the metric is not complete. The example is
only to explain the notion of stability at an elementarty level.
Lecture 3
First Order PDE
The aim of this chapter is to find the general solution and to solve the Cauchy
problem associated with the first order PDE of the form
F (∇u(x), u(x), x) = 0 x ∈ Rn .
3.1 Linear Transport Equation

The transport of a substance in a fluid flowing (one dimenisonal flow) with
constant speed b, with neither source or sink of substance, is given by
ut (x, t) + bux (x, t) + du(x, t) = cuxx (x, t) (x, t) ∈ R × (0, ∞)
where c is the diffusive coefficient of the substance and d is the rate of decay
of the substance. The function u(x, t) denotes the concentration/density of
the substance at (x, t). Note that the case of no diffusion (c = 0) is a linear
first order equation which will be studied in this section.
Example 3.1 (One Dimension with no decay). The one space dimension trans-
port equation is
ut (x, t) + bux (x, t) = 0; (x, t) ∈ R × (0, ∞)
with b ∈ R. The c = d = 0 case describes the transport of an insoluble1

substance O in a fluid flowing with constant speed b. To solve this we consider
two observers, one a fixed observer A and another observer B, moving with
1
assuming no diffusion and no decay of substance.
11
LECTURE 3. FIRST ORDER PDE 12
Figure 3.1: Transport of initial data g
speed b and in the same direction as the substance O. For B, the substance
O would appear stationary while for A, the fixed observer, the substance O
would appear to travel with speed b. What is the equation of the transport
of the “stationary” substance O from the viewpoint of the moving observer
B? The answer to this question lies in identifying the coordinate system
for B relative to A. Fix a point x at time t = 0. After time t, the point
x remains as x for the fixed observer A, while for the moving observer B,
the point x is now x − bt. Therefore, the coordinate system for B is (w, z)
where w = x − bt and z = t. Let v(w, z) describe the motion of O from B’s
perspective. Since B sees O as stationary, the PDE describing the motion
of O is vz (w, z) = 0. Therefore, v(w, z) = g(w), for some arbitrary function
g (sufficiently differentiable), is the solution from B’s perspective. To solve
the problem from A’s perspective, note that
ut = vw wt + vz zt = −bvw + vz and
ux = vw wx + vz zx = vw .
Therefore, ut + bux = −bvw + vz + bvw = vz and, hence, u(x, t) = v(w, z) =
g(w) = g(x − bt) (cf. Fig 3.1). The choice of g is based on our restriction
to be in a classical solution set-up. Note that, for any choice of g, we have
Figure 3.2: Path of substance, over time, placed at x0 with b > 0
g(x) = u(x, 0). The line x − bt = x0 , for some constant x0 , in the xt-plane
tracks the flow of the substance placed at x0 at time t = 0 (cf. Fig 3.2). Also,
observe that 0 = ut + bux = (ux , ut ) · (b, 1) is precisely that the directional
derivative along the vector (b, 1) is zero. This means that u is constant if
we move along the direction (b, 1). Thus, the value of u(x, t) on the line
x − bt = x0 is constant.
Example 3.2 (Transport equation in first quadrant). The transport equation
is
ut (x, t) + bux (x, t) = 0; (x, t) ∈ (0, ∞) × (0, ∞)
with b ∈ R. As before, we obtain u(x, t) = g(x − bt) where g(x) = u(x, 0).
This problem is uniquely solvable in the given region only for b < 0. For
b > 0, g defined on x-axis is not adequate to solve the problem. The problem
is not well-posed! The given data is enough only to solve for u in the region
{(x, t) ∈ (0, ∞) × (0, ∞) | x > bt} when b > 0. To compute u in {(x, t) ∈
(0, ∞) × (0, ∞) | x > bt} we need to provide data on the t-axis (0, t).
Example 3.3 (Transport equation in semi-infinite strip). The transport equa-
tion is
ut (x, t) + bux (x, t) = 0; (x, t) ∈ (0, L) × (0, ∞)
with b ∈ R. As before, we obtain u(x, t) = g(x − bt) where g(x) = u(x, 0). If
b > 0 then the problem is well-posed when the data given on x and t axes.
If b < 0 then the problem is well-posed when the data is given on x-axis and
(L, t)-axis.
3.2 Method of Characteristics

The method of characteristics gives the equation, a system of ODE, of the
characteristic curves. The method of characteristics reduces a first order
PDE to a system of ODE. For illustration, consider the two variable first
order quasi-linear equation:
A(x, y, u)ux + B(x, y, u)uy = C(x, y, u). (3.2.1)
Solving for u(x, y) in the above equation is equivalent to finding the surface
S ≡ {(x, y, u(x, y))} generated by u in R3 . If u is a solution of (3.2.1), at
each (x, y) in the domain of u, then
A(x, y, u)ux + B(x, y, u)uy = C(x, y, u)

A(x, y, u)ux + B(x, y, u)uy − C(x, y, u) = 0
(A(x, y, u), B(x, y, u), C(x, y, u)) · (ux , uy , −1) = 0
(A(x, y, u), B(x, y, u), C(x, y, u)) · (∇u(x, y), −1) = 0.
But (∇u(x, y), −1) is normal to S at the point (x, y) (cf. Appendix B).
Hence, the coefficients (A(x, y, u), B(x, y, u), C(x, y, u)) are perpendicular to
the normal and, therefore, (A(x, y, u), B(x, y, u), C(x, y, u)) lie on the tangent
plane to S at (x, y, u(x, y)).
Definition 3.2.1. A smooth curve in Rn is said to be an integral or charac-

teristic curve w.r.t a given vector field, if the vector field is tangential to the
curve at each of its point.
Definition 3.2.2. A smooth surface in Rn is said to be an integral surface

w.r.t a given vector field, if the vector field is tangential to the surface at each
of its point.
In the spirit of above definition and arguments, finding a solution u to

(3.2.1) is equivalent to determining an integral surface S corresponding to the
coefficient vector field V (x, y) = (A(x, y, u), B(x, y, u), C(x, y, u)) of (3.2.1).
Let s denote the parametrization of the characteristic curves w.r.t V . For
convenience, let z(s) := u(x(s), y(s)). Then the characteristic curves can be
found by solving the system of ODEs
dx dy dz
= A(x(s), y(s), z(s)), = B(x(s), y(s), z(s)), = C(x(s), y(s), z(s)).
ds ds ds
(3.2.2)
The three ODE’s obtained are called characteristic equations. The union of
these characteristic (integral) curves give us the integral surface S. The union
is in the sense that every point in the integral surface belongs to exactly one
characteristic.
Example 3.4 (Linear Transport Equation). The linear transport equation is
already solved earlier using elementary method. Let us solve the same using
the method of characteristics. Consider the linear transport equation in two
variable,
ut + bux = 0, x ∈ R and t ∈ (0, ∞),
where the constant b ∈ R is given. Thus, the given vector field V (x, t) =
(b, 1, 0). The characteristic equations are
dx dt dz
= b, = 1, and = 0.
ds ds ds
Solving the 3 ODE’s, we get
x(s) = bs + c1 , t(s) = s + c2 , and z(s) = c3 .
Note that solving the system of ODEs requires some initial condition. We
have already observe that the solution of the transport equation depended
on the value of u at time t = 0, i.e., the value of u on the curve (x, 0) in
the xt-plane. Thus, the problem of finding a function u solving a first order
PDE such that u is known on a curve Γ in the xy-plane is called the Cauchy
problem.
Example 3.5. Let g be given (smooth enough) function g : R → R. Consider
the linear transport equation

ut + bux = 0 x ∈ R and t ∈ (0, ∞)
(3.2.3)
u(x, 0) = g(x) x ∈ R.
We parametrize the curve Γ with r-variable, i.e., Γ = {γ1 (r), γ2 (r)} =

{(r, 0)}. The characteristic equations are:
dx(r, s) dt(r, s) dz(r, s)

= b, = 1, and =0
ds ds ds
with initial conditions,
x(r, 0) = r, t(r, 0) = 0, and z(r, 0) = g(r).
Solving the ODE’s, we get
x(r, s) = bs + c1 (r), t(r, s) = s + c2 (r)
and z(r, s) = c3 (r) with initial conditions
x(r, 0) = c1 (r) = r
t(r, 0) = c2 (r) = 0, and z(r, 0) = c3 (r) = g(r).

Therefore,
x(r, s) = bs + r, t(r, s) = s, and z(r, s) = g(r).
The idea is to solve for r and s in terms of x, t. Let us set u(x, t) =

z(r(x, t), s(x, t)). In this case we can solve for r and s in terms of x and
t, to get
r(x, t) = x − bt and s(x, t) = t.
Therefore, u(x, t) = z(r, s) = g(r) = g(x − bt).
The above example shows how the information on data curve Γ is reduced
as initial condition for the characteristic ODEs.
Lecture 4
Method of Characteristics:
Continued...
Let us study the Example 3.5 with a different data curve Γ.

Example 4.1. Consider the linear transport equation

ut + bux = 0 x ∈ R and t ∈ (0, ∞)
(4.0.1)
u(bt, t) = g(t) t ∈ (0, ∞).
We parametrize the data curve Γ with r-variable, i.e., Γ = {γ1 (r), γ2 (r)} =
{(br, r)}. The characteristic equations are:

= b, = 1, and =0
ds ds ds
x(r, 0) = br, t(r, 0) = r, and z(r, 0) = g(r).
x(r, s) = bs + c1 (r), t(r, s) = s + c2 (r)
and z(r, s) = c3 (r) with initial conditions
x(r, 0) = c1 (r) = br
t(r, 0) = c2 (r) = r, and z(r, 0) = c3 (r) = g(r).
17
LECTURE 4. METHOD OF CHARACTERISTICS: CONTINUED... 18
Therefore,
x(r, s) = b(s + r), t(r, s) = s + r, and z(r, s) = g(r).
Note that in this case we cannot solve for r and s in terms of x, t.

The above examples suggest that solving a Cauchy problem depends
on the curve on which u is prescribed. One way is to prescribe values of
u on a data curve Γ parametrized as (γ1 (r), γ2 (r)), for r ∈ I ⊆ R, such
that all the characteristic curves (parametrised with s variable) intersect the
(γ1 (r), γ2 (r), u(γ1 , γ2 )) at s = 0. Let u(γ1 (r), γ2 (r)) = g(r). Thus, solving the
Cauchy problem is now reduced to solving for x(r, s), y(r, s), z(r, s) in (3.2.2)
with the initial conditions x(r, 0) = γ1 (r), y(r, 0) = γ2 (r) and z(r, 0) = g(r),
r ∈ I. The system of ODE can be solved uniquely for x(r, s), y(r, s) and
z(r, s), in a neighbourhood of s = 0 and all r ∈ I. At this juncture, one may
ask the following questions:
(i) Can the three solutions be used to define a function z = u(x, y)?
(ii) If yes to (i), then is the solution z = u(x, y) unique for the Cauchy
problem? The answer is yes because two integral surface intersecting
in Γ must contain the same charaterictics (beyond the scope of this
course).
Let us answer (i) in a neighbourhood of (r, 0). Set x(r, 0) = γ1 (r) = x0 ,

y(r, 0) = γ2 (r) = y0 and z(r, 0) = g(r) = z0 . Note that the answer to
(i) is in affirmation if we can solve for r and s in terms of x and y, i.e.,
r = R(x, y) and s = S(x, y) such that R(x0 , y0 ) = 0 and S(x0 , y0 ) = 0. Then
z = z(R(x, y), S(x, y)) = u(x, y). The inverse function theorem tells us that
r and s can be solved in terms of x and y in a neighbourhood of (x0 , y0 ) if
0 0

xr (r, 0) yr (r, 0) γ1 (r) γ2 (r)
J(r, 0) = = 6= 0.
xs (r, 0) ys (r, 0) A(x0 , y0 , z0 ) B(x0 , y0 , z0 )
The quantity J(r, 0) 6= 0 means that the vectors (A(x0 , y0 , z0 ), B(x0 , y0 , z0 ))

and (γ10 (r), γ20 (r)) are not parallel.
What happens in the case of J(r, 0) = 0, i.e., when the associated vectors
are parallel? The condition given on Γ is u(γ1 (r), γ2 (r)) = g(r). If u is a C 1
solution to the Cauchy problem then on differentiation, w.r.t r, the Cauchy
condition yields g 0 (r) = ux (γ1 (r), γ2 (r))γ10 (r) + uy (γ1 (r), γ2 (r))γ20 (r). Since u
is solution, at (x0 , y0 , z0 ), of the algebraic system

A(x0 , y0 , z0 ) B(x0 , y0 , z0 ) ux (x0 , y0 ) C(x0 , y0 , z0 )
=
γ10 (r) γ20 (r) uy (x0 , y0 ) g 0 (r)
a necessary condition is that

A(x0 , y0 , z0 ) B(x0 , y0 , z0 ) C(x0 , y0 , z0 )
rank = 1.
γ10 (r) γ20 (r) g 0 (r)
If the above rank condition is satisfied then the data curve Γ is said to be a
characteristic at (x0 , y0 , z0 ). Thus, we have the following possibilities:
(a) J(r, 0) 6= 0 for all r ∈ I. Note that, when J(r, 0) 6= 0, then the rank
condition is not satisfied1 and, hence, the data curve Γ does not have
any characteristic points (Γ is not parallel at all points). Then, in a
neighborhood of Γ, there exists a unique solution u = u(x, y) of the
Cauchy problem given by the system of ODEs.
(b) J(r0 , 0) = 0, for some r0 ∈ I, and Γ is characteristic at the point P0 =
(γ1 (r0 ), γ2 (r0 ), g(r0 )). Then a C 1 solution may exist in a neighborhood
of P0 .
(c) J(r0 , 0) = 0 for some r0 ∈ I and Γ is not characteristic at P0 . There are
no C 1 solutions in a neighborhood of P0 . There may exist less regular
solutions.
(d) If Γ is a characteristic then there exists infinitely many C 1 solutions in
a neighborhood of Γ.
Example 4.2. Consider the Burgers’ equation given as

ut + uux = 0 x ∈ R and t ∈ (0, ∞)
u(x, 0) = g(x) x ∈ R.
The parametrization of the curve Γ with r-variable, i.e., Γ = {γ1 (r), γ2 (r)} =
{(r, 0)}. The characteristic equations are:
= z, = 1, and =0
ds ds ds
1
because rank is 2 in this case
x(r, 0) = r, t(r, 0) = 0, and z(r, 0) = g(r).
Solving the ODE corresponding to z, we get z(r, s) = c3 (r) with initial

conditions z(r, 0) = c3 (r) = g(r). Thus, z(r, s) = g(r). Using this in the
ODE of x, we get
dx(r, s)
= g(r).
ds
x(r, s) = g(r)s + c1 (r), t(r, s) = s + c2 (r)
with initial conditions
x(r, 0) = c1 (r) = r and t(r, 0) = c2 (r) = 0.
Therefore,
x(r, s) = g(r)s + r, t(r, s) = s and z(r, s) = g(r).
Let us compute J(r, s) = 1 + sg 0 (r) − 0 · 1 = 1 + sg 0 (r) and J(r, 0) = 1 6= 0,

the data curve Γ does not have characteristic points. Hence, one can solve
for r and s, in terms of x, t and z. Thus, we get s = t and r = x − zt.
Therefore, u(x, t) = g(x − tu) is the solution in the implicit form.
(
ut+ uux = 1 x ∈ R and t ∈ (0, ∞)
2
u t4 , t = 2t t > 0.
Note that the data curve is the parabola x = t2 /4. The parametrization
of the curve Γ with r-variable, i.e., Γ = {γ1 (r), γ2 (r)} = {(r2 , 2r)}. The
characteristic equations are:

= z, = 1, and =1
ds ds ds
x(r, 0) = r2 , t(r, 0) = 2r, and z(r, 0) = r.

Solving the ODE corresponding to z, we get z(r, s) = s + c3 (r) with initial

conditions z(r, 0) = c3 (r) = r. Thus, z(r, s) = s + r. Using this in the ODE
of x, we get
dx(r, s)
= s + r.
ds
s2
x(r, s) = + rs + c1 (r), t(r, s) = s + c2 (r)
2
x(r, 0) = c1 (r) = r2 and t(r, 0) = c2 (r) = 2r.
Therefore,
s2
x(r, s) = + rs + r2 , t(r, s) = s + 2r and z(r, s) = s + r.
2
Let us compute J(r, s) = 2r + s − 2(s + r) = −s and J(r, 0) = 0 for all r.
The rank of the matrix, at (r, 0),

r 1 1
= 2 6= 1
2r 2 1
and, hence, Γ does not have any characteristic points. We know in this case
we cannot have a C 1 solution but might have less regular solutions.
q Let us
2
solve for r and s, in terms of x, t and z. Thus, we get s = ∓2 x − t4 and
q q
2 2
r = 2t ± x − t4 . Therefore, u(x, t) = 2t ± x − t4 are two solutions in the
region x > t2 /4 and are not differentiable on Γ.
(
ut+ uux = 1 x ∈ R and t ∈ (0, ∞)
2
u t2 , t = t t > 0.
Note that the data curve is the parabola x = t2 /2. The parametrization
of the curve Γ with r-variable, i.e., Γ = {γ1 (r), γ2 (r)} = {(r2 /2, r)}. The
characteristic equations are:
= z, = 1, and =1
ds ds ds
r2
x(r, 0) = , t(r, 0) = r, and z(r, 0) = r.
2
Solving the ODE corresponding to z, we get z(r, s) = s + c3 (r) with initial
conditions z(r, 0) = c3 (r) = r. Thus, z(r, s) = s + r. Using this in the ODE
of x, we get
dx(r, s)
= s + r.
ds
s2
x(r, s) = + rs + c1 (r), t(r, s) = s + c2 (r)
2
r2
x(r, 0) = c1 (r) = and t(r, 0) = c2 (r) = r.
2
Therefore,
s2 + r 2
x(r, s) = + rs, t(r, s) = s + r and z(r, s) = s + r.
2
Let us compute J(r, s) = r + s − (s + r) = 0 for all r and s. Further, Γ is a
characteristic (at all points) because the rank of the matrix, at (r, 0),

r 1 1
=1
r 1 1
In this case there may exist infinitely many C 1 solutions. For√instance,

u(x, t) = t is a solution satisfying Cauchy data. Also, u(x, t) = ± 2x.
Remark 4.0.1. If the coefficients a and b are independent of u, then the

characteristic curves are lying in the xy-plane. If the coefficients a and b are
constants (independent of both x and u) then the characteristic curves are
straight lines. In the linear case, the characteristics curves will not intersect.
Because, if the curves intersect then, at the point of intersection, they have
the same tangent, which is not possible!
Lecture 5
Classification of Second Order

PDE
A general second order PDE is of the form F (D2 u(x), Du(x), u(x), x) = 0,
for each x ∈ Ω ⊂ Rn and u : Ω → R is the unknown. A Cauchy problem is:
given the knowledge of u on a smooth hypersurface Γ ⊂ Ω, can one find the
solution u of the PDE? The knowledge of u on Γ is said to be the Cauchy
data.
What should be the minimum required Cauchy data for the Cauchy prob-
lem to be solved? Viewing the Cauchy problem as an initial value problem
corresponding to ODE, there is a unique solution to the second order ODE
 00
 y (x) + P (x)y 0 (x) + Q(x)y(x) = 0 x ∈ I
y(x0 ) = y0
y 0 (x0 ) = y00 .

where P and Q are continuous on I (assume I closed interval of R) and for

any point x0 ∈ I. This motivates us to define the Cauchy problem for second
order PDE as:

 F (D2 u(x), Du(x), u(x), x) = 0 x∈Ω
u(x) = g(x) x ∈ Γ (5.0.1)
Du(x) · ν(x) = h(x) x ∈ Γ

where ν is the outward unit normal vector on the hypersurface Γ and g, h

are known functions on Γ.
23
LECTURE 5. CLASSIFICATION OF SECOND ORDER PDE 24
5.1 Semilinear
Consider the Cauchy problem for the second order semilinear PDE in two
variables (x, y) ∈ Ω ⊂ R2 ,


 A(x, y)uxx + 2B(x, y)uxy + C(x, y)uyy = D (x, y) ∈ Ω
u(x, y) = g(x, y) (x, y) ∈ Γ

(5.1.1)

 ux (x, y) = h1 (x, y) (x, y) ∈ Γ
uy (x, y) = h2 (x, y) (x, y) ∈ Γ.

where D(x, y, u, ux , uy ) may be non-linear and Γ is a smooth1 curve in Ω.

