Lecture 01 on 30.09.

2015
Partial Differential Equations
MA201 Mathematics III

M. G. P. Prasad
Department of Mathematics
IIT Guwahati

Lecture 01 on 30.09.2015 Partial Differential Equations 1 / 31

Learning Outcome of this Topic

We learn
1 Introduction to Partial Differential Equations
2 Classification of PDEs
3 Formulation of PDEs
4 Cauchys Problem
5 Well-Posted Problem

Reference: Sneddon Book, Chapter 2, Sections 1, 2, 3

Lecture 01 on 30.09.2015 Partial Differential Equations 2 / 31

Differential Equations

Lecture 01 on 30.09.2015 Partial Differential Equations 3 / 31

Differential Equations (DE)
Ordinary Differential Equation (ODE): It is an equation involving an
unknown function / dependent variable y (say) of one independent
variable x (say) and a finite number of derivatives of y(x).
Examples:
y 0 + y = 0, y 000 + 5(y 0 )2 + y = ex
Partial Differential Equation (PDE): It is an equation involving an unknown
function / dependent variable u (say) of two or more independent
variables x, y, z, t, etc and the partial derivatives of u.
Examples:
xux + yuy = 0, ut = uxx + 3u2 , uxx + uyy + uzz = 0
Notation for Partial Derivatives:
2u 2u
   
u u u
= ux , = = uxx , = = uxy ,
x x2 x x yx y x

2u 3u
    
u u
= = uyx , = = uxyz
xy x y zyx z y x
Lecture 01 on 30.09.2015 Partial Differential Equations 4 / 31
 Partial Differential Equations
Text Books:
Ian N. Sneddon, Elements of Partial Differential Equations, McGraw Hill.
K. Sankara Rao, Introduction to Partial Differential Equations, Third
Edition, Prentice Hall of India.

Lecture 01 on 30.09.2015 Partial Differential Equations 5 / 31

Partial Differential Equations (PDE)

A Partial Differential Equation (PDE) is a relation of the form

F (x, y, . . . , u, ux , uy , . . . , uxx , uxy , uyx , uyy , . . .) = 0

where F is a given function of independent variables x, y, . . ., and of the

unknown function u and of a finite number its partial derivatives.

A PDE is an equation involving an unknown function u (say) and its partial

derivatives.
The independent variables x, y, etc are real variables and u is a real valued
function of two or more independent variables, namely, x, y, etc.

The order of a PDE is defined as the order of the highest derivative occurring
in the equation.

Examples: The order ut uxx = 0 is 2. The order of u3t u7xy is 2.

The order of u + 5uxyx + 3uxxxx = 0 is 4.

Lecture 01 on 30.09.2015 Partial Differential Equations 6 / 31

Classification of PDEs

PDE: F (x1 , x2 , . . . , xn , u, ux1 , ux2 , . . . , ux1 x1 , ux1 x2 , . . .) = 0

where u is an unknown function of independent variables x1 , . . ., xn .
A PDE of order m is said to be quasi-linear if F is linear in the derivatives
of order m with coefficients that depend on x1 , . . ., xn and derivatives of u
of order < m.
Examples: u u2x + uy uxx = 0 and xyu ux + x2 uy + 5u3 = 0 are quasi-linear.
A PDE of order m is said to be semi-linear if F is quasi-linear and the
coefficients of derivatives of order m are functions of independent
variables x1 , . . ., xn alone.
Examples: u u2x + uxx = 0 and xy ux + x2 uy + 5u3 = 0 are semi-linear.
A PDE of order m is said to fully non-linear if it is not linear in the
derivatives of order m.
Examples: u2x + uy = 0 and uxx uyy + ux = 0 are non-linear.

Lecture 01 on 30.09.2015 Partial Differential Equations 7 / 31

Classification of PDEs (Continuation)

A PDE is said to be linear if it is linear in the unknown function u and its

derivatives, with coefficients depending on the independent variables x1 ,
. . ., xn alone.
Examples: ux + uxy = 0 and xux + xy uy + x3 uxx = 0 are linear.

A PDE is said to be homogeneous if each term of the PDE contains the

dependent variable u or its partial derivatives.
Examples: ux + uxy = 0 and xux + y u uy + x2 u = 0 are homogeneous.
A PDE is said to be non-homogeneous / inhomogeneous if it is NOT
homogeneous.
Examples: ux + uxy 2xy = 0 and xux + xy u uy = ex are
non-homogeneous.

