0% found this document useful (0 votes)
471 views50 pages

Linear Quasi

1) The document discusses first order partial differential equations (PDEs), specifically single linear and quasilinear first order PDEs. 2) It defines first order PDEs and provides examples of different types - linear, semilinear, quasilinear, and nonlinear. 3) Characteristic equations and compatibility conditions are introduced as tools for solving first order PDEs. The general solutions of some example equations are presented.

Uploaded by

ashikhmd4467
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
471 views50 pages

Linear Quasi

1) The document discusses first order partial differential equations (PDEs), specifically single linear and quasilinear first order PDEs. 2) It defines first order PDEs and provides examples of different types - linear, semilinear, quasilinear, and nonlinear. 3) Characteristic equations and compatibility conditions are introduced as tools for solving first order PDEs. The general solutions of some example equations are presented.

Uploaded by

ashikhmd4467
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 50

First Order Partial Differential Equations, Part - 1:

Single Linear and Quasilinear First Order Equations


PHOOLAN PRASAD

DEPARTMENT OF MATHEMATICS
INDIAN INSTITUTE OF SCIENCE, BANGALORE

Definition
First order PDE in two independent variables is a relation
F (x, y; u; ux , uy ) = 0
F a known real function from D3 R5 R

(1)

Examples: Linear, semilinear, quasilinear, nonlinear equations ux + uy = 0


ux + uy = ku, c and k are constant
ux + uy = u2
uux + uy = 0
u2x u2y = 0
u2x + u2y + 1 = 0
q
ux + 1 u2y = 0, defined for |uy | 1

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(2)

2 / 50

Symbols for various domains used

In this lecture we denote


by D a domain in R2 where a solution is defined,
by D1 a domain in R2 where the coefficients of a
linear equation are defined and
by D2 is a domain in (x, y, u)-space i.e., R3
finally by D3 a domain in R5 where the function F
of five independent variables is defined.

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

3 / 50

Meaning of a classical (genuine) solution u(x, y)

u is defined on a domain D R2
u C 1 (D)
(x, y, u(x, y), ux (x, y), uy (x, y)) D3 when
(x, y) D
F (x, y; u(x, y); ux (x, y), uy (x, y)) = 0 (x, y) D.
We call a classical solution simply a solution.
Otherwise, generalized or weak solution.

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

4 / 50

Classification
Linear equation:
a(x, y)ux + b(x, y)uy = c1 (x, y)u + c2 (x, y)

(3)

Non linear equation: All other equations with subclasses


Semilinear equation:
a(x, y)ux + b(x, y)uy = c(x, y, u)

(4)

Quasilinear equation:
a(x, y, u)ux + b(x, y, u)uy = c(x, y, u)

(5)

Nonlinear equation: F (x, y; u; ux , uy ) = 0 where F is not linear in


ux , uy .
Properties of solutions of all 4 classes of equations are quite
different.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

5 / 50

Example 2
ux = 0
General solution in D = R2
u = f (y), f is an arbitrary C 1 function.

Function f is uniquely determined if u is prescribed on any curve no


where parallel to x-axis. On a line parallel to x-axis, we can not
prescribe u arbitrarily.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

6 / 50

Example 3

uy = 0

(6)

General solution in D = R2
u = f (x), f C 1(R).

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(7)

7 / 50

Directional derivative
ux = 0 means rate of change of u in direction (1, 0) parallel to xaxis
is zero
i.e. (1, 0).(ux , uy ) = 0
We say ux = 0 is a directional derivative in the direction (1, 0).
Consider a curve with parametric representation x = x(), y = y()
given by ODE
dy
dx
= a(x, y),
= b(x, y)
(8)
d
d

Tangent direction of the curve at (x, y):


(a(x, y), b(x, y))
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(9)
8 / 50

Directional derivative contd..


Rate of change of u(x, y) with as we move along this
curve is
dx
dy
du
= ux
+ uy
(10)
d
d
d
= a(x, y)ux + b(x, y)uy
which is a directional derivative in the direction (a, b)
at (x, y)
If u satisfies PDE aux + buy = c(x, y, u) then
du
= c(x, y, u)
d

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(11)

9 / 50

Characteristic equation and compatibility


condition
For the PDE
a(x, y)ux + b(x, y)uy = c(x, y)
Characteristic equations
dy
dx
= a(x, y),
= b(x, y)
d
d
and compatibility condition
du
= c(x, y)
d
(14) and (15) with a(x, y) 6= 0 characteristic equations give
dy
=
dx
compatibility condition becomes
du
=
dx
A Model Lession FD PDE Part 1

(12)

(13)

(14)

b(x, y)
a(x, y)

(15)

c(x, y)
a(x, y)

(16)

P. Prasad

Department of Mathematics

10 /
50

Example 4

yux xuy = 0

(17)

Characteristic equations are


dy
dx
= y,
= x
(18)
d
d
or
dy
x
= y dy+x dx = 0 x2+y 2 = constant.
dx
y
(19)
The characteristic curves are circles with centre
at (0,0).
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

11 /
50

Example 4 contd..