Also, one of the coefficients A, B or C is identically non-zero (else the PDE
is not of second order). Let r 7→ (γ1 (r), γ2 (r)) be a parametrisation of the
curve Γ. Then we have the compatibility condition that
g 0 (r) = h1 γ10 (r) + h2 γ20 (r).
By computing the second derivatives of u on Γ and considering uxx , uyy

and uxy as unknowns, we have the linear system of three equations in three
unknowns on Γ,
Auxx +2Buxy +Cuyy =D

0 0
γ1 (r)uxx +γ2 (r)uxy = h01 (r)
γ10 (r)uxy +γ20 (r)uyy = h02 (r).
This system of equation is solvable if the determinant of the coefficients are

non-zero, i.e.,
A 2B C
0
γ1 γ20 0 =
6 0.
0 γ10 γ20
Definition 5.1.1. We say a curve Γ ⊂ Ω ⊂ R2 is characteristic (w.r.t

(5.1.1)) if
A(γ20 )2 − 2Bγ10 γ20 + C(γ10 )2 = 0.
where (γ1 (r), γ2 (r)) is a parametrisation of Γ.
Note that the geometry hidden in the above definition is very similar
to that we encountered in first order equation. Since ν = (−γ20 , γ10 ) is the
1
twice differentiable
normal to Γ at each point, the above definition says that the curve is non-
characteristic if
2
X
Aij νi νj = A(γ20 )2 − 2Bγ10 γ20 + C(γ10 )2 6= 0
i,j=1
where A11 = A, A12 = A21 = B and A22 = C. If y = y(x) is a representation

of the curve Γ (locally, if necessary), we have γ1 (r) = r and γ2 (r) = y(r).
Then the characteristic equation reduces as
2
dy dy
A − 2B + C = 0.
dx dx
Therefore, the characteristic curves of (5.1.1) are given by the graphs whose
equation is √
dy B ± B 2 − AC
= .
dx A
Thus, we have three situations arising depending on the sign of the dis-
criminant, B 2 − AC. This classifies the given second order PDE based on
the sign of its discriminant d = B 2 − AC.
Definition 5.1.2. We say a second order PDE is of
(a) hyperbolic type if d > 0,
(b) parabolic type if d = 0 and
(c) elliptic type if d < 0.
The hyperbolic PDE have two families of characteristics, parabolic PDE
has one family of characteristic and elliptic PDE have no characteristic. We
caution here that these names are no indication of the shape of the graph of
the solution of the PDE.
Note that the classification depends on the determinant of the coefficient
matrix
A B
B C
For every (x, y) ∈ Ω, the matrix is symmetric and hence diagonalisable.
If λ1 , λ2 are the diagonal entries, then d = −λ1 λ2 . Thus, a equation is
hyperbolic at a point (x, y) if the eigen values have opposite sign. It is ellipic
if the eigenvalues have same sign and is parabolic if, at least, one of the
eigenvalue is zero.
Example 5.1 (Wave Equation). For a given non-zero c ∈ R, uyy − c2 uxx = 0 is

hyperbolic. Since A = −c2 , B = 0 and C = 1, we have d = B 2 −AC = c2 > 0.
The eigen values of the coefficient matrix are 1, −c2 which have opposite sign.
Example 5.2 (Heat Equation). For a given c ∈ R, uy − cuxx = 0 is parabolic.
Since A = −c, B = 0 and C = 0, thus d = B 2 − AC = 0. The eigen values
of the coefficient matrix are 0, −c has a zero eigenvalue.
Example 5.3 (Laplace equation). uxx +uyy = 0 is elliptic. Since A = 1, B = 0
and C = 1, thus d = B 2 − AC = −1 < 0. The eigen values of the coefficient
matrix are 1, 1 which have same sign.
Example 5.4 (Velocity Potential Equation). In the equation (1 − M 2 )uxx +
uyy = 0, A = (1 − M 2 ), B = 0 and C = 1. Then d = B 2 − AC = −(1 − M 2 ).
The eigen values of the coefficient matrix are 1 − M 2 , 1. Thus, for M > 1
(opposite sign), the equation is hyperbolic (supersonic flow), for M = 1 (zero
eigenvalue) it is parabolic (sonic flow) and for M < 1 (same sign) it is elliptic
(subsonic flow).
Note that the classification of PDE is dependent on its coefficients. Thus,
for constant coefficients the type of PDE remains unchanged throughout
the region Ω. However, for variable coefficients, the PDE may change its
classification from region to region.
Example 5.5. An example is the Tricomi equation , uxx + xuyy = 0. The
discriminant of the Tricomi equation is d = −x. The eigenvalues are 1, x.
Thus, tricomi equation is hyperbolic when x < 0, elliptic when x > 0 and
degenerately parabolic when x = 0, i.e., y-axis. Such equations are called
mixed type.
Example 5.6. Let us compute the family of characteristic curves of second
order PDE, whenever they exist. For instance, recall that elliptic equation
will not have any real characteristic curves.
(i) For a given non-zero c ∈ R, uyy −c2 uxx = 0. We have already noted that
the equation is hyperbolic and, hence, should admit two characteristic
curves. Recall that the characteristic curves are given by the equation
√ √
dy B ± B 2 − AC ± c2 ∓1
= = 2
= .
dx A −c c
Thus, cy ± x = a constant is the equation for the two characteristic
curves. Note that the characteristic curves y = ∓x/c + y0 are boundary
of two cones in R2 with vertex at (0, y0 ).
(ii) For any given c ∈ R, consider uy −cuxx = 0. We have already noted that
the equation is parabolic and, hence, should admit one characteristic
curve. The characteristic curve is given by the equation
√
dy B ± B 2 − AC
= = 0.
dx A
Thus, y = a constant is the equation of the characteristic curve. i.e.,
any horizontal line in R2 is a charateristic curve.
(iii) We have already noted that the equation uxx + uyy = 0 is elliptic and,
hence, will have no real characteristics.
(iv) The equation uxx + xuyy = 0 is of mixed type. In the region x > 0, the
characteristic curves are y ∓ 2x3/2 /3 = a constant.
Lecture 6
Classification of SOPDE:
Continued
6.1 Quasilinear
The notion of classification of second order semilinear PDE could be gen-
eralised to quasilinear PDE A(x, u(x), Du(x)), non-linear PDE and system
of ODE. However, in these cases the classification may also depend on the
solution u. The solutions to characteristic equation for a quasilinear equation
depends on the solution considered.
Example 6.1. Consider the quasilinear PDE uxx −uuyy = 0. The discriminant
is d = u. The eigenvalues are 1, −u(x). It is hyperbolic for {u > 0}1 , elliptic
when {u < 0} and parabolic when {u = 0}.
Example 6.2. Consider the quasilinear PDE
(c2 − u2x )uxx − 2ux uy uxy + (c2 − u2y )uyy = 0
where c > 0. Then d = B 2 − AC = c2 (u2x + u2y − c2 ) = c2 (|∇u|2 − c2 ). It is
hyperbolic if |∇u| > c, parabolic if |∇u| = c and elliptic if |∇u| < c.
6.2 Why Characteristic Curves?

Recall that a second order ODE
y 00 (x) + P (x)y 0 (x) + Q(x)y(x) = 0, x ∈ I
1
The notation {u > 0} means {x ∈ Ω | u(x) > 0}
29
LECTURE 6. CLASSIFICATION OF SOPDE: CONTINUED 30
can have other types of boundary conditions, in addition to the initial (or
Cauchy) condition
y(x0 ) = y0 and y 0 (x0 ) = y00 ,
such as, if I = (a, b) then
(a) Dirichlet condition

y(a) = y0 and y(b) = y1 .
(b) Neumann condition
y 0 (a) = y0 and y 0 (b) = y1 .
(c) Periodic condition
y(a) = y(b) and y 0 (a) = y 0 (b).
(d) Robin condition
αy(a) + βy 0 (a) = y0 and γy(b) + δy 0 (b) = y1 .
In contrast to Initial (Cauchy boundary) condition, boundary value problems

may be ill-posed. These boundary conditions also can be generalised to
second order PDE. The classification described tells us the right amount
of initial/boundary condition to be imposed for a second order PDE to be
well-posed.
For hyperbolic, which has two real characteristics, requires as many ini-
tial condition as the number of characteristics emanating from initial time
and as many boundary conditions as the number of characteristics that pass
into the spatial boundary. Thus, hyperbolic equations will take the Cauchy
boundary conditions on an open surface. For parabolic, which has exactly
one real characteristic, we need one boundary condition at each point of the
spatial boundary and one initial condition at initial time. Thus, parabolic
equation will take either Dirichlet or Neumann on an open surface. For ellip-
tic, which has no real characteristic curves, we need one boundary condition
at each point of the spatial boundary. Thus, elliptic equations will take either
Dirichlet or Neumann in a closed surface enclosing the domain of interest.
6.3 Cauchy Boundary Condition

Recall that for a second order Cauchy problem we need to know both u
and its normal derivative on a data curve Γ contained in Ω. However, the
Cauchy problem for Laplacian (more generally for elliptic equations) is not
well-posed. In fact, the Cauchy problem for Laplace equation on a bounded
domain Ω is over-determined.
Example 6.3 (Hadamard). Consider the Cauchy problem for Laplace equa-
tion 
 uxx + uyy = 0
u(0, y) = cosk2ky
ux (0, y) = 0,

where k > 0 is an integer. It is easy to verify that there is a unique solution

cosh(kx) cos(ky)
uk (x, y) =
k2
of the Cauchy problem. Note that for any x0 > 0,
cosh(kx0 )
|uk (x0 , nπ/k)| = .
k2
Since, as k → ∞, nπ/k → 0 and |uk (x0 , nπ/k)| → ∞ the Cauchy problem is
not stable, and hence not well-posed.
Exercise 1. Show that the Cauchy problem for Laplace equation

 uxx + uyy = 0
u(x, 0) =0
uy (x, 0) = k −1 sin kx,

where k > 0, is not well-posed. (Hint: Compute explicit solution using sep-
aration of variable. Note that, as k → ∞, the Cauchy data tends to zero
uniformly, but the solution does not converge to zero for any y 6= 0. There-
fore, a small change from zero Cauchy data (with corresponding solution
being zero) may induce bigger change in the solution.)
This issue of ill-posedness of the Cauchy problem is very special to second
order elliptic equations. In general, any hyperbolic equation Cauchy problem
is well-posed, as long as the hyperbolicity is valid in the full neighbourhood
of the data curve.
Example 6.4. Consider the Cauchy problem for the second order hyperbolic
equation  2
 y uxx − yuyy + 21 uy = 0 y > 0
u(x, 0) = f (x)
uy (x, 0) = g(x).

The general solution to this problem can be computed as

2 3/2 2 3/2
u(x, y) = F x + y +G x− y .
3 3
On y = 0 u(x, 0) = F (x) + G(x) = f (x). Further,

1/2 0 2 3/2 1/2 0 2 3/2
uy (x, y) = y F x + y −y G x− y
3 3
and uy (x, 0) = 0. Thus, the Cauchy problem has no solution unless g(x) = 0.
If g ≡ 0 then the solution is

2 3/2 2 3/2 2 3/2
u(x, y) = F x + y −F x− y +f x− y
3 3 3
for arbitrary F ∈ C 2 . Therefore, when g ≡ 0 the solution is not unique.

The Cauchy problem is not well-posed because the equation is hyperbolic
(B 2 − AC = y 3 ) not in the full neighbourhood of the data curve {y = 0}.
Lecture 7
Classification of SOPDE:
Continued
7.1 Invariance of Discriminant

The classification of second order semilinear PDE is based on the discriminant
B 2 −AC. In this section, we note that the classification is independent of the
choice of coordinate system (to represent a PDE). Consider the two-variable
semilinear PDE
A(x, y)uxx +2B(x, y)uxy +C(x, y)uyy = D(x, y, u, ux , uy ) (x, y) ∈ Ω (7.1.1)
where the variables (x, y, u, ux , uy ) may appear non-linearly in D and Ω ⊂ R2 .

Also, one of the coefficients A, B or C is identically non-zero (else the PDE is
not of second order). We shall observe how (7.1.1) changes under coordinate
transformation.
Definition 7.1.1. For any PDE of the form (7.1.1) we define its discrimi-
nant as B 2 − AC.
Let T : R2 → R2 be the coordinate transformation with components

T = (w, z), where w, z : R2 → R. We assume that w(x, y), z(x, y) are such
that w, z are both continuous and twice differentiable w.r.t (x, y), and the
Jacobian J of T is non-zero,

wx wy
J = 6= 0.
zx zy
33
We compute the derivatives of u in the new variable,
ux = uw wx + uz zx ,
uy = uw wy + uz zy ,
uxx = uww wx2 + 2uwz wx zx + uzz zx2 + uw wxx + uz zxx
uyy = uww wy2 + 2uwz wy zy + uzz zy2 + uw wyy + uz zyy
uxy = uww wx wy + uwz (wx zy + wy zx ) + uzz zx zy + uw wxy + uz zxy
Substituting above equations in (7.1.1), we get
a(w, z)uww + 2b(w, z)uwz + c(w, z)uzz = d(w, z, u, uw , uz ).
where D transforms in to d and
a(w, z) = Awx2 + 2Bwx wy + Cwy2 (7.1.2)

b(w, z) = Awx zx + B(wx zy + wy zx ) + Cwy zy (7.1.3)
c(w, z) = Azx2 + 2Bzx zy + Czy2 . (7.1.4)
Note that the coefficients in the new coordinate system satisfy
b2 − ac = (B 2 − AC)J 2 .
Since J 6= 0, we have J 2 > 0. Thus, both b2 − ac and B 2 − AC have the

same sign. Thus, the sign of the discriminant is invariant under coordinate
transformation. All the above arguments can be carried over to quasilinear
and non-linear PDE.
7.2 Standard or Canonical Forms

The advantage of above classification helps us in reducing a given PDE into
simple forms. Given a PDE, one can compute the sign of the discriminant
and depending on its clasification we can choose a coordinate transformation
(w, z) such that
(i) For hyperbolic, a = c = 0 or b = 0 and a = −c.
(ii) For parabolic, c = b = 0 or a = b = 0. We conveniently choose

c = b = 0 situation so that a 6= 0 (so that division by zero is avoided in
the equation for characteristic curves).
(iii) For elliptic, b = 0 and a = c.