Lecture 01 on 30.09.2015 Partial Differential Equations 8 / 31

Classification of PDEs in terms of linearity:

Classification of PDEs in terms of homogeneity:

Lecture 01 on 30.09.2015 Partial Differential Equations 9 / 31

Well Known Examples of PDEs:

2 2
Laplace Operator: 2 = = 2 + +
x1 x2n
Let u = u(x1 , . . . , xn ).
Laplace Equation:
u = 0 or 2 u = 0
Poisson Equation:
u = f (x1 , . . . , xn ) or 2 u = f (x1 , . . . , xn )

Let u = u(x1 , . . . , xn , t).

Wave Equation:
ut = ku, where k > 0 is a constant
Heat/ Diffusion Equation:
utt = c2 u, where c is a non-zero real constant

Lecture 01 on 30.09.2015 Partial Differential Equations 10 / 31

Well Known Examples of PDEs (Continuation):

Schrodingers Equation:
i ut + u = 0
Transport Equation:
n
X
ut = ci uxi = C u where C = (c1 , . . . , cn ) Rn
i=1

Let u = u(x, t).

Airys Equation:
ut + uxxx = 0
Beam Equation:
ut + uxxxx = 0
Inviscid Burgers Equation:

ut + u ux = 0

Lecture 01 on 30.09.2015 Partial Differential Equations 11 / 31

In MA201-PDE: What types of equations we are going
to study?

First Order PDEs in two independent variables

Linear
Quasi-Linear
Non-Linear

Linear Second Order PDEs

Wave Equation
Heat Equation
Laplace Equation

Lecture 01 on 30.09.2015 Partial Differential Equations 12 / 31

First Order PDE in Two Independent
Variables

Lecture 01 on 30.09.2015 Partial Differential Equations 13 / 31

First Order PDE in Two Independent Variables
Notation: Take x and y as independent variables and z = z(x, y) as
z z
dependent variable. Set p = = zx and q = = zy .
x y

Linear First Order PDE:

P (x, y) p + Q(x, y) q + R(x, y) z = S(x, z)
Semi-Linear First Order PDE:
P (x, y) p + Q(x, y) q = R(x, y, z)
It is linear in the highest order terms with coefficients are functions of
independent variables alone. However it need not be linear in z.
Quasi-Linear First Order PDE:
P (x, y, z) p + Q(x, y, z) q = R(x, y, z)
It is linear in the highest order terms with coefficients are functions of
independent variables and the derivatives of u of order < 1.
Lecture 01 on 30.09.2015 Partial Differential Equations 14 / 31
How First Order PDEs occur?
( FORMULATION OF PDEs )

Lecture 01 on 30.09.2015 Partial Differential Equations 15 / 31

 Formulation of PDE
( By Eliminating Arbitrary Constants )

Lecture 01 on 30.09.2015 Partial Differential Equations 16 / 31

Curves/Surfaces: Expression involved with parameters (arbitrary constants)

Let
f (x, y, z, a, b) = 0 (1)
3
represent two parameter family of surfaces in R , where a and b are arbitrary
constants (real parameters).
We would like to formulate a PDE which is satisfied by (1).
Step 1: Differentiating partially (1) with respect to x and y, we get

f f
+ p = 0 (2)
x z
f f
+ q = 0 (3)
y z
Step 2: Eliminating arbitrary constants a and b in (1), (2) and (3), we get a
relation of the form
F (x, y, z, p, q) = 0
which is a PDE for the unknown function z of two independent variables x and
y.
Lecture 01 on 30.09.2015 Partial Differential Equations 17 / 31
Example: Formulating PDE by eliminating arbitrary
constants

The equation
x2 + y 2 + (z c)2 = r2 (4)
where r > 0 and c are arbitrary constants, represents the set of all spheres
whose center lie on the z-axis. Form the PDE by eliminating arbitrary
constants from (4)
Differentiating (4) partially with respect to x and y, we get

x + (z c) p = 0 (5)

y + (z c) q = 0 (6)
Eliminating arbitrary constant c from (5) and (6), we obtain a first order PDE:

xq yp = 0.

Lecture 01 on 30.09.2015 Partial Differential Equations 18 / 31

 Formulation of PDE
( By Eliminating Arbitrary Functions )

Lecture 01 on 30.09.2015 Partial Differential Equations 19 / 31

Curves/Surfaces: Expression involved with arbitrary functions (Unknown function of a known function)

Let
z = f (g(x, y)) (7)
where f is an unknown function and g is a known function of x and y.
We would like to formulate a PDE which is satisfied by (7).
Step 1: Differentiating partially (7) with respect to x and y, we get

p = f 0 (g) gx (8)

q = f 0 (g) gy (9)
0
Step 2: Eliminating function f (g) from (8) and (9), we obtain

gy p gx q = 0

which is a PDE for the unknown function z.