Compatibility conditions along these curves are


du
= 0 u = constant.
d

(20)

Hence value of u at (x, y) = value of u at (x, y).


u is an even function of x and also of y.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

12 /
50

Example 4 contd..
Will this even function be of the form
u = f (x2 + y 4 )?
The information
u = constant on the circles x2 + y 2 = constant
u = f (x2 + y 2 )
where f C 1 (R) is arbitrary.
Every solution is of this form.
u is an even function of x and y but of a special
form.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

13 /
50

Linear first order PDE


a(x, y)ux + b(x, y)uy = c1 (x, y)u + c2 (x, y)

(21)

Let w(x, y) be any solution of the nonhomogeneous


equation (21). Set u = v + w(x, y)
v satisfies the homogeneous equation
a(x, y)vx + b(x, y)vy = c1 (x, y)v

(22)

Let f (x, y) be a general solution of (22)


u = f (x, y) + w(x, y)
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(23)
14 /
50

Example 5: Equation with constant coefficients


aux + buy = c, a, b, c are constants

(24)

For the homogeneous equation, c = 0, characteristic


equation (with a 6= 0)
dy
b
= ay bx = constant
dx a

(25)

du
= 0 u = constant
dx

(26)

Along these

Hence u = f (ay bx) is general solution of the


homogeneous equation.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

15 /
50

Example 5: Equation with constant coefficients


contd..
For the nonhomogeneous equation, the compatibility condition
du
c
c
= u = const + x
dx
a
a

(27)

The constant here is constant along the characteristics


ay bx = const.
Hence general solution
c
u = f (ay bx) + x.
a

(28)

Alternatively u = ac x is a particular solution. Hence the result.


Solution of a PDE contains arbitrary elements. For a first order
PDE, it is an arbitrary function.
In applications - additional condition Cauchy problem.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

16 /
50

The Cauchy Problem for F (x, y; u; ux , uy ) = 0


Piecewise smooth curve
: x = x0 (), y = y0 (), I R
a given function u0 () on
Find a solution u(x, y) in a neighbourhood of
The solution takes the prescribed value u0 () on
i.e
u(x0 (), y0 ()) = u0 ()

(29)

Existence and uniqueness of solution of a Cauchy


problem requires restrictions on , the function F and
the Cauchy data u0 ().
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

17 /
50

The Cauchy Problem contd..

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

18 /
50

Example 6a
Solve yux xuy = 0 in R2
with u(x, 0) = x, x R
The solution must be an even function of x and y.
But the Cauchy data is an odd function.
Solution does not exist.

Example 6b
Solve yux xuy = 0 in a domain D
u(x, 0) = x, x R+

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(30)

19 /
50

Example 6b conti....

Solution is u(x, y) = (x2 + y 2 )1/2 , verify with partial


derivatives
x
y
ux = 2
, uy = 2
2
1/2
(x + y )
(x + y 2 )1/2

(31)

Solution is determined in R2 \ {(0, 0)}.


A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

20 /
50

Example 7
Cauchy problem: Solve
2ux + 3uy = 1
with data u| = u(, ) = u0 () on
: x y = 0; , = constant
Characteristic curve through an arbitrary point P 0 (x0 , y 0 ) in
(x, y)- plane.
dx
dy
= 2,
=3
d
d
x = x0 + 2, y = y 0 + 3
is a straight line
3x 2y = 3x0 2y 0
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

21 /
50

Example 7 contd..