If the given second order PDE (7.1.1) is such that A = C = 0, then
6 0) gives
(7.1.1) is of hyperbolic type and a division by 2B (since B =
uxy = D̃(x, y, u, ux , uy )
where D̃ = D/2B. The above form is the first standard form of second order
hyperbolic equation. If we introduce the linear change of variable X = x + y
and Y = x − y in the first standard form, we get the second standard form
of hyperbolic PDE
uXX − uY Y = D̂(X, Y, u, uX , uY ).
If the given second order PDE (7.1.1) is such that A = B = 0, then
(7.1.1) is of parabolic type and a division by C (since C 6= 0) gives
uyy = D̃(x, y, u, ux , uy )
where D̃ = D/C. The above form is the standard form of second order
parabolic equation.
If the given second order PDE (7.1.1) is such that A = C and B = 0,
then (7.1.1) is of elliptic type and a division by A (since A 6= 0) gives
uxx + uyy = D̃(x, y, u, ux , uy )
where D̃ = D/A. The above form is the standard form of second order
elliptic equation.
Note that the standard forms of the PDE is an expression with no mixed
derivatives.
7.3 Reduction to Standard Form

Consider the second order semilinear PDE (7.1.1) not in standard form. We
look for transformation w = w(x, y) and z = z(x, y), with non-vanishing
Jacobian, such that the reduced form is the standard form.
If B 2 − AC > 0, we have two characteristics. We are looking for the
coordinate system w and z such that a = c = 0. This implies from equation
(7.1.2) and (7.1.4) that we need to find w and z such that
√
wx −B ± B 2 − AC zx
= = .
wy A zy
Therefore, we need to find w and z such that along the slopes of the charac-
teristic curves, √
dy B ± B 2 − AC −wx
= = .
dx A wy
This means that, using the parametrisation of the characteristic curves,
wx γ10 (r) + wy γ20 (r) = 0 and w0 (r) = 0. Similarly for z. Thus, w and z
are chosen such that they are constant on the characteristic curves.
The characteristic curves are found by solving
√
dy B ± B 2 − AC
=
dx A
and the coordinates are then chosen such that along the characteristic curve
w(x, y) √
2
and z(x, y) = a constant. Note that wx zy − wy zx =
= a constant
2
wy zy A B − AC 6= 0.
Example 7.1. For a non-zero constant c ∈ R, let us reduce to canonical
form the PDE uxx − c2 uyy = 0. Note that A = 1, B = 0, C = −c2 and
B 2 − AC = c2 and the equation is hyperbolic. The characteristic curves are
given by the equation
√
dy B ± B 2 − AC
= = ±c.
dx A
Solving we get y ∓ cx = a constant. Thus, w = y + cx and z = y − cx. Now
writing

= c2 (uww − 2uwz + uzz )
= uww + 2uwz + uzz
2
−c uyy = −c2 (uww + 2uwz + uzz )
Substituting into the given PDE, we get
0 = 4c2 uwz
= uwz .
Example 7.2. Let us reduce the PDE uxx − x2 yuyy = 0 given in the region
{(x, y) | x ∈ R, x 6= 0, y > 0} to its canonical form. Note that A = 1, B = 0,
C = −x2 y and B 2 − AC = x2 y. In the given region x2 y > 0, hence the

equation is hyperbolic. The characteristic curves are given by the equation
√
dy B± B 2 − AC √
= = ±x y.
dx A
√ √
Solving we get x2 /2 ∓ 2 y = a constant. Thus, w = x2 /2 + 2 y and
√
z = x2 /2 − 2 y. Now writing
ux = uw wx + uz zx = x(uw + uz )
1
uy = uw wy + uz zy = √ (uw − uz )
y
= x2 (uww + 2uwz + uzz ) + uw + uz
1 1
= (uww − 2uwz + uzz ) − √ (uw − uz )
y 2y y
x2
−x2 yuyy = −x2 (uww − 2uwz + uzz ) + √ (uw − uz )
2 y

√ √
2 2 y + x2 2 y − x2
0 = 4x uwz + √ uw + √ uz
2 y 2 y
√ √ √
= 8x2 yuwz + (x2 + 2 y)uw + (2 y − x2 )uz .
√
Note that w + z = x2 and w − z = 4 y. Now, substituting x, y in terms of
w, z, we get

2 2 w−z w−z
0 = 2(w − z )uwz + w + z + uw + − w − z uz
2 2

3w + z w + 3z
= uwz + uw − uz .
4(w2 − z 2 ) 4(w2 − z 2 )
In the parabolic case, B 2 − AC = 0, we have a single characteristic. We

are looking for a coordinate system such that b = c = 0.
Example 7.3. Let us reduce the PDE e2x uxx + 2ex+y uxy + e2y uyy = 0 to its
canonical form. Note that A = e2x , B = ex+y , C = e2y and B 2 − AC = 0.
The PDE is parabolic. The characteristic curves are given by the equation
dy B ey
= = x.
dx A e
Solving, we get e−y − e−x = a constant. Thus, w = e−y − e−x . Now, we
choose z such that wx zy − wy zx 6= 0. For instance, z = x is one such choice.
Then
ux = e−x uw + uz
uy = −e−y uw
uxx = e−2x uww + 2e−x uwz + uzz − e−x uw
uyy = e−2y uww + e−y uw
uxy = −e−y (e−x uww − uwz )
ex e−y uzz = (e−y − e−x )uw
Replacing x, y in terms of w, z gives
w
uzz = uw .
1 + wez
In the elliptic case, B 2 − AC < 0, we have no real characteristics. Thus,
we choose w, z to be the real and imaginary part of the solution of the
characteristic equation.
Example 7.4. Let us reduce the PDE x2 uxx + y 2 uyy = 0 given in the region
{(x, y) ∈ R2 | x > 0, y > 0} to its canonical form. Note that A = x2 ,
B = 0, C = y 2 and B 2 − AC = −x2 y 2 < 0. The PDE is elliptic. Solving the
characteristic equation
dy iy
=±
dx x
we get ln x ± i ln y = c. Let w = ln x and z = ln y. Then
ux = uw /x
uy = uz /y
uxx = −uw /x2 + uww /x2
uyy = −uz /y 2 + uzz /y 2
Substituting into the PDE, we get
uww + uzz = uw + uz .
Example 7.5. Let us reduce the PDE uxx + 2uxy + 5uyy = xux to its canonical
form. Note that A = 1, B = 1, C = 5 and B 2 − AC = −4 < 0. The PDE is
elliptic. The characteristic equation is
dy
= 1 ± 2i.
dx
Solving we get x − y ± i2x = c. Let w = x − y and z = 2x. Then
ux = uw + 2uz
uy = −uw
uxx = uww + 4uwz + 4uzz
uyy = uww
uxy = −(uww + 2uwz )
Substituting into the PDE, we get
uww + uzz = x(uw + 2uz )/4.
Replacing x, y in terms of w, z gives

z
uww + uzz = (uw + 2uz ).
8
Lecture 8
The Laplacian
We introducedPin Lecture 1 the Laplacian to be the trace of the Hessain

∂2
matrix, ∆ := ni=1 ∂x 2 . The Laplace operator usually appears in physical
i
models associated with dissipative effects (except wave equation). The im-
portance of Laplace operator can be realised by its appearance in various
physical models. For instance, in the
∂
(a) heat equation ∂t
− ∆,
∂2
(b) the wave equation ∂t2
− ∆,
∂
(c) and the Schrödinger’s equation i ∂t + ∆.
8.1 Properties of Laplacian

∂2
Pn
In cartesian coordiantes, the n-dimensional Laplacian is ∆ := i=1 ∂x2i . Note
d2
that in one dimension, i.e. n = 1, ∆ = .
In two dimension polar coordi-
dx2
nates, the Laplacian is given as
1 ∂2

1 ∂ ∂
∆ := r + 2 2,
r ∂r ∂r r ∂θ
where r is the magnitude component (0 ≤ r < ∞) and θ is the direction

component (0 ≤ θ < 2π). The direction component is also called the azimuth
angle or polar angle. This is easily seen by using the relation x = r cos θ and
41
LECTURE 8. THE LAPLACIAN 42
∂x ∂y ∂u
y = r sin θ. Then ∂r
= cos θ, ∂r
= sin θ and ∂r
= cos θ ∂u
∂x
+ sin θ ∂u
∂y
. Also,
∂ 2u 2
2 ∂ u
2
2 ∂ u ∂ 2u
= cos θ 2 + sin θ 2 + 2 cos θ sin θ .
∂r2 ∂x ∂y ∂x∂y
∂x ∂y ∂u
Similarly, ∂θ
= −r sin θ, ∂θ
= r cos θ, ∂θ
= r cos θ ∂u
∂y
− r sin θ ∂u
∂x
and
1 ∂ 2u 2
2 ∂ u
2
2 ∂ u ∂ 2u 1 ∂u
2 2
= sin θ 2
+ cos θ 2
− 2 cos θ sin θ − .
r ∂θ ∂x ∂y ∂x∂y r ∂r
∂2u 1 ∂2u ∂2u ∂2u 1 ∂u
Therefore, ∂r2
+ r2 ∂θ2
= ∂x2
+ ∂y 2
− r ∂r
and, hence,
∂ 2u 1 ∂ 2 u 1 ∂u
∆u = 2 + 2 2 + .
∂r r ∂θ r ∂r
Further, in three dimension cylindrical coordinates, the Laplacian is given as
1 ∂2 ∂2

1 ∂ ∂
∆ := r + 2 2+ 2
r ∂r ∂r r ∂θ ∂z
where r ∈ [0, ∞), θ ∈ [0, 2π) and z ∈ R. In three dimension spherical

coordinates, the Laplacian is given as
∂2

1 ∂ 2 ∂ 1 ∂ ∂ 1
∆ := 2 r + 2 sin φ + 2 2
r ∂r ∂r r sin φ ∂φ ∂φ r sin φ ∂θ2
where r ∈ [0, ∞), φ ∈ [0, π] (zenith angle or inclination) and θ ∈ [0, 2π)
(azimuth angle).
Theorem 8.1.1. Let n ≥ 2 and u be a radial function, i.e., u(x) = v(r)

where x ∈ Rn and r = |x|, then
d2 v(r) (n − 1) dv(r)
∆u(x) = + .
dr2 r dr
Proof. Note that
p
∂r ∂|x| ∂( x21 + . . . + x2n ) 1 xi
= = = (x21 + . . . + x2n )−1/2 (2xi ) = .
∂xi ∂xi ∂xi 2 r
Thus,
n n
X ∂ ∂u(x) X ∂ dv(r) xi
∆u(x) = =
i=1
∂xi ∂xi i=1
∂xi dr r
n
X ∂ 1 dv(r) n dv(r)
= xi +
i=1
∂xi r dr r dr
n
x2i d dv(r) 1

X n dv(r)
= +
i=1
r dr dr r r dr
n
x2i 1 d2 v(r)

X 1 dv(r) n dv(r)
= − +
i=1
r r dr2 r2 dr r dr
r2 1 d2 v(r)

1 dv(r) n dv(r)
= 2
− 2 +
r r dr r dr r dr
2
d v(r) 1 dv(r) n dv(r)
= − +
dr2 r dr r dr
d2 v(r) (n − 1) dv(r)
= + .
dr2 r dr
Hence the result proved.
More generally, the Laplacian in Rn may be written in polar coordinates
as
∂2 n−1 ∂ 1
∆ := 2 + + 2 ∆Sn−1
∂r r ∂r r
where ∆Sn−1 is a second order differential operator in angular variables only.
The angular part of Laplacian is called the Laplace-Beltrami operator acting
on Sn−1 (unit sphere of Rn ) with Riemannian metric induced by the standard
Euclidean metric in Rn .
8.2 Boundary Conditions

Let Ω be a bounded open subset of Rn with boundary denoted as ∂Ω. To
make the over-determined Cauchy problem of an elliptic equation well-posed,
it is reasonable to specify one of the following conditions on the boundary
∂Ω:
(i) (Dirichlet condition) u = g;
(ii) (Neumann condition) ∇u·ν = g, where ν(x) is the unit outward normal
of x ∈ ∂Ω;
(iii) (Robin condition) ∇u · ν + cu = g for any c > 0.
(iv) (Mixed condition) u = g on Γ1 and ∇u·ν = h on Γ2 , where Γ1 ∪Γ2 = ∂Ω

and Γ1 ∩ Γ2 = ∅.
The elliptic equation with Neumann boundary condition naturally im-

poses a compatibility condition. By Gauss divergence theorem (cf. Corol-
lary A.0.5), if u is a solution of the Neumann problem then u satisfies, for
every connected component ω of Ω,
Z Z
∆u = ∇u · ν (Using GDT)
ωZ Z∂ω
− f = g.
ω ∂ω
The second equality is called the compatibility condition. Thus, for an inho-
mogeneous Laplace equation with Neumann boundary condition, the given
data f, g must necessarily satisfy the compatibility condition. Otherwise, the
Neumann problem does not make any sense.
The aim of this chapter is to solve, for any open bounded subset Ω ⊂ Rn ,

−∆u(x) = f (x) in Ω
one of the above inhomogeneous boudary condition on ∂Ω.
By the linearity of Laplacian, u = v + w where v is a solution of

∆v(x) =0 in Ω
one of the above inhomogeneous boudary condition on ∂Ω,
and w is a solution of

−∆w(x) = f (x) in Ω
one of the above homogeneous boudary condition on ∂Ω.
Therefore, we shall solve for u by solving for v and w separately.

8.3 Harmonic Functions

d 2
The one dimensional Laplace equation is an ODE ( dx 2 ) and is solvable with
solutions u(x) = ax + b for some constants a and b. But in higher dimensions

solving Laplace equation is not so simple. For instance, a two dimensional
Laplace equation
uxx + uyy = 0
has the trivial solution, u(x, y) = ax + by + c, all one degree polynomials of
two variables. In addition, xy, x2 − y 2 , x3 − 3xy 2 , 3x2 y − y 3 , ex sin y and
ex cos y are all solutions to the two variable Laplace equation. In Rn , it is
trivial to check that all polynomials up to degree one, i.e.
X
aα x α
|α|≤1
is a solution to ∆u = 0 in Rn . But we also have functions of higher degree

and functions not expressible in terms of elementaryQfunctions as solutions
to Laplace equation. For instance, note that u(x) = ni=1 xi is a solution to
∆u = 0 in Rn .
Definition 8.3.1. Let Ω be an open subset of Rn . A function u ∈ C 2 (Ω) is

said to be harmonic on Ω if ∆u(x) = 0 in Ω.
Gauss was the first to deduce some important properties of harmonic

functions and thus laid the foundation for Potential theory and Harmonic
Analysis. Due to the linearity of ∆, sum of any finite number of har-
monic functions is harmonic and a scalar multiple of a harmonic function is
harmonic.
In two dimension, one associates with a harmonic function u(x, y), a
conjugate harmonic function, v(x, y) defined as the solution of a first order
system of PDE called the Cauchy-Riemann equations,
ux (x, y) = vy (x, y) and uy (x, y) = −vx (x, y).
Harmonic functions and holomorphic functions (differentiable complex func-

tions) are related in the sense that, for any pair (u, v), harmonic and its conju-
gate, gives a holomorphic function f (z) = u(x, y) + iv(x, y) where z = x + iy.
Conversely, for any holomorphic function f , its real part and imaginary part
are conjugate harmonic functions. This observation gives us more examples
of harmonic functions, for instance, since all complex polynomials f (z) = z m

are holomorphic we have (using the polar coordinates) u(r, θ) = rm cos mθ
and v(r, θ) = rm sin mθ are harmonic functions in R2 for all m ∈ N. Simi-
larly, since f (z) = log z = ln r + iθ is holomorphic in certain region, we have
u(r, θ) = ln r and v(r, θ) = θ are harmonic in R2 \ (0, 0) and R2 \ {θ = 0},
respectively.
Lecture 9
Properties of Harmonic
Functions
In this section we shall study properties of harmonic functions. We shall as-

sume the divergence theorems from multivariable calculus (cf. Appendix A).
Also, note that if u is a harmonic function on Ω then, by Gauss divergence
theorem (cf. Theorem A.0.4),
Z
∂u
dσ = 0.
∂Ω ∂ν
Theorem 9.0.1 (Maximum Principle). Let Ω be an open, bounded subset of

Rn . Let u ∈ C(Ω) be harmonic in Ω. Then
max u(y) = max u(y).

y∈Ω y∈∂Ω
Proof. Since ∂Ω ⊂ Ω, we have max∂Ω u ≤ maxΩ u. It only remains to prove

the other equality. For the given harmonic function u and for a fixed ε > 0,
we set vε (x) = u(x) + ε|x|2 , for each x ∈ Ω. For each x ∈ Ω, ∆vε =
∆u + 2nε > 0. Recall that1 if a function v attains local maximum at a point
x ∈ Ω, then in each direction its second order partial derivative vxi xi (x) ≤ 0,
for all i = 1, 2, . . . , n. Therefore ∆v(x) ≤ 0. Thus, we argue that vε does not
attain (even a local) maximum in Ω. But vε has to have a maximum in Ω,
1
v ∈ C 2 (a, b) has a local maximum at x ∈ (a, b) then v 0 (x) = 0 and v 00 (x) ≤ 0
47
LECTURE 9. PROPERTIES OF HARMONIC FUNCTIONS 48
hence it should be attained at some point x? ∈ ∂Ω, on the boundary. For all
x ∈ Ω,
u(x) ≤ vε (x) ≤ vε (x? ) = u(x? ) + ε|x? |2 ≤ max u(x) + ε max |x|2 .

x∈∂Ω x∈∂Ω
The above inequality is true for all ε > 0. Thus, u(x) ≤ maxx∈∂Ω u(x), for all
x ∈ Ω. Therefore, maxΩ u ≤ maxx∈∂Ω u(x). and hence we have equality.
9.0.1 Existence and Uniqueness of Solution

A consequence of the maximum principle is the uniqueness of the harmonic
functions.
Theorem 9.0.2 (Uniqueness of Harmonic Functions). Let Ω be an open,
bounded subset of Rn . Let u1 , u2 ∈ C 2 (Ω) ∩ C(Ω) be harmonic in Ω such that
u1 = u2 on ∂Ω, then u1 = u2 in Ω.
Proof. Note that u1 − u2 is a harmonic function and hence, by maximum
principle, should attain its maximum on ∂Ω. But u1 − u2 = 0 on ∂Ω. Thus
u1 − u2 ≤ 0 in Ω. Now, repeat the argument for u2 − u1 , we get u2 − u1 ≤ 0
in Ω. Thus, we get u1 − u2 = 0 in Ω.
In fact harmonic functions satisfy a much stronger maximum principle
whose proof is beyond the scope of this course.
Theorem 9.0.3 (Strong Maximum Principle). Let Ω be an open, connected
(domain) subset of Rn . Let u be harmonic in Ω and M := maxy∈Ω u(y).
Then
u(x) < M ∀x ∈ Ω
or u ≡ M is constant in Ω.
By the strong maximum principle (cf. Theorem 9.0.3), if Ω is connected
and g ≥ 0 and g(x) > 0 for some x ∈ ∂Ω then u(x) > 0 for all x ∈ Ω.
Theorem 9.0.4. Let Ω be an open bounded connected subset of Rn and
g ∈ C(∂Ω). Then the Dirichlet problem (9.0.1) has atmost one solution
u ∈ C 2 (Ω) ∩ C(Ω). Moreover, if u1 and u2 are solution to the Dirichlet
problem corresponding to g1 and g2 in C(∂Ω), respectively, then
(a) (Comparison) g1 ≥ g2 on ∂Ω and g1 (x0 ) > g2 (x0 ) for some x ∈ ∂Ω
implies that u1 > u2 in Ω.
(b) (Stability) |u1 (x) − u2 (x)| ≤ maxy∈∂Ω |g1 (y) − g2 (y)| for all x ∈ Ω.
Proof. The fact that there is atmost one solution to the Dirichlet problem
follows from the Theorem 9.0.2. Let w = u1 − u2 . Then w is harmonic.
(a) Note that w = g1 − g2 ≥ 0 on ∂Ω. Since g1 (x0 ) > g2 (x0 ) for some
x0 ∈ ∂Ω, then w(x) > 0 for all x ∈ ∂Ω. This proves the comparison
result.
(b) Again, by maximum principle, we have
±w(x) ≤ max |g1 (y) − g2 (y)|∀x ∈ Ω.

y∈∂Ω
This proves the stability result.
We remark that the uniqueness result is not true for unbounded domains.
Example 9.1. Let u ∈ C 2 (Ω) ∩ C(Ω) be a solution of the Dirichlet problem