Lecture 01 on 30.09.2015 Partial Differential Equations 20 / 31

Example 1: Formulating PDE by eliminating arbitrary functions

Form the PDE by eliminating arbitrary functions from

z = f (x2 + y 2 ) . (10)

Here f is an arbitrary function and g is known function g(x, y) = x2 + y 2 so

that z = f (g(x, y)).
Differentiating (10) partially with respect to x and y, we get

p = f 0 (g) 2x (11)

q = f 0 (g) 2y (12)
df
where f 0 (g) = dg .
Eliminating arbitrary constant f 0 (g) from (19) and (20), we obtain a first order
PDE:
xq yp = 0.

Lecture 01 on 30.09.2015 Partial Differential Equations 21 / 31

Example 2: Formulating PDE by eliminating arbitrary functions

Form the PDE by eliminating arbitrary functions from

z = f (x ay) + g(x + ay) (13)
where a > 0 is a constant and f and g are arbitrary functions.
Differentiating (13) partially with respect to x and y, we get

z
= f 0 (x ay) + g 0 (x + ay) (14)
x
z
= f 0 (x ay) (a) + g 0 (x + ay) (a) (15)
y
2z
= f 00 (x ay) + g 00 (x + ay) (16)
x2
2z
= f 00 (x ay) (a2 ) + g 00 (x + ay) (a2 ) (17)
y 2
Eliminating arbitrary function f 00 and g 00 from (16) and (17), we obtain

2z 2
2 z
= a .
y 2 x2
Lecture 01 on 30.09.2015 Partial Differential Equations 22 / 31
Curves/Surfaces: Expression involved with arbitrary function (Unknown function of two known functions)

Let
f (g(x, y, z), h(x, y, z)) = 0 (18)
where g and h are known functions of x, y, and z and f is an
unknown/arbitrary function of g and h.
We would like to formulate a PDE which is satisfied by (18).
Step 1: Differentiating partially (18) with respect to x and y, we get
   
f g g z f h h z
+ + + =0 (19)
g x z x h x z y
   
f g g z f h h z
+ + + =0 (20)
g y z y h y z y
Step 2: Eliminating function f /g and f /h from (19) and (20), we obtain
 
 g + p g h + p h 
 x z x z  = 0
 g
 y + q g
z
h h 
y + q z 
 
g h 
(g, h) (g, h) (g, h) (g, h) g h g h  x
p +q = where = =  g x 

(y, z) (z, x) (x, y) (x, y) x y y x  y hy 
Lecture 01 on 30.09.2015 Partial Differential Equations 23 / 31
Example 3: Formulating PDE by eliminating arbitrary functions

Form the PDE by eliminating the arbitrary function from

(x2 + y 2 + z 2 , lx + my + nz) = 0.
Set g(x, y, z) = x2 + y 2 + z 2 and h(x, y, z) = lx + my + nz. Then,

(g, h) g h g h
= = 2yn 2zm
(y, z) y z z y

(g, h) g h g h
= = 2zl 2xn
(z, x) z x x z
(g, h) g h g h
= = 2xm 2yl
(x, y) x y y x
(g, h) (g, h) (g, h)
p +q = 2(ny mz)p + 2(lz nx)q = 2(nx ly)
(y, z) (z, x) (x, y)

(ny mz)p + (lz nx)q = (nx ly)

Lecture 01 on 30.09.2015 Partial Differential Equations 24 / 31

Geometrical Problem

Suppose S : z = z(x, y) is a surface with the following property:

At any point P (x0 , y0 , z(x0 , y0 )) on the surface, the tangent plane to the
surface S at the point P passes through the origin.
What differential equation the surface should satisfy?

The equation of the tangent plane at (x0 , y0 , z(x0 , y0 )) is

 
zx (x0 , y0 ) x x0 + zy (x0 , y0 ) y y0 z z(x0 , y0 ) = 0.

Since this plane passes through the origin (0, 0, 0), we have

zx (x0 , y0 )x0 zy (x0 , y0 )y0 + z(x0 , y0 ) = 0. (21)

For equation (21) to hold for all (x0 , y0 ) in the domain of z, z must satisfy

xzx + yzy z = 0,

which is a first-order PDE.