If P 0 lies on , x0 = , y 0 = , the characteristic


curves are
x = + 2, y = + 3, = const, varies
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

22 /
50

Example 7 contd..
Compatibility condition
du
= 1 u = u0 (x0 , y 0 ) +
d

(32)

u = u(, ) + = u0 () +

(33)

When P 0 lies on

Solution of the Cauchy problem


Solve x = + 2, y = + 3 for and
=

x y
2y 3x
, =
2 3
2 3

(34)

Substitute in expression for u


x y
u(x, y) =
+ u0
2 3
A Model Lession FD PDE Part 1

P. Prasad

2y 3x
2 3

Department of Mathematics

(35)
23 /
50

Example 7 contd.. Existence and Uniqueness

The solution exists as long as 2 3 6= 0 i.e., the


datum curve is not a characteristic curve
Uniqueness:
Compatibility condition carries information on the
variation of u along a characteristic in unique way.
This leads to uniqueness.
What happens when 2 3 = 0?

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

24 /
50

Example 8: Characteristic Cauchy problem


2 3 = 0 datum curve is a characteristic curve
Choose = 2, = 3 x = 2, y = 3

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

25 /
50

Example 8: Characteristic Cauchy problem


contd...
The characteristic Cauchy problem: Solve
2ux + 3uy = 1
with data

1
u(2, 3) = u0 () = u0 ( x)
2

Since
du0 ()
d
=
u(2, 3) = 2ux + 3uy = 1, using PDE
d
d

(36)

The Cauchy data u0 cannot be prescribed arbitrarily on .


u0 () = = 21 x, ignoring constant of integration.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

26 /
50

Example 8 contd..
u = 21 x is a particular solution satisfying the
Cauchy data and g(3x 2y) is solution of the
homogeneous equation.
Hence
1
u = x + g(3x 2y), g C 1 and g(0) = 0 (37)
2
is a solution of the Cauchy problem.
Since g is any C 1 function with g(0) = 0, solution
of the Characteristic Cauchy problem is not unique.
We verify an important theorem in general,
solution of a characteristic Cauchy problem
does not exist and if exists, it is not unique.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

27 /
50

Quasilinear equation
a(x, y, u)ux + b(x, y, u)uy = c(x, y, u)

(38)

Since a and b depend on u, it is not possible to interpret

+ b(x, y, u)
(39)
a(x, y, u)
x
y
as a directional derivative in (x, y)- plane.
We substitute a known solution u(x, y) for u in a and b, then at

any point (x, y), it represents directional derivative


in the
direction given by
dy
dx
= a(x, y, u(x, y)),
= b(x, y, u(x, y))
(40)
d
d
Along these characteristic curves, we get compatibility condition
du
= c(x, y, u(x, y))
(41)
d
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

28 /
50

Quasilinear equation contd..


(41) is true for every solution u(x, y). The Characteristic
equations
dx
dy
= a(x, y, u),
= b(x, y, u)
d
d

(42)

along with the Compatibility condition


du
= c(x, y, u)
d
forms closed system.
Solution of 3 equations form a 2-parameter family of curves in
(x, y, u)-space are called Monge Curves and their
projections on (x, y)-plane are two parameter family of
characteristic curves.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

29 /
50

Method of solution of a Cauchy problem


Solve
a(x, y, u)ux + b(x, y, u)uy = c(x, y, u)

(43)

in a domain D containing
: x = x0 (), y = y0 ()
with Cauchy data
u(x0 (), y0 ()) = u0 ()

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(44)

30 /
50

Method of solution of a Cauchy problem contd..


Solve

dy
du
dx
= a(x, y, u),
= b(x, y, u),
= c(x, y, u)
d
d
d

(45)

(x, y, u)|=0 = (x0 (), y0 (), u0 ())

(46)

x = x(, x0 (), y0 (), u0 ()) X(, )


y = y(, x0 (), y0 (), u0 ()) Y (, )
u = u(, x0 (), y0 (), u0 ()) U (, )

(47)

Solving the first two for = (x, y), = (x, y) we get the solution
u = U ((x, y), (x, y)) u(x, y)

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(48)

31 /
50

Method of solution of a Cauchy problem contd..

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

32 /
50

Quasilinear equations conti...


Theorem:
x0 (), y0 (), u0 () C 1 (I), say I = [0, 1]
a(x, y, u), b(x, y, u), c(x, y, u)
C 1 (D2 ), where D2 is a domain in (x, y, u) space
D2 contains curve in (x, y, u)-space
: x = x0 (), y = y0 (), u = u0 (), I
dy0
d a(x0 (), y0 (), u0 ())

dx0
d b(x0 (), y0 (), u0 ())

6= 0, I

There exists a unique solution of the Cauchy problem in a domain D


containing I.