∆u(x) = 0 x∈Ω
(9.0.1)
u(x) = g(x) x ∈ ∂Ω.
Let Ω = {x ∈ Rn | |x| > 1} and g ≡ 0. Obviously, u = 0 is a solution. But

we also have a non-trivial solution
(
ln |x| n=2
u(x) = 2−n
|x| − 1 n ≥ 3.
Example 9.2. Consider the problem (9.0.1) with g ≡ 0 and Ω = {x ∈ Rn |

xn > 0}. Obviously, u = 0 is a solution. But we also have a non-trivial
solution u(x) = xn .
We have shown above that if a solution exists for (9.0.1) then it is unique
(cf. Theorem 9.0.2). So it only remains to show the existence of solution
of (9.0.1), for any given domain Ω. In the modern theory, there are three
different methods to address the question of existence, viz., Perron’s Method,
Layer Potential (Integral Equations) and L2 methods which are beyond the
scope of this course.
Example 9.3 (Non-existence of Solutions). In 1912, Lebesgue gave an example

of a domain on which the classical Dirichlet problem is not solvable. The
domain is
Ω := {(x, y, z) ∈ R3 | r2 + z 2 < 1; r > e−1/2z for z > 0}.
Note that Ω is the unit ball in R3 with a sharp inward cusp, called Lebesgue
spine, at the origin (0, 0, 0).
Example 9.4. There are domains with inward cusps for which the classical
problem is solvable. For instance, consider
Ω := {(x, y, z) ∈ R3 | r2 + z 2 < 1; r > z 2k for z > 0},
for any positive integer k. The proof of this fact involves the theory of
capacities, beyond the scope of this course.
Remark 9.0.5 (Neumann Boundary Condition). The Neumann problem is

stated as follows: Given f : Ω → R and g : ∂Ω → R, find u : Ω → R such
that
−∆u = f in Ω
∂u (9.0.2)
∂ν
= g on ∂Ω
where ∂u
∂ν
:= ∇u · ν and ν = (ν1 , . . . , νn ) is the outward pointing unit normal
vector field of ∂Ω. Thus, the boundary imposed is called the Neumann
boundary condition. The solution of a Neumann problem is not necessarily
unique. If u is any solution of (9.0.2), then u + c for any constant c is also a
solution of (9.0.2). More generally, for any v such that v is constant on the
connected components of Ω, u + v is a solution of (9.0.2).
Lecture 10
Sturm-Liouville Problems
10.1 Eigen Value Problems

Definition 10.1.1. Let L denote a linear differential operator and I ⊂ R.
Then we say Ly(x) = λy(x) on I is an eigenvalue problem (EVP) corre-
sponding to L when both λ and y : I → R are unknown.
2
Example 10.1. For instance, if L = −d dx2
then its corresponding eigenvalue
problem is −y 00 = λy.
If λ ∈ R is fixed then one can obtain a general solution. But, in an EVP1
we need to find all λ ∈ R for which the given ODE is solvable. Note that
y ≡ 0 is a trivial solution, for all λ ∈ R.
Definition 10.1.2. A λ ∈ R, for which the EVP corresponding to L admits
a non-trivial solution yλ is called an eigenvalue of the operator L and yλ is
said to be an eigen function corresponding to λ. The set of all eigenvalues of
L is called the spectrum of L.
Exercise 2. Note that for a linear operator L, if yλ is an eigen function
corresponding to λ, then αyλ is also an eigen function corresponding to λ,
for all α ∈ R.
As a consequence of the above exercise note that, for a linear operator
L, the set of all eigen functions corresponding to λ forms a vector space
Wλ , called the eigen space corresponding to λ. Let V denote the set of all
solutions of the EVP corresponding to a linear operator L. Necessarily, 0 ∈ V
and V ⊂ C 2 (I). Note that V = ∪λ Wλ where λ’s are eigenvalues of L.
1
compare an EVP with the notion of diagonalisation of matrices from Linear Algebra
51
LECTURE 10. STURM-LIOUVILLE PROBLEMS 52
Exercise 3. Show that any second order ODE of the form
y 00 + P (x)y 0 + Q(x)y(x) = R(x)
can be written in the form

d dy
p(x) + q(x)y(x) = r(x).
dx dx
(Find p, q and r in terms of P , Q and R).
Proof. Let us multiply the original equation by a function p(x), which will
be chosen appropriately later. Thus, we get
p(x)y 00 + p(x)P (x)y 0 + p(x)Q(x)y(x) = p(x)R(x).
We shall choose µ such that
p0 = p(x)P (x).
R
Hence, p(x) = e P (x) dx . Thus, by setting q(x) = p(x)Q(x) and r(x) =
p(x)R(x), we have the other form.
10.2 Sturm-Liouville Problems

Given a finite interval (a, b) ⊂ R, the Sturm-Liouville (S-L) problem is given
as  d dy

 dx p(x) dx + q(x)y + λr(x)y = 0 x ∈ (a, b)
c1 y(a) + c2 y 0 (a) = 0. c21 + c22 > 0 (10.2.1)
d1 y(b) + d2 y 0 (b) = 0 d21 + d22 > 0.

The function y(x) and λ are unknown quantities. The pair of boundary
conditions given above is called separated. The boundary conditions corre-
sponds to the end-point a and b, respectively. Note that both c1 and c2
cannot be zero simultaneously and, similar condition on d1 and d2 .
Definition 10.2.1. The Sturm-Liouville problem with separated boundary
conditions is said to be regular if:
(a) p, p0 , q, r : [a, b] → R are continuous functions
(b) p(x) > 0 and r(x) > 0 for x ∈ [a, b].

We say the S-L problem is singular if either the interval (a, b) is un-
bounded or one (or both) of the regularity condition given above fails.
We say the S-L problem is periodic if p(a) = p(b) and the separated
boundary conditions are replaced with the periodic boundary condition y(a) =
y(b) and y 0 (a) = y 0 (b).
Example 10.2. Examples of regular S-L problem:
(a)
−y 00 (x) = λy(x) x ∈ (0, a)

y(0) = y(a) = 0.
We have chosen c1 = d1 = 1 and c2 = d2 = 0. Also, q ≡ 0 and p ≡ r ≡ 1.
(b)
−y 00 (x) = λy(x) x ∈ (0, a)

y 0 (0) = y 0 (a) = 0.
We have chosen c1 = d1 = 0 and c2 = d2 = 1. Also, q ≡ 0 and p ≡ r ≡ 1.
(c) 
 −y 00 (x) = λy(x) x ∈ (0, a)
y 0 (0) = 0
cy(a) + y 0 (a) = 0,

where c > 0 is a constant.
(d)  0
 − (x2 y 0 (x)) = λy(x) x ∈ (1, a)
y(1) = 0
y(a) = 0,

where p(x) = x2 , q ≡ 0 and r ≡ 1.
Remark 10.2.2. In a singular Sturm-Liouville problem, the boundary con-

dition at an (or both) end(s) is dropped if p vanishes in (or both) the cor-
responding end(s). This is because when p vanishes, the equation at that
point is no longer second order. Note that dropping a boundary condition
corresponding to a end-point is equivalent to taking both constants zero (for
instance, c1 = c2 = 0, in case of left end-point).
Example 10.3. Examples of singular S-L problem:

(a) For each n = 0, 1, 2, . . ., consider the Bessel’s equation

( 2
− (xy 0 (x))0 = − nx + λx y(x) x ∈ (0, a)
y(a) = 0,
where p(x) = r(x) = x, q(x) = −n2 /x. This equation is not regular
because p(0) = r(0) = 0 and q is not continuous in the closed interval
[0, a], since q(x) → −∞ as x → 0. Note that there is no boundary
condition corresponding to 0.
(b) The Legendre equation

0
− (1 − x2 )y 0 (x) = λy(x) x ∈ (−1, 1)

with no boundary condition. Here p(x) = 1 − x2 , q ≡ 0 and r ≡ 1. This

equation is not regular because p(−1) = p(1) = 0. Note that there is no
boundary conditions because p vanishes at both the end-points.
Example 10.4. Examples of periodic S-L problem:


 −y 00 (x) = λy(x) x ∈ (−π, π)
y(−π) = y(π)
 0
y (−π) = y 0 (π).
Lecture 11
Spectral Results
We shall now state without proof the spectral theorem for regular S-L prob-
lem. Our aim, in this course, is to check the validity of the theorem through
some examples.
Theorem 11.0.1. For a regular S-L problem, there exists an increasing se-
quence of eigenvalues 0 < λ1 < λ2 < λ3 < . . . < λk < . . . with λk → ∞, as
k → ∞.
Exercise 4. Let Wk = Wλk be the eigen space corresponding λk . Show that
for regular S-L problem Wk is one dimensional, i.e., corresponding to each
λk , there cannot be two or more linearly independent eigen vectors.
Example 11.1. Consider the boundary value problem,
y 00 + λy = 0 x ∈ (0, a)

y(0) = y(a) = 0.
This is a second order ODE with constant coeffcients. √ Its characteristic
equation is m2 + λ = 0. Solving for m, we get m = ± −λ. Note that
the λ can be either zero, positive or negative. If λ = 0, then y 00 = 0 and
the general solution is y(x) = αx + β, for some constants α and β. Since
y(0) = y(a) = 0 and a 6= 0, we get α = β = 0. Thus, we have no non-trivial
solution corresponding to λ = 0. √ √
If λ < 0, then ω = −λ > 0. Hence y(x) = αe ωx + βe− ωx . Using the
boundary condition y(0) = y(a) = 0, we get α = β = 0 and hence we have
no non-trivial solution corresponding
√ to negative√
λ’s. √
If λ > 0, then m = ±i λ and y(x) = α cos( λx) + β sin( λx). √ Using
the boundary condition y(0) = 0, we get α = 0 and y(x) = β sin( λx).
55
LECTURE 11. SPECTRAL RESULTS 56
√
Using y(a) = 0 (and β = 0 yields trivial solution), we assume sin( λa) = 0.
Thus, λ = (kπ/a)2 for each non-zero k ∈ N (since λ > 0). Hence, for each
k ∈ N, there is a solution (yk , λk ) with

kπx
yk (x) = sin ,
a
and λk = (kπ/a)2 . Notice the following properties of the eigenvalues λk and

eigen functions yk
(i) We have discrete set of λ’s such that 0 < λ1 < λ2 < λ3 < . . . and
λk → ∞, as k → ∞.
(ii) The eigen functions yλ corresponding to λ form a subspace of dimension

one.
In particular, in the above example, when a = π the eigenvalues, for each

k ∈ N, are (yk , λk ) where yk (x) = sin(kx) and λk = k 2 .
Theorem 11.0.2. For a periodic S-L problem, there exists an increasing

sequence of eigenvalues 0 < λ1 < λ2 < λ3 < . . . < λk < . . . with λk → ∞,
as k → ∞. Moreover, W1 = Wλ1 , the eigen space corresponding to the first
eigen value is one dimensional.
Example 11.2. Consider the boundary value problem,

 00
 y + λy = 0 in (−π, π)
y(−π) = y(π)
 0
y (−π) = y 0 (π).
√
The characteristic equation is m2 +λ = 0. Solving for m, we get m = ± −λ.
Note that the λ can be either zero, positive or negative.
If λ = 0, then y 00 = 0 and the general solution is y(x) = αx + β, for some
constants α and β. Since y(−π) = y(π), we get α = 0. Thus, for λ = 0, y ≡
a constant is the only non-trivial solution. √ √
If λ < 0, then ω = −λ > 0. Hence y(x) = αe ωx + βe− ωx . Using
the boundary condition y(−π) = y(π), we get α = β and using the other
boundary condition, we get α = β = 0. Hence we have no non-trivial solution
corresponding to negative λ’s.
√ √ √
If λ > 0, then m = ±i λ and y(x) = α cos( λx) + β sin( λx). Using
the boundary condition, we get
√ √ √ √
α cos(− λπ) + β sin(− λπ) = α cos( λπ) + β sin( λπ)
and
√ √ √ √
−α sin(− λπ) + β cos(− λπ) = −α sin( λπ) + β cos( λπ).
√ √
Thus, β sin( λπ) = α sin( λπ) = 0. √
For a non-trivial solution, we must have sin( λπ) = 0. Thus, λ = k 2 for
each non-zero k ∈ N (since λ > 0).
Hence, for each k ∈ N ∪ {0}, there is a solution (yk , λk ) with
yk (x) = αk cos kx + βk sin kx,
and λk = k 2 .
Lecture 12
Singular Sturm-Liouville
Problem
Singular S-L, in general, have continuous spectrum. However, the examples

we presented viz. Bessel’s equation and Legendre equation have a discrete
spectrum, similar to the regular S-L problem.
12.0.1 EVP of Legendre Operator

Consider the Legendre equation

d 2 dy
(1 − x ) + λy = 0 for x ∈ [−1, 1].
dx dx
Note that, equivalently, we have the form
(1 − x2 )y 00 − 2xy 0 + λy = 0 for x ∈ [−1, 1].
The function p(x) = 1 − x2 vanishes at the endpoints x = ±1.
Definition 12.0.1. A point x0 is a singular point of
y 00 (x) + P (x)y 0 (x) + Q(x)y(x) = R(x)
if either P or Q (or both) are not analytic at x0 . A singular point x0 is said

to be regular if (x − x0 )P (x) and (x − x0 )2 Q(x) are analytic at x0 .
59
LECTURE 12. SINGULAR STURM-LIOUVILLE PROBLEM 60
The end points x = ±1 are regular singular point. The coefficients P (x) =
−2x λ
1−x2
and R(x) = 1−x 2 are analytic at x = 0, with radius of convergence 1.
We look for power series form of solutions y(x) = ∞ k

P
k=0 ak x . Differentiating
(twice) the series term by term, substituting in the Legendre equation and
equating like powers of x, we get a2 = −λa2
0
, a3 = (2−λ)a
6
1
and for k ≥ 2,
(k(k + 1) − λ)ak
ak+2 = .
(k + 2)(k + 1)
Thus, the constants a0 and a1 can be fixed arbitrarily and the remaining
constants are defined as per the above relation. For instance, if a1 = 0, we
get the non-trivial solution of the Legendre equation as
∞
X
y 1 = a0 + a2k x2k
k=1
and if a0 = 0, we get the non-trivial solution as

∞
X
y2 = a1 x + a2k+1 x2k+1 ,
k=1
provided the series converge. Note from the recurrence relation that if a
coefficient is zero at some stage, then every alternate coefficient, subsequently,
is zero. Thus, there are two possibilities of convergence here:
(i) the series terminates after finite stage to become a polynomial
(ii) the series does not terminate, but converges.
Suppose the series does not terminate, say for instance, in y1 . Then
a2k 6= 0, for all k. Consider the ratio
a2(k+1) x2(k+1) 2k(2k + 1)x2 2kx2

lim = x2 .
= k→∞
lim = lim

k→∞ a2k x2k (2k + 2)(2k + 1) k→∞ (2k + 2)
The term involving λ tends to zero. Therefore, by ratio test, y1 converges in
x2 < 1 and diverges in x2 > 1. Also, it can be shown that when x2 = 1 the
series diverges (beyond the scope of this course).
Since, Legendre equation is a singular S-L problem, we try to find solution
y such that y and its derivative y 0 are continuous in the closed interval [−1, 1].
Thus, the only such possible solutions will be terminating series becoming
polynomials.
Note that, for k ≥ 2,
(k(k + 1) − λ)ak
ak+2 = .
(k + 2)(k + 1)
Hence, for any n ≥ 2, if λ = n(n+1), then an+2 = 0 and hence every alternate
term is zero. Also, if λ = 1(1 + 1) = 2, then a3 = 0. If λ = 0(0 + 1) = 0, then
a2 = 0. Thus, for each n ∈ N ∪ {0}, we have λn = n(n + 1) and one of the
solution y1 or y2 is a polynomial. Thus, for each n ∈ N ∪ {0}, we have the
eigen value λn = n(n + 1) and the Legendre polynmial Pn of degree n which
is a solution to the Legendre equation.
12.0.2 EVP of Bessel’s Operator

Consider the EVP, for each fixed n = 0, 1, 2, . . .,
( 2
− (xy 0 (x))0 = − nx + λx y(x) x ∈ (0, a)
y(a) = 0.
As before, since this is a singular S-L problem we shall look for solutions y
such that y and its derivative y 0 are continuous in the closed interval [0, a].
We shall assume that the eigenvalues are all real1 . Thus, λ may be zero,
positive or negative.
When λ = 0, the given ODE reduces to the Cauchy-Euler form
0 n2
− (xy 0 (x)) + y(x) = 0
x
or equivalently,
x2 y 00 (x) + xy 0 (x) − n2 y(x) = 0.
The above second order ODE with variable coefficients can be converted to
an ODE with constant coefficients by the substitution x = es (or s = ln x).
Then, by chain rule,
dy dy ds dy
y0 = = = e−s
dx ds dx ds
1
needs proof
and
0
d2 y dy

00 −s dy −s d −s dy −2s
y =e =e e =e − .
ds ds ds ds2 ds
Therefore,
y 00 (s) − n2 y(s) = 0,
where y is now a function of the new variable s. For n = 0, the general
solution is y(s) = αs+β, for some arbitrary constants. Thus, y(x) = α ln x+
β. The requirement that both y and y 0 are continuous on [0, a] forces α = 0.
Thus, y(x) = β. But y(a) = 0 and hence β = 0, yielding the trivial solution.
Now, let n > 0 be positive integers. Then the general solution is y(s) =
αens + βe−ns . Consequently, y(x) = αxn + βx−n . Since y and y 0 has to be
continuous on [0, a], β = 0. Thus, y(x) = αxn . Now, using the boundary
condition y(a) = 0, we get α = 0 yielding the trivial solution. Therefore,
λ = 0 is not an eigenvalue for all n = 0, 1, 2, . . ..
When λ > 0, the given ODE reduces to
x2 y 00 (x) + xy 0 (x) + (λx2 − n2 )y(x) = 0.
√
Using the change of variable s2 = λx2 , we get y 0 (x) = λy 0 (s) and y 00 (x) =
λy 00 (s). Then the given ODE is transformed into the Bessel’s equation
s2 y 00 (s) + sy 0 (s) + (s2 − n2 )y(s) = 0.
Using the power series form of solution, we know that the general solution
of the Bessel’s equation is
y(s) = αJn (s) + βYn (s),
where Jn and Yn are the Bessel
√ functions
√ of first and second kind, respectively.
Therefore, y(x) = αJn ( λx) + βYn ( λx). √ The continuity assumptions of
0
y and y force that√β = 0, because Yn ( λx) is discontinuous at x = 0.
√ y(x) = αJn ( λx). Using the boundary condition y(a) = 0, we get
Thus,
Jn ( λa) = 0.
Theorem 12.0.2. For each non-negative integer n, Jn has infinitely many
positive zeroes.
√ For each n ∈ N ∪ {0}, let 2

znm be the m-th zero of Jn , m ∈ N. Hence
λa = znm and so λnm = znm /a2 and the corresponding eigen functions are
ynm (x) = Jn (znm x/a).
For λ < 0, there are no eigen values. Observing this fact is beyond the
scope of this course, hence we assume this fact.
Lecture 13
Orthogonality of Eigen
Functions
Observe that for a regular S-L problem the differential operator can be writ-
ten as
−1 d d q(x)
L= p(x) − .
r(x) dx dx r(x)
Let V denote the set of all solutions of (10.2.1). Necessarily, 0 ∈ V and
V ⊂ C 2 (a, b). We define the inner product 1 h·, ·i : V × V → R on V as,
Z b
hf, gi := r(x)f (x)g(x) dx.
a
Definition 13.0.1. We say two functions f and g are perpendicular or

p with weight r if hf, gi = 0. We say f is of unit length if its norm
orthogonal
kf k = hf, f i = 1.
Theorem 13.0.2. With respect to the inner product defined above in V , the
eigen functions corresponding to distinct eigenvalues of the S-L problem are
orthogonal.
Proof. Let yi and yj are eigen functions corresponding to distinct eigenvalues