Lecture 01 on 30.09.2015 Partial Differential Equations 25 / 31
Cauchys problem or IVP for first-order PDEs
If
(a) x0 (), y0 () and z0 () are functions which, together with their first
derivatives, are continuous in the interval I = (1 , 2 ); and
(b) F (x, y, z, p, q) is a continuous function of x, y, z, p and q in a certain region
U of the xyzpq space,
then it is required to establish the existence of a function (x, y) with the
following properties:
1 (x, y) and its partial derivatives with respect to x and y are continuous
functions of x and y in a region R of the xy space.
2 For
 all values of x and y lying  in R, the point
x, y, (x, y), x (x, y), y (x, y) lies in U and


F x, y, (x, y), x (x, y), y (x, y) = 0.


3 For all belonging to the interval I, the point x0 (), y0 () belongs to
the region R, and  
x0 (), y0 () = z0 ().

Lecture 01 on 30.09.2015 Partial Differential Equations 26 / 31

Cauchys problem or IVP for first-order PDEs
From the point of view of geometry:

Find a surface z = (x, y) which passes through the smooth curve whose
parametric equation is

x = x0 (), y = y0 (), z = z0 (), I,

and at every point of which the direction (p, q, 1) of the normal is such that

F (x, y, z, p, q) = 0.

This is one of the various forms of
Cauchys problem.

The curve : x0 (), y0 (), z0 () , M , is called the initial curve of the
problem and the equation
 
x0 (), y0 () = z0 ()

is called the initial condition (or side condition) of the problem.

Lecture 01 on 30.09.2015 Partial Differential Equations 27 / 31
One such Theorem for Existence of Solutions to Cauchys Problem:

Theorem
If g(y) and all its derivatives are continuous for |y y0 | < , if x0 is a given
number, and z0 = g(y0 ), q0 = g 0 (y0 ), and if F (x, y, z, q) and all its partial
derivatives are continuous in a region S defined by

|x x0 | < , |y y0 | < , |q q0 | < ,

then there exists a unique function (x, y) such that

(a) (x, y) and all its partial derivatives are continuous in a region R defined
by |x x0 | < 1 , |y y0 | < 1 ;
(b) For all (x, y) R, z = (x, y) is a solution of the equation
 
z z
= F x, y, z, ;
x y

(c) For all values of y in the interval |y y0 | < 1 , (x0 , y) = g(y).

The above theorem is due to Sonia Kowalewski and is also known as

Cauchy-Kowalewski theorem.
Lecture 01 on 30.09.2015 Partial Differential Equations 28 / 31
About Sonia Kowalewski (Sofia Kovalevskaya)

In 1874 she presented three papers-on partial differential equations, on the dynamics of Saturns rings and on
elliptic integrals-to the University of Gottingen as her doctoral dissertation. With the support of Weierstrass, this
earned her a doctorate in mathematics summa cum laude, bypassing the usual required lectures and
examinations. The lunar crater Kovalevskaya is named in her honor. The Alexander Von Humboldt Foundation
of Germany bestows a bi-annual Sofia Kovalevskaya Award to promising young researchers.
Lecture 01 on 30.09.2015 Partial Differential Equations 29 / 31
Well-posed Problem (In the sense of Hadamard)

The Cauchys problem (PDE + side condition) is said to be well-posed if it

satisfies the following criteria:
1 The solution must exist (Existence).
2 The solution should be unique (Uniqueness).
3 The solution should depend continuously on the initial and/or boundary
data (Stability).

If one or more of the conditions above does not hold, we say that the problem
is ill-posed.

Lecture 01 on 30.09.2015 Partial Differential Equations 30 / 31

Interpretation for Integral Surfaces of Quasi-Linear 1st
Order PDE in 2 independent variables
Let f (x, y, z, p, q) = a(x, y, z)p + b(x, y, z)q c(x, y, z) = 0 be a quasi-linear
first order PDE in two independent variables.
We assume that the coefficient functions a, b and c are continuously
differentiable function in a domain D of xyz-space.
Surfaces corresponding to solutions of above PDE are called integral
surfaces.
Suppose z = z(x, y) is a smooth surface in R3 .
Then, the normal vector to the surface is the gradient vector and is given by
(zx , zy , 1) = (p, q, 1). That is, (p, q, 1) are the direction ratios of the
normal to the surface.
The above given PDE is equivalent to
h(a, b, c) (p, q, 1)i = 0 .
It says that the normal of an integral surface at any point is perpendicular to
the vector (a, b, c) corresponding to that point.
Thus integral surfaces are surfaces that at each point the vector (a, b, c) is
tangential to the surface.
Lecture 01 on 30.09.2015 Partial Differential Equations 31 / 31