λi and λj . We need to show that hyi , yj i = 0. Recall that L is the S-L operator
1
a generalisation of the usual scalar product of vectors
63
LECTURE 13. ORTHOGONALITY OF EIGEN FUNCTIONS 64
and hence Lyk = λk yk , for k = i, j. Consider

Z b
λi hyi , yj i = hLyi , yj i = rLyi yj dx
a
Z b Z b
d dyi
= − p(x) yj (x) − q(x)yi yj dx
a dx dx a
Z b
dyi dyj (x)
= p(x) dx − [p(b)yi0 (b)yj (b) − p(a)yi0 (a)yj (b)]
a dx dx
Z b
− q(x)yi yj dx
a
Z b
d dyj
+ p(b)yj0 (b)yi (b) − p(a)yj0 (a)yi (b)

= − yi (x) p(x)
a dx dx
Z b
0 0
− [p(b)yi (b)yj (b) − p(a)yi (a)yj (b)] − q(x)yi yj dx
a
= hyi , Lyj i + p(b) yj0 (b)yi (b) − yi0 (b)yj (b)

−p(a) yj0 (a)yi (a) − yi0 (a)yj (a)

= λj hyi , yj i + p(b) yj0 (b)yi (b) − yi0 (b)yj (b)

−p(a) yj0 (a)yi (a) − yi0 (a)yj (a) .

Thus,
(λi −λj )hyi , yj i = p(b) yj0 (b)yi (b) − yi0 (b)yj (b) −p(a) yj0 (a)yi (a) − yi0 (a)yj (a) .

For regular S-L problem, the boundary condition corresponding to the end-
point b is the system of equations
d1 yi (b) + d2 yi0 (b) = 0
d1 yj (b) + d2 yj0 (b) = 0
such that d21 + d22 = 0. Therefore, the determinant of the coefficient matrix
yi (b)yj0 (b) − yj (b)yi0 (b) = 0. Similar, argument is also valid for the boundary
condition corresponding to a. Thus, (λi − λj )hyi , yj i = 0. But λi − λj 6= 0,
hence hyi , yj i = 0.
For periodic S-L problem, p(a) = p(b), yk (a) = yk (b) and yk0 (a) = yk0 (b),
for k = i, j. Then the RHS vanishes and hyi , yj i = 0.
For singular S-L problems such that either p(a) = 0 or p(b) = 0 or both
happens, then again RHS vanishes. This is because if p(a) = 0, we drop the
boundary condition corresponding to the end-point a.
Let us examine the orthogonality of the eigenvectors computed in the

examples earlier.
Example 13.1. We computed in Example 11.1 the eigenvalues and eigenvec-
tors of the regular S-L problem,
y 00 + λy = 0 x ∈ (0, a)

y(0) = y(a) = 0
to be (yk , λk ) where
kπx
yk (x) = sin
a
and λk = (kπ/a)2 , for each k ∈ N. For m, n ∈ N such that m 6= n, we need
to check that ym and yn are orthogonal. Since r ≡ 1, we consider
Z a mπx nπx
hym (x), yn (x)i = sin sin dx
0 a a
Exercise 5. Show that, for any n ≥ 0 and m positive integer,
(i) (
Z π
π, for m = n
cos nt cos mt dt =
−π 0, for m 6= n.
(ii) (
Z π
π, for m = n
sin nt sin mt dt =
−π 0, for m 6= n.
(iii) Z π
sin nt cos mt dt = 0.
−π
cos
√ kt sin
√ kt
Consequently, show that π
and π
are of unit length.
Proof. (i) Consider the trigonometric identities
cos((n + m)t) = cos nt cos mt − sin nt sin mt (13.0.1)
and
cos((n − m)t) = cos nt cos mt + sin nt sin mt. (13.0.2)
Adding (13.0.1) and (13.0.2), we get

1
cos((n + m)t) + cos((n − m)t) = cos nt cos mt.
2
Integrating both sides from −π to π, we get
Z π
1 π
Z
cos nt cos mt dt = (cos((n + m)t) + cos((n − m)t)) dt.
−π 2 −π
But Z π
1
cos kt dt = sin kt|π−π = 0, for k 6= 0.
−π k
Thus, (
Z π
π, for m = n
cos nt cos mt dt =
−π 0, for m 6= n.
Further, √
k cos ktk = hcos kt, cos kti1/2 = π.
cos
√ kt
Therefore, π
is of unit length.
(ii) Subtract (13.0.1) from (13.0.2) and use similar arguments as above.
(iii) Arguments are same using the identities (13.0.1) and (13.0.2) corre-
sponding to sin.
Exercise 6. Show that

Z π (
2π, for m = n
eimt e−int dt =
−π 0, for m 6= n.
Example 13.2. We computed in Example 11.2 the eigenvalues and eigenvec-
tors of the periodic S-L problem,
 00
 y + λy = 0 in (−π, π)
y(−π) = y(π)
 0
y (−π) = y 0 (π)
to be, for each k ∈ N ∪ {0}, (yk , λk ) where
yk (x) = αk cos kx + βk sin kx,
and λk = k 2 . Again r ≡ 1 and the orthogonality follows from the exercise
above.
Example 13.3. The orthogonality of Legendre polynomial and Bessel function

must have been discussed in your course on ODE. Recall that the Legendre
polynomials has the property
Z 1 (
0, if m 6= n
Pm (x)Pn (x) dx = 2
−1 2n+1
, if m = n
and the Bessel functions have the property

Z 1 (
0, if m 6= n
xJn (zni x)Jn (znj x) dx = 1
0 [J (z )]2 ,
2 n+1 ni
if m = n
where zni is the i-th positive zero of the Bessel function (of order n) Jn .
13.1 Eigen Function Expansion

Observe that an eigenvector yk , for any k, can be normalised (unit norm) in
its inner-product by dividing yk by its norm kyk k. Thus, yk /kyk k, for any
k, is a unit vector. For instance, in view of Exercise 5, cos
√ kt and sin
π
√ kt are
π
functions of unit length.
Definition 13.1.1. Any given function f : (a, b) → R is said to be have the
eigen function expansion corresponding to the S-L problem (10.2.1), if
∞
X
f (x) ≈ ak y k ,
k=0
for some constants ak and yk are the normalised eigenvectors corresponding

to (10.2.1).
We are using the “≈” symbol to highlight the fact that the issue of con-
vergence of the series is ignored.
If the eigenvectors (or eigen functions) yk involves only sin or cos terms,
as in regular S-L problem (cf. Example 11.1), then the series is called Fourier
Sine or Fourier Cosine series.
If the eigen functions yk involve both sin and cos, as in periodic S-L
problem (cf. Example 11.2), then the series is called Fourier series. In the
case of the eigen functions being Legendre polynomial or Bessel function, we
call it Fourier-Legendre or Fourier-Bessel series, respectively.
Lecture 14
Fourier Series
At the end of previous chapter, we introduced the Fourier series of a function

f . A natural question that arises at this moment is: what classes of functions
admit a Fourier series expansion? We attempt to answer this question in this
chapter.
14.1 Periodic Functions

We isolate the properties of the trigonometric functions, viz., sin, cos, tan
etc.
Definition 14.1.1. A function f : R → R is said to be periodic of period
T , if T > 0 is the smallest number such that
f (t + T ) = f (t) ∀t ∈ R.
Such functions are also called T -periodic functions.
Example 14.1. The trigonometric functions sin t and cos t are 2π-periodic
functions, while sin 2t and cos 2t are π-periodic functions.
Given a L-periodic real-valued function g on R, one can always construct
a T -periodic function as: f (t) = g(Lt/T ). For instance, f (t) = sin 2πt

T
is a
T -periodic function.

2π(t + T ) 2πt 2πt
sin = sin + 2π = sin .
T T T
In fact, for any positive integer k, sin 2πkt and cos 2πkt

T T
are T -periodic
functions.
69
LECTURE 14. FOURIER SERIES 70
Exercise 7. If f : R → R is a T -periodic function, then show that

(i) f (t − T ) = f (t), for all t ∈ R.
(ii) f (t + kT ) = f (t), for all k ∈ Z.
(iii) g(t) = f (αt + β) is (T /α)-periodic, where α > 0 and β ∈ R.
Exercise 8. Show that for a T -periodic integrable function f : R → R,
Z α+T Z T
f (t) dt = f (t) dt ∀α ∈ R.
α 0
14.2 Fourier Coefficients and Fourier Series

Without loss of generality, to simplify our computation, let us assume that
f is a 2π-periodic1 function on R. Suppose that f : (−π, π) → R, extended
to all of R as a 2π-periodic function, is such that the infinite series
∞
X
a0 + (ak cos kt + bk sin kt)
k=1
converges uniformly2 to f . Then,

∞
X
f (t) = a0 + (ak cos kt + bk sin kt). (14.2.1)
k=1
and integrating both sides of (14.2.1), from −π to π, we get

∞
Z π Z π !
X
f (t) dt = a0 + (ak cos kt + bk sin kt) dt
−π −π k=1
∞
!
Z π X
= a0 (2π) + (ak cos kt + bk sin kt) dt
−π k=1
Since the series converges uniformly to f , the interchange of integral and

series is possible. Therefore,
Z π ∞ Z π
X
f (t) dt = a0 (2π) + (ak cos kt + bk sin kt) dt
−π k=1 −π
1
similar idea will work for any T -periodic function
2
note the uniform convergence hypothesis
From Exercise 5, we know that

Z π Z π
sin kt dt = cos kt dt = 0, ∀k ∈ N.
−π −π
Hence, Z π
1
a0 = f (t) dt.
2π −π
To find the coefficients ak , for each fixed k ∈ N, we multiply both sides

of (14.2.1) by cos kt and integrate from −π to π. Consequently,
Z π Z π
f (t) cos kt dt = a0 cos kt dt
−π −π
∞ Z
X π
+ (aj cos jt cos kt + bj sin jt cos kt) dt
j=1 −π
Z π
= ak cos kt cos kt dt = πak .
−π
Similar argument, after multiplying by sin kt, gives the formula for bk . Thus,
we have derived , for all k ∈ N,
1 π
Z
ak = f (t) cos kt dt
π −π
1 π
Z
bk = f (t) sin kt dt
π −π
Z π
1
a0 = f (t) dt.
2π −π
These are the formulae for Fourier coefficients of a 2π-periodic functions f ,

in terms of f . Similarly, if f is a T -periodic function extended to R, then its
Fourier series is
∞
X 2πkt 2πkt
f (t) = a0 + ak cos + bk sin ,
k=1
T T
where Z T
2 2πkt
ak = f (t) cos dt (14.2.2a)
T 0 T
Z T
2 2πkt
bk = f (t) sin dt (14.2.2b)
T 0 T
1 T
Z
a0 = f (t) dt. (14.2.2c)
T 0
The above discussion motivates us to give the following definition.
Definition 14.2.1. If f : R → R is any T -periodic integrable function then
we define the Fourier coefficients of f , a0 , ak and bk , for all k ∈ N, by (14.2.2)
and the Fourier series of f is given by
∞
X 2πkt 2πkt
f (x) ≈ a0 + ak cos + bk sin . (14.2.3)
k=1
T T
Note the use of “≈” symbol in (14.2.3). This is because we have the
following issues once we have the definition of Fourier series of f , viz.,
(a) Will the Fourier series of f always converge?
(b) If it converges, will it converge to f ?
(c) If so, is the convergence point-wise or uniform3 .
Answering these question, in all generality, is beyond the scope of this
course. However, we shall state some results, in the next section, that will
get us in to working mode. We end this section with some simple examples
on computing Fourier coefficients of functions.
Example 14.2. Consider the constant function f ≡ c on (−π, π). Then
Z π
1
a0 = c dt = c.
2π −π
For each k ∈ N, Z π
1
ak = c cos kt dt = 0
π −π
and Z π
1
bk = c sin kt dt = 0.
π −π
3
because our derivation of formulae for Fourier coefficients assumed uniform conver-
gence of the series
Example 14.3. Consider the trigonometric function f (t) = sin t on (−π, π).
Then Z π
1
a0 = sin x dt = 0.
2π −π
1
ak = sin t cos kt dt = 0
π −π
and (
π
0 k 6= 1
Z
1
bk = sin t sin kt dt =
π −π 1 k = 1.
Similarly, for f (t) = cos t on (−π, π), all Fourier coefficients are zero, except
a1 = 1.
Example 14.4. Consider the function f (t) = t on (−π, π). Then
Z π
1
a0 = t dt = 0.
2π −π
For each k ∈ N,
1 π
Z Z π
1
ak = t cos kt dt = − sin kt dt + (π sin kπ − (−π) sin k(−π))
π −π kπ −π
and hence ak = 0, for all k.

1 π
Z Z π
1
bk = t sin kt dt = cos kt dt − (π cos kπ − (−π) cos k(−π))
π −π kπ −π
1 (−1)k+1 2
= 0 − π(−1)k + π(−1)k =
kπ k
Therefore, t as a 2π-periodic function defined in (−π, π) has the Fourier series
expansion
∞
X (−1)k+1
t≈2 sin kt
k=1
k
Example 14.5. Let us consider the same function f (t) = t, as in previous
example, but defined on (0, π). Viewing this as π-periodic function, we com-
pute
1 π
Z
π
a0 = t dt = .
π 0 2
For each k ∈ N,
2 π
Z Z π
2
ak = t cos 2kt dt = − sin 2kt dt + (π sin 2kπ − 0)
π 0 2kπ 0

1 1
= (cos 2kπ − cos(0)) = 0
kπ 2k
and
2 π
Z Z π
2
bk = t sin 2kt dt = cos 2kt dt − (π cos 2kπ − 0)
π 0 2kπ 0

1 1 −1
= (sin 2kπ − sin(0)) − π = .
kπ 2k k
Therefore, t as a π-periodic function defined on (0, π) has the Fourier series

expansion
∞
π X1
t≈ − sin 2kt.
2 k=1 k
Note that difference in Fourier expansion of the same function when the
periodicity changes.
Exercise 9. Find the Fourier coefficients and Fourier series of the function
(
0 if t ∈ (−π, 0]
f (t) =
t if t ∈ (0, π).
Theorem 14.2.2 (Riemann-Lebesgue Lemma). Let f be a continuous func-

tion in [−π, π]. Show that the Fourier coefficients of f converges to zero,
i.e.,
lim ak = lim bk = 0.
k→∞ k→∞
Proof. Observe that |ak | and |bk | are bounded sequences, since
Z π
{|ak |, |bk |} ≤ |f (t)| dt < +∞.
−π
We need to show that these bounded sequences, in fact, converges to zero.

Now set x = t − π/k and hence
Z π Z π−π/k
bk = f (t) sin kt dt = f (x + π/k) sin(kx + π) dx
−π −π−π/k
Z π−π/k
= − f (x + π/k) sin kx dx.
−π−π/k
Therefore, after reassigning x as t,

Z π Z π−π/k
2bk = f (t) sin kt dt − f (t + π/k) sin kt dt
−π −π−π/k
Z −π Z π−π/k
= − f (t + π/k) sin kt dt + (f (t) − f (t + π/k)) sin kt dt
−π−π/k −π
Z π
+ f (t) sin kt dt
π−π/k
= I1 + I2 + I3 .
Thus, |2bk | ≤ |I1 | + |I2 | + |I3 |. Consider

Z π

|I3 | = f (t) sin kt dt
π−π/k
Z π
≤ |f (t)| dt
π−π/k

π Mπ
≤ max |f (t)| = .
t∈[−π,π] k k
Similar estimate is also true for I1 . Let us consider,
Z
π−π/k
|I2 | = (f (t) − f (t + π/k)) sin kt dt

−π

π
≤ max |f (t) − f (t + π/k)| 2π −
t∈[−π,π−π/k] k
By the uniform continuity of f on [−π, π], the maximum will tend to zero as
k → ∞. Hence |bk | → 0. Exactly, similar arguments hold for ak .
Lecture 15
Fourier Series: Continued...
15.1 Piecewise Smooth Functions

Definition 15.1.1. A function f : [a, b] → R is said to be piecewise contin-
uously differentiable if it has a continuous derivative f 0 in (a, b), except at
finitely many points in the interval [a, b] and at each these finite points, the
right-hand and left-hand limit for both f and f 0 exist.
Example 15.1. Consider f : [−1, 1] → R defined as f (t) = |t| is continuous.
It is not differentiable at 0, but it is piecewise continuously differentiable.
Example 15.2. Consider the function f : [−1, 1] → R defined as

−1, for − 1 < t < 0,

f (t) = 1, for 0 < t < 1,

0, for t = 0, 1, −1.

It is not continuous, but is piecewise continuous. It is also piecewise contin-

uously differentiable.
Exercise 10 (Riemann-Lebesgue Lemma). Let f be a piecewise continuous
function in [−π, π] such that
Z π
|f (t)| dt < +∞.
−π
Show that the Fourier coefficients of f converges to zero, i.e.,

lim ak = lim bk = 0.
k→∞ k→∞
77
LECTURE 15. FOURIER SERIES: CONTINUED... 78
Theorem 15.1.2. If f is a T -periodic piecewise continuously differentiable

function, then the Fourier series of f converges to f (t), for every t at which
f is smooth. Further, at a non-smooth point t0 , the Fourier series of f will
converge to the average of the right and left limits of f at t0 .
Corollary 15.1.3. If f : R → R is a continuously differentiable (derivative
f 0 exists and is continuous) T -periodic function, then the Fourier series of f
converges to f (t), for every t ∈ R.
Example 15.3. For a given constant c 6= 0, consider the piecewise function
(
0 if t ∈ (−π, 0)
f (t) =
c if t ∈ (0, π).
Then, Z π
1 c
a0 = c dt = .
2π 0 2
1
ak = c cos kt dt = 0
π 0
and
π
c(1 + (−1)k+1 )
Z
1 c 1
bk = c sin kt dt = (− cos kπ + cos(0)) = .
π 0 π k kπ
Therefore,
∞
c X c(1 + (−1)k+1 )
f (t) ≈ + sin kt.
2 k=1 kπ
The point t0 = 0 is a non-smooth point of the function f . Note that the
right limit of f at t0 = 0 is c and the left limit of f at t0 = 0 is 0. Note that
the Fourier series of f ay t0 = 0 converges to c/2, the average of c and 0.
15.2 Complex Fourier Coefficients

The Fourier series of a 2π-periodic function f : R → R as given in (14.2.1),
can be recast in complex number notation using the formulae
eıt + e−ıt eıt − e−ıt
cos t = , sin t = .
2 2i
Note that we can rewrite the Fourier series expansion of f as

∞
a0 X
f (t) = + (ak cos kt + bk sin kt)
2 k=1
with a factor 2 in denominator of a0 and make the formulae of the Fourier

coefficient having uniform factor. Thus,
∞
a0 X
f (t) = + (ak cos kt + bk sin kt)
2 k=1
∞
a0 X ak ıkt −ıkt
ıbk ıkt −ıkt

= + e +e − e −e
2 k=1
2 2
∞
a0 X ak − ıbk ıkt ak + ıbk −ıkt
f (t) = + e + e
2 k=1
2 2
∞
X
ck eıkt + c−k e−ıkt

= c0 +
k=1
∞
X
= ck eıkt .
k=−∞
Exercise 11. Given a 2π-periodic function f such that

∞
X
f (t) = ck eıkt , (15.2.1)
k=−∞
where the convergence is uniform. Use the integral formulae from Exercise 6
to show that, for all k ∈ Z,
Z π
1
ck = f (t)e−ıkt dt.
2π −π
Proof. Fix a k. To find the coefficient ck , multiply both sides of (15.2.1) by
e−ıkt and integrate from −π to π.
Using the real Fourier coefficients one can write down the complex Fourier
coefficients using the relations
a0 ak − ıbk ak + ıbk
c0 = , ck = and c−k =
2 2 2
and if one can compute directly the complex Fourier series of a periodic
function f , then one can write down the real Fourier coefficients using the
formula,
a0 = 2c0 , ak = ck + c−k and bk = ı(ck − c−k ).
Exercise 12. Find the complex Fourier coefficients (directly) of the function
f (t) = t for t ∈ (−π, π] extended to R periodically with period 2π. Use the
complex Fourier coefficients to find the real Fourier coefficients of f .
Proof. We use the Z π

1
ck = te−ıkt dt.
2π −π
for all k = 0, ±1, ±2, . . .. For k = 0, we get c0 = 0. For k 6= 0, we apply

integration by parts to get
π Z π
1 ıt −ıkt ı −ıkt
ck = e − e dt
2π k −π −π k
1 h ıπ −ıkπ ıπ ıkπ i ı ı
= e + e = cos kπ = (−1)k .
2π k k k k
Hence the real Fourier coefficients are
(−1)k+1
a0 = 0, ak = (ck + c−k ) = 0 and bk = ı(ck − c−k ) = 2 .
k
15.3 Orthogonality
Let V be the class of all 2π-periodic real valued continuous function on R.
Exercise 13. Show that V is a vector space over R.
We introduce an inner product on V . For any two elements f, g ∈ V , we
define: Z π
hf, gi := f (t)g(t) dt.
−π
The inner product generalises to V the properties of scalar product on Rn .

Exercise 14. Show that the inner product defined on V as

Z π
hf, gi = f (t)g(t) dt
−π
satisfies the properties of a scalar product in Rn , viz., for all f, g ∈ V ,
(a) hf, gi = hg, f i.
(b) hf + g, hi = hf, hi + hg, hi.
(c) hαf, gi = αhf, gi ∀α ∈ R.
(d) hf, f i ≥ 0 and hf, f i = 0 implies that f ≡ 0.
Definition 15.3.1. We say two functions f and g are perpendicular or

orthogonal
p if hf, gi = 0. We say f is of unit length if its norm kf k =
hf, f i = 1.
Consider, for k ∈ N, the following elements in V
1 cos kt sin kt
e0 (t) = √ , ek (t) = √ and fk (t) = √ .
2π π π
Example 15.4. e0 , ek and fk are all of unit length. he0 , ek i = 0 and he0 , fk i =
0. Also, hem , en i = 0 and hfm , fn i = 0, for m 6= n. Further, hem , fn i = 0
for all m, n. Check and compare these properties with the standard basis
vectors of Rn !
In this new formulation, we can rewrite the formulae for the Fourier
coefficients as:
1 1 1
a0 = √ hf, e0 i, ak = √ hf, ek i and bk = √ hf, fk i.
2π π π
and the Fourier series of f has the form,

∞
1 1 X
f (t) = hf, e0 i √ + √ (hf, ek i cos kt + hf, fk i sin kt) .
2π π k=1
15.3.1 Odd and Even functions

Definition 15.3.2. We say a function f : R → R is odd if f (−t) = −f (t)
and even if f (−t) = f (t).
Example 15.5. All constant functions are even functions. For all k ∈ N, sin kt
are odd functions and cos kt are even functions.
Exercise 15. Any odd function is always orthogonal to an even function.
The Fourier series of an odd or even functions will contain only sine or
cosine parts, respectively. The reason being that, if f is odd
hf, 1i = 0 and hf, cos kti = 0
and hence a0 = 0 and ak = 0, for all k. If f is even
hf, sin kti = 0
and bk = 0, for all k.
15.4 Fourier Sine-Cosine Series

Let f : (0, T ) → R be a piecewise smooth function. To compute the Fourier
Sine series of f , we extend f , as an odd function fo , to (−T, T )
(
f (t), for t ∈ (0, T )
fo (t) =
−f (−t) , for t ∈ (−T, 0).
Note that fo is a 2T -periodic function and is an odd function. Since fo is

odd, the cosine coefficients ak and the constant term a0 vanishes in Fourier
series expansion of fo . The restriction of the Fourier series of fo to f in the
interval (0, T ) gives the Fourier sine series of f . We derive the formulae for
Fourier sine coefficient of f .
∞
X πkt
f (t) = bk sin where (15.4.1)
k=1
T
1 T
Z
1 πkt πkt
bk = fo , sin = fo (t) sin dt
T T T −T T
Z 0 Z T
1 πkt πkt
= −f (−t) sin dt + f (t) sin dt
T −T T 0 T
Z 0 Z T
1 πkt πkt
= −f (t) sin dt + f (t) sin dt
T T T 0 T
2 T
Z
πkt
= f (t) sin dt.
T 0 T
Example 15.6. Let us consider the function f (t) = t on (0, π). To compute the
Fourier sine series of f , we extend f to (−π, π) as an odd function fo (t) = t
on (−π, π). For each k ∈ N,
2 π
Z Z π
2
bk = t sin kt dt = cos kt dt − (π cos kπ − 0)
π 0 kπ 0
(−1)k+1 2

2 1 k+1
= (sin kπ − sin(0)) + π(−1) = .
kπ k k
Therefore, the Fourier sine series expansion of f (t) = t on (0, π) is

∞
X (−1)k+1
t≈2 sin kt
k=1
k
Compare the result with Example 14.4.

For computing the Fourier cosine series of f , we extend f as an even
function to (−T, T ),
(
f (t), for t ∈ (0, T )
fe (t) =
f (−t) , for t ∈ (−T, 0).
The function fe is a 2T -periodic function extended to all of R. The Fourier

series of fe has no sine coefficients, bk = 0 for all k. The restriction of the
Fourier series of fe to f in the interval (0, T ) gives the Fourier cosine series
of f . We derive the formulae for Fourier cosine coefficient of f .
∞
X πkt
f (t) = a0 + ak cos (15.4.2)
k=1
T
where Z T
2 πkt
ak = f (t) cos dt
T 0 T
and
1 T
Z
a0 = f (t) dt.
T 0
Example 15.7. Let us consider the function f (t) = t on (0, π). To compute
the Fourier cosine series of f , we extend f to (−π, π) as an even function
fe (t) = |t| on (−π, π). Then,
1 π
Z
π
a0 = t dt = .
π 0 2
For each k ∈ N,
2 π
Z Z π
2
ak = t cos kt dt = − sin kt dt + (π sin kπ − 0)
π 0 kπ 0
2[(−1)k − 1]

2 1
= (cos kπ − cos(0)) = .
kπ k k2π
Therefore, the Fourier cosine series expansion of f (t) = t on (0, π) is
∞
π X (−1)k − 1
t≈ +2 2π
cos kt.
2 k=1
k
Compare the result with the Fourier series of the function f (t) = |t| on
(−π, π).
15.5 Fourier Transform and Integral

Recall that we had computed the Fourier series expansion of periodic func-
tions. The periodicity was assumed due to the periodicity of sin and cos
functions. The question we shall address in this section is: Can we generalise
the notion of Fourier series of f , to non-periodic functions?
The answer is a “yes”! Note that the periodicity of f is captured by the
integer k appearing in the arguments of sin and cos. To generalise the notion
of Fourier series to non-periodic functions, we shall replace k, a positive
integer, with a real number ξ. Note that when we replace k with ξ, the
sequences ak , bk become functions of ξ, a(ξ) and b(ξ) and the series form is
replaced by an integral form over R.
Definition 15.5.1. If f : R → R is a piecewise continuous function which

vanishes outside a finite interval, then its Fourier integral is defined as
Z ∞
f (t) = (a(ξ) cos ξt + b(ξ) sin ξt) dξ,
0
where
1 ∞
Z
a(ξ) = f (t) cos ξt dt
π −∞
1 ∞
Z
b(ξ) = f (t) sin ξt dt.
π −∞
Lecture 16
Standing Waves: Separation of

Variable
The method of separation of variables was introduced by d’Alembert (1747)

and Euler (1748) for the wave equation. This technique was also employed
by Laplace (1782) and Legendre (1782) while studying the Laplace equation
and also by Fourier while studying the heat equation.
Recall the set-up of the vibrating string given by the equation utt = uxx ,
we have normalised the constant c. Initially at time t, let us say the string
has the shape of the graph of v, i.e., u(x, 0) = v(x). The snapshot of the
vibrating string at each time are called the “standing waves”. The shape of
the string at time t0 can be thought of as some factor (depending on time)
of v. This observation motivates the idea of “separation of variable”, i.e.,
u(x, t) = v(x)w(t), where w(t) is the factor depending on time, which scales
v at time t to fit with the shape of u(x, t).
The fact that endpoints are fixed is given by the boundary condition
u(0, t) = u(L, t) = 0.
We are also given the initial position u(x, 0) = g(x) (at time t = 0) and initial
velocity of the string at time t = 0, ut (x, 0) = h(x). Given g, h : [0, L] → R
such that g(0) = g(L) = 0 and h(0) = h(L), we need to solve the initial value
87
LECTURE 16. STANDING WAVES: SEPARATION OF VARIABLE 88
problem


 utt (x, t) − c2 uxx (x, t) =0 in (0, L) × (0, ∞)
u(x, 0) = g(x) in [0, L]



ut (x, 0) = h(x) in [0, L] (16.0.1)
u(0, t) = φ(t) in (0, ∞)




u(L, t) = ψ(t) in (0, ∞),

where φ, ψ, g, h satisfies the compatibility condition
g(0) = φ(0), g 00 (0) = φ00 (0), h(0) = φ0 (0)
and
g(L) = ψ(0), g 00 (L) = ψ 00 (0), h(L) = ψ 0 (0).
Let φ = ψ ≡ 0. Let us seek for solutions u(x, t) whose variables can be
separated. Let u(x, t) = v(x)w(t). Differentiating and substituting in the
wave equation, we get
v(x)w00 (t) = c2 v 00 (x)w(t)
Hence
w00 (t) v 00 (x)
= .
c2 w(t) v(x)
Since RHS is a function of x and LHS is a function t, they must equal a
constant, say λ. Thus,
v 00 (x) w00 (t)
= 2 = λ.
v(x) c w(t)
Using the boundary condition u(0, t) = u(L, t) = 0, we get
v(0)w(t) = v(L)w(t) = 0.
If w ≡ 0, then u ≡ 0 and this cannot be a solution to (16.0.1). Hence, w 6≡ 0

and v(0) = v(L) = 0. Thus, we need to solve the eigen value problem for the
second order differential operator.
00
v (x) = λv(x), x ∈ (0, L)
v(0) = v(L) = 0,
Note that the λ can be either zero, positive or negative. If λ = 0, then

00
v = 0 and the general solution is v(x) = αx + β, for some constants α and
β. Since v(0) = 0, we get β = 0, and v(L) = 0 and L 6= 0 implies that α = 0.

Thus, v ≡ 0 and hence u ≡ √ 0. But, this
√
cannot be a solution to (16.0.1).
λx − λx
If λ > 0, then v(x) = αe + βe . Equivalently,
√ √
v(x) = c1 cosh( λx) + c2 sinh( λx)
such that α = (c1 + c2 )/2 and β = (c1 − c2 )/2. Using the boundary condition
v(0) = 0, we get c1 = 0 and hence
√
v(x) = c2 sinh( λx).
√
Now using v(L) = 0, we have c2 sinh λL = 0. Thus, c2 = 0 and v(x) = 0.
We have seen this cannot be a solution. √
Finally, if λ < 0, then set ω = −λ. We need to solve the simple
harmonic oscillator problem
00
v (x) + ω 2 v(x) = 0 x ∈ (0, L)
v(0) = v(L) = 0.
The general solution is
v(x) = α cos(ωx) + β sin(ωx).
Using v(0) = 0, we get α = 0 and hence v(x) = β sin(ωx). Now using

v(L) = 0, we have β sin ωL = 0. Thus, either β = 0 or sin ωL = 0. But
β = 0 does not yield a solution. Hence ωL = kπ or ω = kπ/L, for all non-
zero k ∈ Z. Since ω > 0, we can consider only k ∈ N. Hence, for each k ∈ N,
there is a solution (vk , λk ) for the eigen value problem with

kπx
vk (x) = βk sin ,
L
for some constant bk and λk = −(kπ/L)2 . It now remains to solve w for each
of these λk . For each k ∈ N, we solve for wk in the ODE
wk00 (t) + (ckπ/L)2 wk (t) = 0.

ckπt ckπt
wk (t) = ak cos + bk sin .
L L
For each k ∈ N, we have

ckπt ckπt kπx
uk (x, t) = ak cos + bk sin sin
L L L
for some constants ak and bk . The situation corresponding to k = 1 is called

the fundamental mode and the frequency of the fundamental mode is
√ p
c −λ1 1 cπ c T /ρ
= = = .
2π 2π L 2L 2L
The frequency of higher modes are integer multiples of the fundamental fre-
quency. Note that the frequency of the vibration is related to eigenvalues of
the second order differential operator.
The general solution of (16.0.1), by principle of superposition, is
∞
X ckπt ckπt kπx
u(x, t) = ak cos + bk sin sin .
k=1
L L L
Note that the solution is expressed as series, which raises the question of
convergence of the series. Another concern is whether all solutions of (16.0.1)
have this form. We ignore these two concerns at this moment.
Since we know the initial position of the string as the graph of g, we get
∞
X kπx
g(x) = u(x, 0) = ak sin .
k=1
L
This expression is again troubling and rises the question: Can any arbitrary
function g be expressed as an infinite sum of trigonometric functions? An-
swering this question led to the study of “Fourier series”. Let us also, as
usual, ignore this concern for time being. Then, can we find the the con-
lπx
stants ak with knowledge of g. By multiplying sin L both sides of the
expression of g and integrating from 0 to L, we get
Z L Z L "X ∞ #
lπx kπx lπx
g(x) sin dx = ak sin sin dx
0 L 0 k=1
L L
∞ Z L
X kπx lπx
= ak sin sin dx
k=1 0 L L
Therefore, the constants ak are given as
2 L
Z
kπx
ak = g(x) sin .
L 0 L
Finally, by differentiating u w.r.t t, we get

∞
X ckπ ckπt ckπt kπx
ut (x, t) = bk cos − ak sin sin .
k=1
L L L L
Employing similar arguments and using ut (x, 0) = h(x), we get

∞
X bk kcπ kπx
h(x) = ut (x, 0) = sin
k=1
L L
and hence Z L
2 kπx
bk = h(x) sin .
kcπ 0 L
16.1 Elliptic Equations

Theorem 16.1.1 (Laplacian in 2D Rectangle). Let Ω = {(x, y) ∈ R2 | 0 <
x < a and 0 < y < b} be a rectangle in R2 . Let g : ∂Ω → R which vanishes
on three sides of the rectangle, i.e., g(0, y) = g(x, 0) = g(a, y) = 0 and
g(x, b) = h(x) where h is a continuous function h(0) = h(a) = 0. Then there
is a unique solution to (9.0.1) on this rectangle with given boundary value g.
Proof. We begin by looking for solution u(x, y) whose variables are separated,
i.e., u(x, y) = v(x)w(y). Substituting this form of u in the Laplace equation,
we get
v 00 (x)w(y) + v(x)w00 (y) = 0.
Hence
v 00 (x) w00 (y)
=− .
v(x) w(y)
Since LHS is function of x and RHS is function y, they must equal a constant,
say λ. Thus,
v 00 (x) w00 (y)
=− = λ.
v(x) w(y)
Using the boundary condition on u, u(0, y) = g(0, y) = g(a, y) = u(a, y) =

0, we get v(0)w(y) = v(a)w(y) = 0. If w ≡ 0, then u ≡ 0 which is not a
solution to (9.0.1). Hence, w 6≡ 0 and v(0) = v(a) = 0. Thus, we need to
solve, 00
v (x) = λv(x), x ∈ (0, a)
v(0) = v(a) = 0,
the eigen value problem for the second order differential operator. Note that
the λ can be either zero, positive or negative.
If λ = 0, then v 00 = 0 and the general solution is v(x) = αx + β, for some
constants α and β. Since v(0) = 0, we get β = 0, and v(a) = 0 and a 6= 0
implies that α = 0. Thus, v ≡ 0 and hence u ≡ 0. But, this can not be a
solution to (9.0.1). √ √
If λ > 0, then v(x) = αe λx + βe− λx . Equivalently,
√ √
v(x) = c1 cosh( λx) + c2 sinh( λx)
such that α = (c1 + c2 )/2 and β = (c1 − c2 )/2. Using the boundary condition
v(0) = 0, we get c1 = 0 and hence
√
v(x) = c2 sinh( λx).
√
Now using v(a) = 0, we have c2 sinh λa = 0. Thus, c2 = 0 and v(x) = 0.
We have seen this cannot √ be a solution.
If λ < 0, then set ω = −λ. We need to solve
00
v (x) + ω 2 v(x) = 0 x ∈ (0, a)
(16.1.1)
v(0) = v(a) = 0.
v(x) = α cos(ωx) + β sin(ωx).
Using the boundary condition v(0) = 0, we get α = 0 and hence v(x) =

β sin(ωx). Now using v(a) = 0, we have β sin ωa = 0. Thus, either β = 0
or sin ωa = 0. But β = 0 does not yield a solution. Hence ωa = kπ or
ω = kπ/a, for all non-zero k ∈ Z. Hence, for each k ∈ N, there is a solution
(vk , λk ) for (16.1.1), with

kπx
vk (x) = βk sin ,
a
for some constant βk and λk = −(kπ/a)2 . We now solve w corresponding to

each λk . For each k ∈ N, we solve for wk in the ODE
2
wk00 (y) = kπ
a
wk (y), y ∈ (0, b)
w(0) = 0.
Thus, wk (y) = ck sinh(kπy/a). Therefore, for each k ∈ N,

kπx kπy
uk = δk sin sinh
a a
is a solution to (9.0.1). The general solution is of the form (principle of

superposition) (convergence?)
∞
X kπx kπy
u(x, y) = δk sin sinh .
k=1
a a
The constant δk are obtained by using the boundary condition u(x, b) = h(x)
which yields
∞
X kπb kπx
h(x) = u(x, b) = δk sinh sin .
k=1
a a
Since h(0) = h(a) = 0, the function h admits a Fourier Sine series. Thus
δk sinh kπb
a
is the k-th Fourier sine coefficient of h, i.e.,
−1 Z a
kπb 2 kπx
δk = sinh h(x) sin .
a a 0 a
Theorem 16.1.2 (2D Disk). Let Ω = {(x, y) ∈ R2 | x2 + y 2 < R2 } be the

disk of radius R in R2 . Let g : ∂Ω → R is a continuous function. Then there
is a unique solution to (9.0.1) on the unit disk with given boundary value g.
Proof. Given the nature of the domain, we shall use the Laplace operator in
polar coordinates,
1 ∂2

1 ∂ ∂
∆ := r + 2 2
r ∂r ∂r r ∂θ
where r is the magnitude component and θ is the direction component. Then

∂Ω is the circle of radius one. Then, solving for u(x, y) in the Dirichlet
problem is to equivalent to finding U (r, θ) : Ω → R such that
 1∂ 1 ∂2U
 r ∂r r ∂U
∂r
+ r2 ∂θ2 = 0 in Ω
U (r, θ + 2π) = U (r, θ) in Ω (16.1.2)
U (R, θ) = G(θ) on ∂Ω

where U (r, θ) = u(r cos θ, r sin θ), G : [0, 2π) → R is G(θ) = g(cos θ, sin θ).
Note that both U and G are 2π periodic w.r.t θ. We will look for solution
U (r, θ) whose variables can be separated, i.e., U (r, θ) = v(r)w(θ) with both
v and w non-zero. Substituting it in the polar form of Laplacian, we get
v d2 w

w d dv
r + 2 2 =0
r dr dr r dθ
and hence
1 d2 w

−r d dv
r = .
v dr dr w dθ2
Since LHS is a function of r and RHS is a function of θ, they must equal a
constant, say λ. We need to solve the eigen value problem,
00
w (θ) − λw(θ) = 0 θ∈R
w(θ + 2π) = w(θ) ∀θ.
Note that the λ can be either zero, positive or negative. If λ = 0, then

w00 = 0 and the general solution is w(θ) = αθ + β, for some constants α and
β. Using the periodicity of w,
αθ + β = w(θ) = w(θ + 2π) = αθ + 2απ + β
implies that α = 0. Thus, the pair λ = 0 and w(θ) = β is a solution. If

λ > 0, then √ √
w(θ) = αe λθ + βe− λθ .
If either of α and β is non-zero, then w(θ) → ±∞ as θ → ∞, which contra-
dicts the periodicity of w. Thus,√α = β = 0 and w ≡ 0, which cannot be a
solution. If λ < 0, then set ω = −λ and the equation becomes
00
w (θ) + ω 2 w(θ) = 0 θ∈R
w(θ + 2π) = w(θ) ∀θ
Its general solution is
w(θ) = α cos(ωθ) + β sin(ωθ).
Using the periodicity of w, we get ω = k where k is an integer. For each

k ∈ N, we have the solution (wk , λk ) where
λk = −k 2 and wk (θ) = αk cos(kθ) + βk sin(kθ).
For the λk ’s, we solve for vk , for each k = 0, 1, 2, . . .,

d dvk
r r = k 2 vk .
dr dr
For k = 0, we get v0 (r) = α ln r + β. But ln r blows up as r → 0, but any

solution U and, hence v, on the closed unit disk (compact subset) has to be
bounded. Thus, we must have the α = 0. Hence v0 ≡ β. For k ∈ N, we need
to solve for vk in
d dvk
r r = k 2 vk .
dr dr
ds d d ds
Use the change of variable r = es . Then es dr = 1 and dr = ds dr
= e1s ds
d
.
d d s ks −ks k −k −k
Hence r dr = ds . vk (e ) = αe + βe . vk (r) = αr + βr . Since r blows
up as r → 0, we must have β = 0. Thus, vk = αrk . Therefore, for each
k = 0, 1, 2, . . .,
Uk (r, θ) = ak rk cos(kθ) + bk rk sin(kθ).
∞
a0 X
ak rk cos(kθ) + bk rk sin(kθ) .

U (r, θ) = +
2 k=1
To find the constants, we must use U (R, θ) = G(θ). If G ∈ C 1 [0, 2π], then G
admits Fourier series expansion. Therefore,
∞
a0 X k
R ak cos(kθ) + Rk bk sin(kθ)

G(θ) = +
2 k=1
where Z π
1
ak = k G(θ) cos(kθ) dθ,
R π −π
Z π
1
bk = k G(θ) sin(kθ) dθ.
R π −π
Using this in the formula for U and the uniform convergence of Fourier series,
we get
" ∞
#
1 π
Z
1 X r k
U (r, θ) = G(η) + (cos kη cos kθ + sin kη sin kθ) dη
π −π 2 k=1 R
" ∞
#
1 π
Z
1 X r k
= G(η) + cos k(η − θ) dη.
π −π 2 k=1 R
Using the relation

∞
" ∞
# " #
r i(η−θ)
X r k X r k e
cos k(η − θ) = Re ei(η−θ) = Re R
k=1
R k=1
R 1 − Rr ei(η−θ)
2
R − rR cos(η − θ)
= −1
R2
+ r2 − 2rR cos(η − θ)
rR cos(η − θ) − r2
=
R2 + r2 − 2rR cos(η − θ)
in U (r, θ) we get
π
R2 − r 2
Z
G(η)
U (r, θ) = dη.
2π −π R2 + r2 − 2rR cos(η − θ)
Note that the formula derived above for U (r, θ) can be rewritten in Carte-
sian coordinates and will have the form
R2 − |x|2
Z
g(y)
u(x) = 2
dy.
2πR SR (0) |x − y|
This can be easily seen, by setting y = R(x10 cos η +x20 sin η), we get dy = Rdη
and |x − y|2 = R2 + r2 − 2rR cos(η − θ). This is called the Poisson formula.
More generally, the unique solution to the Dirichlet problem on a ball of
radius R centred at x0 in Rn is given by Poisson formula
R2 − |x − x0 |2
Z
g(y)
u(x) = n
dy.
ωn R SR (x0 ) |x − y|
We will derive this general form later (cf. (??)).

Theorem 16.1.3 (3D Sphere). Let Ω = {(x, y, z) ∈ R3 | x2 + y 2 + z 2 < 1}

be the unit sphere in R3 . Let g : ∂Ω → R is a continuous function. Then
there is a unique solution to (9.0.1) on the unit sphere with given boundary
value g.
Proof. Given the nature of domain, the Laplace operator in spherical coor-
dinates,
∂2

1 ∂ 2 ∂ 1 ∂ ∂ 1
∆ := 2 r + 2 sin φ + 2 2 .
r ∂r ∂r r sin φ ∂φ ∂φ r sin φ ∂θ2
where r is the magnitude component, φ is the inclination (zenith or elevation)
in the vertical plane and θ is the azimuth angle (in the direction in horizontal
plane). Solving for u in (9.0.1) is equivalent to finding U (r, φ, θ) : Ω → R
such that

1 ∂ 2 ∂U 1 ∂ ∂U

 r2 ∂r r ∂r + r2 sin φ ∂φ sin φ ∂φ

1 ∂2U (16.1.3)
 + r2 sin 2 φ ∂θ 2 =0 in Ω

U (1, φ, θ) = G(φ, θ) on ∂Ω
where U (r, φ, θ) and G(φ, θ) are appropriate spherical coordinate function
corresponding to u and g. We will look for solution U (r, φ, θ) whose variables
can be separated, i.e., U (r, φ, θ) = v(r)w(φ)z(θ) with v, w and z non-zero.
Substituting it in the spherical form of Laplacian, we get
vw d2 z

wz d 2 dv vz d dw
r + sin φ + =0
r2 dr dr r2 sin φ dφ dφ r2 sin2 φ dθ2
and hence
1 d2 z

1 d 2 dv −1 d dw
r = sin φ − .
v dr dr w sin φ dφ dφ z sin2 φ dθ2
Since LHS is a function of r and RHS is a function of (φ, θ), they must equal
a constant, say λ. If Azimuthal symmetry is present then z(θ) is constant
and hence dz
dθ
= 0. We need to solve for w,
sin φw00 (φ) + cos φw0 (φ) + λ sin φw(φ) = 0, φ ∈ (0, π)
dx
Set x = cos φ. Then dφ
= − sin φ.
dw d2 w dw
w0 (φ) = − sin φ and w00 (φ) = sin2 φ 2 − cos φ
dx dx dx
In the new variable x, we get the Legendre equation

(1 − x2 )w00 (x) − 2xw0 (x) + λw(x) = 0 x ∈ [−1, 1].
We have already seen that this is a singular problem (while studying S-L
problems). For each k ∈ N ∪ {0}, we have the solution (wk , λk ) where
λk = k(k + 1) and wk (φ) = Pk (cos φ).
For the λk ’s, we solve for vk , for each k = 0, 1, 2, . . .,

d 2 dvk
r = k(k + 1)vk .
dr dr
For k = 0, we get v0 (r) = −α/r + β. But 1/r blows up as r → 0 and U must
be bounded in the closed sphere. Thus, we must have the α = 0. Hence
v0 ≡ β. For k ∈ N, we need to solve for vk in

d 2 dvk
r = k(k + 1)vk .
dr dr
ds d d ds
Use the change of variable r = es . Then es dr = 1 and dr = ds dr
= e1s ds
d
.
d d
Hence r dr = ds . Solving for m in the quadratic equation m2 + m = k(k + 1).
m1 = k and m2 = −k − 1. vk (es ) = αeks + βe(−k−1)s . vk (r) = αrk + βr−k−1 .
Since r−k−1 blows up as r → 0, we must have β = 0. Thus, vk = αrk .
Therefore, for each k = 0, 1, 2, . . .,
Uk (r, φ, θ) = ak rk Pk (cos φ).
∞
X
U (r, φ, θ) = ak rk Pk (cos φ).
k=0
Since we have azimuthal symmetry, G(φ, θ) = G(φ). To find the constants,

we use U (1, φ, θ) = G(φ), hence
∞
X
G(φ) = ak Pk (cos φ).
k=0
Using the orthogonality of Pk , we have

2k + 1 π
Z
ak = G(φ)Pk (cos φ) sin φ dφ.
2 0
Lecture 17
Parabolic: Heat Equation
Theorem 17.0.1 (Heat Flow on a Bar). Let Ω = (0, L) be a homogeneous

rod of length L insulated along sides and its ends are kept at zero temperature.
The temperature zero at the end points of the rod is given by the Dirichlet
boundary condition u(0, t) = u(L, t) = 0. The initial temperature of the rod,
at time t = 0, is given by u(x, 0) = g(x), where g : [0, L] → R be such that
g(0) = g(L) = 0. Then there is a solution u of

 ut (x, t) − c2 uxx (x, t) = 0 in (0, L) × (0, ∞)
u(0, t) = u(L, t) = 0 in (0, ∞)
u(x, 0) = g(x) on [0, L]

where c is a constant.
Proof. We begin with the ansatz that u(x, t) = v(x)w(t) (variable separated).
Substituting u in separated form in the equation, we get
v(x)w0 (t) = c2 v 00 (x)w(t)
and, hence,
w0 (t) v 00 (x)
= .
c2 w(t) v(x)
Since LHS, a function of t, and RHS, a function x, are equal they must be
equal to some constant, say λ. Thus,
w0 (t) v 00 (x)
= = λ.
c2 w(t) v(x)
99
LECTURE 17. PARABOLIC: HEAT EQUATION 100
Therefore, we need to solve two ODE to obtain v and w,

w0 (t) = λc2 w(t) and v 00 (x) = λv(x).
We first solve the eigenvalue problem involving v. For each k ∈ N, there
is a pair (λk , vk ) which solves the eigenvalue problem involving v, where
kπx
2 2

λk = −(kπ) /L and vk (x) = sin L . For each k ∈ N, we solve for wk to
get
ln wk (t) = λk c2 t + ln α,
2
where α is integration constant. Thus, wk (t) = αe−(kcπ/L) t . Hence,

kπx −(kcπ/L)2 t
uk (x, t) = vk (x)wk (t) = βk sin e ,
L
for some constants βk , is a solution to the heat equation. By superposition
principle, the general solution is
∞ ∞
X X kπx −(kcπ/L)2 t
u(x, t) = uk (x, t) = βk sin e .
k=1 k=1
L
We now use the initial temperature of the rod, given as g : [0, L] → R to

compute the constants. Since u(x, 0) = g(x),
∞
X kπx
g(x) = u(x, 0) = βk sin .
k=1
L
Further, g(0) = g(L) = 0. Thus, g admits a Fourier Sine expansion and

hence its coefficients βk are given as
2 L
Z
kπx
βk = g(x) sin .
L 0 L
Theorem 17.0.2 (Circular Wire). Let Ω be a circle (circular wire) of radius

one insulated along its sides. Let the initial temperature of the wire, at time
t = 0, be given by a 2π-periodic function g : R → R. Then there is a solution
u(r, θ) of

 ut (θ, t) − c2 uθθ (θ, t) = 0 in R × (0, ∞)
u(θ + 2π, t) = u(θ, t) in R × (0, ∞)
u(θ, 0) = g(θ) on R × {t = 0}

where c is a constant.
Proof. Note that u(θ, t) is 2π-periodic in θ-variable, i.e., u(θ + 2π, t) = u(θ, t)
for all θ ∈ R and t ≥ 0. We begin with ansatz u(θ, t) = v(θ)w(t) with
variables separated. Substituting for u in the equation, we get
w0 (t) v 00 (θ)
= = λ.
c2 w(t) v(θ)
For each k ∈ N ∪ {0}, the pair (λk , vk ) is a solution to the eigenvalue problem
where λk = −k 2 and
vk (θ) = ak cos(kθ) + bk sin(kθ).

2
For each k ∈ N ∪ {0}, we get wk (t) = αe−(kc) t . For k = 0
u0 (θ, t) = a0 /2 (To maintain consistency with Fourier series)
and for each k ∈ N, we have

2 c2 t
uk (θ, t) = [ak cos(kθ) + bk sin(kθ)] e−k .
Therefore, the general solution is

∞
a0 X 2 2
u(θ, t) = + [ak cos(kθ) + bk sin(kθ)] e−k c t .
2 k=1
We now use the initial temperature on the circle to find the constants. Since
u(θ, 0) = g(θ),
∞
a0 X
g(θ) = u(θ, 0) = + [ak cos(kθ) + bk sin(kθ)] .
2 k=1
Further, g is 2π-periodic and, hence, admits a Fourier series expansion. Thus,

1 π
Z
ak = g(θ) cos(kθ) dθ
π −π
and Z π
1
bk = g(θ) sin(kθ) dθ.
π −π
Note that as t → ∞ the temperature of the wire approaches a constant

a0 /2.
17.1 Inhomogeneous Equation

In this section we solve the inhomogeneous heat equation, using Duhamel’s
principle. The Duhamel’s principle states that one can obtain a solution of
the inhomogeneous IVP for heat from its homogeneous IVP. The motivation
for Duhamel’s principle is given in Appendix C.
For a given f , let u(x, t) be the solution of the inhomogeneous heat equa-
tion,

 ut (x, t) − c2 ∆u(x, t) = f (x, t) in Ω × (0, T )
u(x, t) = 0 in ∂Ω × (0, T ) (17.1.1)
u(x, 0) = 0 in Ω.

As a first step, for each s ∈ (0, ∞), consider w(x, t; s) as the solution of the
homogeneous problem (auxiliary)
 s
 wt (x, t) − c2 ∆ws (x, t) = 0 in Ω × (s, T )
ws (x, t) = 0 in ∂Ω × (s, T )
ws (x, s) = f (x, s) on Ω × {s}.

Since t ∈ (s, T ), introducing a change of variable r = t−s, we have ws (x, t) =

w(x, t − s) which solves

 wt (x, r) − c2 ∆w(x, r) = 0 in Ω × (0, T − s)
w(x, r) = 0 in ∂Ω × (0, T − s)
w(x, 0) = f (x, s) on Ω.

Theorem 17.1.1 (Duhamel’s Principle). The function u(x, t) defined as
Z t Z t
s
u(x, t) := w (x, t) ds = w(x, t − s) ds
0 0
solves (17.1.1).
Proof. Suppose w is C 2,1 (Rn × (0, T )), we get
∂ t
Z
ut (x, t) = w(x, t − s) ds
∂t 0
Z t
d(t)
= wt (x, t − s) ds + w(x, t − t)
0 dt
d(0)
− w(x, t − 0)
dt
Z t
= wt (x, t − s) ds + w(x, 0)
0
Z t
= wt (x, t − s) ds + f (x, t).
0
Similarly, Z t
∆u(x, t) = ∆w(x, t − s) ds.
0
Thus,
Z t
2
wt (x, t − s) − c2 ∆w(x, t − s) ds

ut − c ∆u = f (x, t) +
0
= f (x, t).
Lecture 18
Travelling Waves
Consider the wave equation utt = c2 uxx on R × (0, ∞), describing the vibra-
tion of an infinite string. We have already seen in Chapter 5 that the equation
is hyperbolic and has the two characteristics x ± ct= a constant. Introduce
the new coordinates w = x + ct, z = x − ct and set u(w, z) = u(x, t). Thus,
we have the following relations, using chain rule:
ux = uw wx + uz zx = uw + uz
ut = uw wt + uz zt = c(uw − uz )
uxx = uww + 2uzw + uzz
utt = c2 (uww − 2uzw + uzz )
In the new coordinates, the wave equation satisfies uwz = 0. Integrating1

this twice, we have u(w, z) = F (w) + G(z), for some arbitrary functions F
and G. Thus, u(x, t) = F (x + ct) + G(x − ct) is a general solution of the
wave equation.
Consider the case where G is chosen to be zero function. Then u(x, t) =
F (x+ct) solves the wave equation. At t = 0, the solution is simply the graph
of F and at t = t0 the solution is the graph of F with origin translated to the
left by ct0 . Similarly, choosing F = 0 and G = F , we have u(x, t) = F (x−ct)
also solves wave equation and at time t is the translation to the right of the
graph of F by ct. This motivates the name “travelling waves” and “wave
equation”. The graph of F is shifted to right or left with a speed of c.
1
We are assuming the function is integrable, which may be false
105
LECTURE 18. TRAVELLING WAVES 106
Now that we have derived the general form of the solution of wave equa-
tion, we return to understand the physical system of a vibrating infinite
string. The initial shape (position at initial time t = 0) of the string is given
as u(x, 0) = g(x), where the graph of g on R2 describes the shape of the
string. Since we need one more data to identify the arbitrary functions, we
also prescribe the initial velocity of the string, ut (x, 0) = h(x).
Another interesting property that follows from the general solution is
that for any four points A, B, C and D that form a rectangle bounded by
characteristic curves in R × R+ then u(A) + u(C) = u(B) + u(D) because
u(A) = F (α) + G(β), u(C) = F (γ) + G(δ), u(B) = F (α) + G(δ) and u(D) =
F (γ) + G(β).
Theorem 18.0.1. Given g ∈ C 2 (R) and h ∈ C 1 (R), there is a unique C 2
solution u of the Cauchy initial value problem (IVP) of the wave equation,

 utt (x, t) − c2 uxx (x, t) = 0 in R × (0, ∞)
u(x, 0) = g(x) in R (18.0.1)
ut (x, 0) = h(x) in R,

which is given by the d’Alembert’s formula

Z x+ct
1 1
u(x, t) = (g(x + ct) + g(x − ct)) + h(y) dy. (18.0.2)
2 2c x−ct
Proof. The general solution is u(x, t) = F (x + ct) + G(x − ct) with F, G ∈

C 2 (R). Using the initial position we get
F (x) + G(x) = g(x).
Thus, g should be C 2 (R). Now, ut (x, t) = c (F 0 (w) − G0 (z)) and putting

t = 0, we get
1
F 0 (x) − G0 (x) = h(x).
c
Thus, h should be C (R). Now solving for F 0 and G0 , we get 2F 0 (x) =
1
g 0 (x) + h(x)/c. Similarly, 2G0 (x) = g 0 (x) − h(x)/c. Integrating2 both these
equations, we get
1 x
Z
1
F (x) = g(x) + h(y) dy + c1
2 c 0
2
assuming they are integrable and the integral of their derivatives is itself
and Z x
1 1
G(x) = g(x) − h(y) dy + c2 .
2 c 0
Since F (x) + G(x) = g(x), we get c1 + c2 = 0. Therefore, the solution to the
wave equation is given by (18.0.2).
Aliter. Let us derive the d’Alembert’s formula in an alternate way. Note
that the wave equation can be factored as

∂ ∂ ∂ ∂
+c −c u = utt − c2 uxx = 0.
∂t ∂x ∂t ∂x
∂ ∂

We set v(x, t) = ∂t − c ∂x u(x, t) and hence
vt (x, t) + cvx (x, t) = 0 in R × (0, ∞).
Notice that the above first order PDE obtained is in the form of homogeneous
linear transport equation (cf. (??)), which we have already solved. Hence,
for some smooth function φ,
v(x, t) = φ(x − ct)
and φ(x) := v(x, 0). Using v in the original equation, we get the inhomoge-
neous transport equation,
ut (x, t) − cux (x, t) = φ(x − ct).
Recall the formula for inhomogenoeus transport equation (cf. (??))
Z t
u(x, t) = g(x − at) + φ(x − a(t − s), s) ds.
0
Since u(x, 0) = g(x) and a = −c, in our case the solution reduces to,
Z t
u(x, t) = g(x + ct) + φ(x + c(t − s) − cs) ds
0
Z t
= g(x + ct) + φ(x + ct − 2cs) ds
0
−1 x−ct
Z
= g(x + ct) + φ(y) dy
2c x+ct
1 x+ct
Z
= g(x + ct) + φ(y) dy.
2c x−ct
But φ(x) = v(x, 0) = ut (x, 0) − cux (x, 0) = h(x) − cg 0 (x) and substituting
this in the formula for u, we get
1 x+ct
Z
u(x, t) = g(x + ct) + (h(y) − cg 0 (y)) dy
2c x−ct
1
= g(x + ct) + (g(x − ct) − g(x + ct))
2
1 x+ct
Z
+ h(y) dy
2c x−ct
1 x+ct
Z
1
= (g(x − ct) + g(x + ct)) + h(y) dy
2 2c x−ct
For c = 1, the d’Alembert’s formula takes the form

1 x+t
Z
1
u(x, t) = (g(x − t) + g(x + t)) + h(y) dy.
2 2 x−t
A useful observation from the d’Alembert’s formula is that the regularity of
u is same as the regularity of its initial value g.
18.1 Domain of Dependence and Influence

Note that the solution u(x, t) depends only on the interval [x − ct, x + ct]
because g takes values only on the end-points of this interval and h takes
values between this interval. The interval [x − ct, x + ct] is called the domain
of dependence. Thus, the region of R × (0, ∞) on which the value of u(x, t)
depends forms a triangle with base [x − ct, x + ct] and vertex at (x, t). The
domain of dependence of (x, t) is marked in x-axis by the characteristic curves
passing through (x, t).
Given a point p on the x-axis what values of u on (x, t) will depend on
the value of g(p) and h(p). This region turns out to be a cone with vertex
at p and is called the domain of influence. The domain of influence is the
region bounded by the characteristic curves passing through p.
If the initial data g and h are supported in the interval Bx0 (R) then the
solution u at (x, t) is supported in the region Bx0 (R + ct). Consequently, if g
and h have compact support then the solution u has compact support in R
for all time t > 0. This phenomenon is called the finite speed of propagation.
Appendices
109
Appendix A
Divergence Theorem
Definition A.0.1. For an open set Ω ⊂ Rn , its boundary ∂Ω is said to be C k

(k ≥ 1) if, for every point x ∈ ∂Ω, there is a r > 0 and a C k diffeomorphism1
φ : Br (x) → B1 (0) such that
1. φ(∂Ω ∩ Br (x)) ⊂ B1 (0) ∩ {x ∈ Rn | xn = 0} and
2. φ(Ω ∩ Br (x)) ⊂ B1 (0) ∩ {x ∈ Rn | xn > 0}
The boundary ∂Ω is said to be C ∞ if ∂Ω is C k , for all k ∈ N, and ∂Ω is
analytic if φ is analytic.
Equivalently, ∂Ω is C k if, for every point x ∈ ∂Ω, there exists a neigh-
bourhood Ux of x and a C k function φ : Rn−1 → R such that
Ω ∩ Bx = {x ∈ Bx | xn > φ(x1 , x2 , . . . , xn−1 )}.
Theorem A.0.2. Let Ω be an open bounded subset of Rn with C 1 boundary.

If v ∈ C 1 (Ω) then Z Z
∂v
dx = vνi dσ
Ω ∂xi ∂Ω
where ν = (ν1 , . . . , νn ) is the unit vector pointing outward and dσ is the
surface measure of ∂Ω.

∂v
The hypothesis that Ω is bounded can be relaxed provided |v| and ∂x i

decays as |x| → ∞. Much weaker hypotheses on ∂Ω and v are considered in
geometric measure theory.
1 −1
φ exists and both φ and φ−1 are k-times continuously differentiable
111
APPENDIX A. DIVERGENCE THEOREM 112
Theorem A.0.3 (Integration by parts). Let Ω be an open bounded subset of

Rn with C 1 boundary. If u, v ∈ C 1 (Ω) then
Z Z Z
∂v ∂u
u dx + v dx = uvνi dσ.
Ω ∂xi Ω ∂xi ∂Ω
Hint. Set v := uv in the theorem above.
Theorem A.0.4 (Gauss). Let Ω be an open bounded subset of Rn with C 1

boundary. If V = (v1 , . . . , vn ) on Ω is a vector field such that vi ∈ C 1 (Ω), for
all 1 ≤ i ≤ n, then Z Z
∇ · V dx = V · ν dσ. (A.0.1)
Ω ∂Ω
The divergence of a vector field is the measure of the magnitude (outgoing

nature) of all source (of the vector field) and absorption in the region. The
divergence theorem was discovered by C. F. Gauss in 18132 which relates the
outward flow (flux) of a vector field through a closed surface to the behaviour
of the vector field inside the surface (sum of all its “source” and “sink”). The
divergence theorem is the mathematical formulation of the conservation law.
Theorem A.0.5 (Green’s Identities). Let Ω be an open bounded subset of

Rn with C 1 boundary. If u, v ∈ C 2 (Ω) then
(i) Z Z
∂u
(v∆u + ∇v · ∇u) dx = v dσ,
Ω ∂Ω ∂ν
∂u
where ∂ν
:= ∇u · ν;
(ii) Z Z
∂u ∂v
(v∆u − u∆v) dx = v −u dσ.
Ω ∂Ω ∂ν ∂ν
Hint. Apply divergence theorem to V = v∇u to get the first formula. To get
second formula apply divergence theorem for both V = v∇u and V = u∇v
and subtract one from the other.
2
J. L. Lagrange might have discovered this, before Gauss, in 1762
Appendix B
Normal Vector of a Surface
Let S(x, y, z) = 0 be the equation of a surface G in R3 . Fix p0 = (x0 , y0 , z0 ) ∈

G. What is the normal vector at p0 ? Fix an arbitrary curve C lying on G
and passing through p0 . Let r(t) = (x(t), y(t), z(t)) be the parametric form
of C with r(t0 ) = p0 . Since C lies on G, S(r(t)) = S(x(t), y(t), z(t)) = 0, for
all t. Differentiating w.r.t t (using chain rule),
∂S dx(t) ∂S dy(t) ∂S dz(t)

+ + = 0
∂x dt ∂y dt ∂z dt
(Sx , Sy , Sz ) · (x0 (t), y 0 (t), z 0 (t)) = 0
∇S(r(t)) · r0 (t) = 0.
In particular, ∇S(p0 ) · r0 (t0 ) = 0. Since r0 (t0 ) is the slope of the tangent,

at t0 , to the curve C, the vector ∇S(p0 ) is perpendicular to the tangent vector
at p0 . Since the argument is valid for any curve in G that passes through
p0 , ∇S(p0 ) is normal vector to the tangent plane at p0 . If, in particular,
the equation of the surface is given as S(x, y, z) = u(x, y) − z, for some
u : R2 → R, then
∇S(p0 ) = (Sx (p0 ), Sy (p0 ), Sz (p0 ))

= (ux (x0 , y0 ), uy (x0 , y0 ), −1) = (∇u(x0 , y0 ), −1).
113
APPENDIX B. NORMAL VECTOR OF A SURFACE 114
Appendix C
Duhamel’s Principle
Consider the first order inhomogeneous ODE

0
x (t) + ax(t) = f (t) in (0, ∞)
(C.0.1)
x(0) = x0 .
Multiplying the integration factor eat both sides, we get
[eat x(t)]0 = eat f (t)
and Z t
−at
x(t) = e eas f (s) ds + ce−at .
0
Using the initial condition x(0) = x0 , we get
Z t
−at
x(t) = x0 e + ea(s−t) f (s) ds.
0
Notice that x0 e−at is a solution of the homogeneous ODE. Thus, the solution
x(t) can be given as
Z t
x(t) = S(t)x0 + S(t − s)f (s) ds
0
where S(t) is a solution operator of the linear equation, given as S(t) = e−at .
Consider the second order inhomogeneous ODE
 00
 x (t) + a2 x(t) = f (t) in (0, ∞)
x(0) = x0 (C.0.2)
0
x (0) = x1 .

115
APPENDIX C. DUHAMEL’S PRINCIPLE 116
We introduce a new function y such that
x0 (t) = ay(t).
Then
f (t)
y 0 (t) = − ax(t)
a
and the second order ODE can be rewritten as a system of first order ODE
X 0 (t) + AX(t) = F (t)
where X = (x, y), F = (0, f /a) and

0 −a
A=
a 0
with the initial condition X0 := X(0) = (x0 , x1 /a). We introduce the matrix
(At)n
exponential eAt = ∞ At
P
n=1 n! . Then, multiplying the integration factor e
both sides, we get
[eAt X(t)]0 = eAt F (t)
and Z t
−At
X(t) = X0 e + eA(s−t) F (s) ds.
0
Notice that X0 e−At is a solution of the homogeneous ODE. Thus, the solution
X(t) can be given as
Z t
X(t) = S(t)X0 + S(t − s)F (s) ds
0
where S(t) is a solution operator of the linear equation, given as S(t) = e−At .
Bibliography
117
BIBLIOGRAPHY 118
Index
characteristic curve, 14 method of characteristics, 14
directional derivative, 2 Neumann boundary condition, 50

divergence, 3 Neumann Problem, 50
elliptic PDE, 25 parabolic PDE, 25
equation
velocity potential, 26 tensor, 3
heat, 26
Laplace, 26
tricomi, 26
wave, 26
Gauss divergence result, 112

gradient, 3
Green’s identities, 112
Hadamard, 9
wellposed, 9
harmonic function, 45
Hessian matrix, 3
hyperbolic PDE, 25
integral curve, 14
integral surface, 14
Laplace operator, 41
Laplace-Beltrami operator, 43
Laplacian, 3
maximum principle
strong, 48
weak, 47
119

Mso203b PDF

Încărcat de

Informații document

Descriere originală:

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Mso203b PDF

Încărcat de

Drepturi de autor:

Formate disponibile

Partial Differential Equations

November 13, 2015

3 First Order PDE 11

4 Method of Characteristics: Continued... 17

5 Classification of Second Order PDE 23

6 Classification of SOPDE: Continued 29

7 Classification of SOPDE: Continued 33

9 Properties of Harmonic Functions 47

12 Singular Sturm-Liouville Problem 59

13 Orthogonality of Eigen Functions 63

15 Fourier Series: Continued... 77

16 Standing Waves: Separation of Variable 87

17 Parabolic: Heat Equation 99

18 Travelling Waves 105

A Divergence Theorem 111

B Normal Vector of a Surface 113

C Duhamel’s Principle 115

N denotes the set of natural numbers

Ω denotes an open subset of Rn , not necessarily bounded

∂Ω denotes the boundary of Ω

R denotes the set of real numbers

Rn denotes the n-dimensional Euclidean space

C(X) is the class of all continuous functions on X

C k (X) is the class of C k functions which admit a continuous extension to

C k (X) is the class of all k-times (k ≥ 1) continuously differentiable functions

C ∞ (X) is the class of all infinitely differentiable functions on X

C j,k (X × Y ) is the class of all j-times (j ≥ 0) continuously differentiable

BVP Boundary value problem

IVP Initial value problem

w.r.t with respect to

A partial differential equation (PDE) is an equation involving an unknown

(i) Modelling the problem or deriving the mathematical equation (formu-

(ii) Solving the equation (PDE). What constitutes as a solution to a PDE?

1.1 Multi-Index Notations

∂u u(x + hξ) − u(x)

∂u u(x + hei ) − u(x)

A multi-index α = (α1 , . . . , αn ) is a n-tuple where αi , for each 1 ≤ i ≤

The partial differential operator of order α is denoted as

For each k ∈ N, Dk u(x) := {Dα u(x) | |α| = k}. The k = 1 case is

D1 u(x) = D(1,0,...,0) u(x), D(0,1,0,...,0) u(x), . . . , D(0,0,...,0,1) u(x)

The operator D1 is called the gradient operator and is denoted as D or ∇.

Example 1.1. Let u(x, y) : R2 → R be defined as u(x, y) = ax2 + by 2 . Then

1.2 Classification of PDE

(iii) A k-th order PDE is quasilinear if F has the form

i.e., the coefficient of its highest (k-th) order derivative depends on u

(iv) A k-th order PDE is fully nonlinear if it depends nonlinearly on the

(ii) xux + yuy = x2 + y 2 is linear.

(iii) utt − c2 uxx = f (x, t) is linear.

(iv) y 2 uxx + xuyy = 0 is linear.

(vi) ut + uux + uxxx = 0 is semilinear.

(vii) u2tt + uxxxx = 0 is semilinear.

(viii) ux + uuy − u2 = 0 is quasilinear.

(ix) uux + uy = 2 is quasilinear.

2.1 Solution of PDE

• and u satisfies the equation (1.2.1).

Example 2.1. Consider the first order equation ux (x, y) = 0 in R2 . Freezing

Recall that the family of solutions of an ODE is indexed by constants. In

2.2 Well-posedness of PDE

(a) has a solution (existence);

(b) the solution is unique (uniqueness);

Example 2.6. The initial value problem (IVP)

sup{|uε (x, t) − u(x, t)|} = ε2 sup {|sin(x/ε) sin(t/ε)|} = ε2 .