Do not worry about the complex statement in the


theorem. As long as the datum curve is not
tangential to a characteristic curve and the functions
involved are smooth, the solution exist and is unique
(see previous slide).
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

33 /
50

Example 9
Cauchy problem
ux + uy = u
u(x, 0) = 1
x0 = , y0 = 0, u0 = 1

(49)

Step 1. Characteristic curves


dx
=1x=+
d
dy
=1y=
d
Step 2. Therefore
= y, = x y
Step 3. Compatibility condition
du
= u u = u0 ()e = e
d
Step 4. Solution u = ey
exists on D = R2
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(50)

34 /
50

Example 10
Cauchy problem
ux + uy = u2
u(x, 0) = 1
x0 = , y0 = 0, u0 = 1

(51)

Step 1. Characteristic equations give


x = + , y =
Step 2. Compatibility condition gives
1
du
= u2 u =
d
u0 ()
1
Step 3. Solution u = 1y
exists on the domain D = y < 1 and u + as y 1.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

35 /
50

Example 11
Cauchy problem
uux + uy = 0
u(x, 0) = x, 0 x 1

(52)

x = , y = 0, u = , 0 1 at = 0
Step 1. Characteristic equations and compatibility
condition
dx
dy
du
= u,
= 1,
=0
(53)
d
d
d
Step 2. Quasilinear equations, characteristics depend
on the solution
u=
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

36 /
50

Example 11 contd..
Step 3.
x = ( + 1), y =
Step 4. From solution of characteristic equations
x
= y and = y+1
x
Step 5. Solution is u = y+1
, but D =?
Step 6. Characteristic curves are straight lines
x
= , 0 |
y+1
which meet at the point (1, 0).
Step 7. u is constant on these characteristics (see next
slide).
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

37 /
50

Solution is determined in a wedged shaped region


in (x, y)-plane including the lines y = 0 and y = x + 1.
Figure:

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

38 /
50

Example 12
Cauchy problem
uux + uy = 0
u(x, 0) = 12 , 0 x 1

(54)

Step 1. Parametrization of Cauchy data


1
x0 = , y0 = 0, u0 = , 0 1.
2
Step 2. The compatibility condition gives
1
u = constant = .
2
Step 3. The characteristic curves are
y 2x = 2, 0 1.
on which solution has the same value u = 12 .
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(55)
39 /
50

Example 12 contd..
Step 4. The solution u = 12 of the Cauchy problem is determined in an
infinite strip 2x y 2x 2 in (x, y)-plane.

Important: From examples 11 and 12, we notice that the domain,


where solution of a Cauchy problem for a quasilinear equation is
determined, depends on the Cauchy data.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

40 /
50

General solution
General solution contains an arbitrary function.
Theorem : If (x, y, u) = C1 and (x, y, u) = C2 be
two independent first integrals of the ODEs
dy
du
dx
=
=
a(x, y, u)
b(x, y, u)
c(x, y, u)

(56)

and 2u + u2 6= 0, the general solution of the PDE


aux + buy = c is given by
h((x, y, u), (x, y, u)) = 0

(57)

where h is an arbitrary function.


A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

41 /
50

Example 13
uux + uy = 0

(58)

dy
du
dx
=
=
(59)
u
1
0
Note 0 appearing in a denominator to be properly interpreted
u = C1
x C1 y = C2
x uy = C2

(60)

General solution is given by


(u, x uy) = 0
or u = f (x uy)

(61)

where h and f arbitrary functions.


Note : Solution of this nonlinear equation may be very difficult.
Numerical method is generally used.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

42 /
50

Example 14
Consider the differential equation
1
(y + 2ux)ux (x + 2uy)uy = (x2 y 2 )
2

(62)

The characteristic equations and the compatibility condition are


dx
dy
du
=
= 1 2
y + 2ux
(x + 2uy)
(x y 2 )
2

(63)

To get one first integral we derive from these


2du
xdx + ydy
= 2
2
2
2u(x y )
x y2

(64)

which immediately leads to


(x, y, u) x2 + y 2 4u2 = C1
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

(65)
43 /
50

Example 14 contd..
For another independent first integral we derive a
second combination
2du
ydx + xdy
=
y 2 x2
x2 y 2

(66)

(x, y, u) xy + 2u = C2

(67)

which leads to

The general integral of the equation (55)is given by


h(x2 + y 2 4u2 , xy + 2u) = 0
x2 + y 2 4u2 = f (xy + 2u)

(68)

where h or f are arbitrary functions of their arguments.


A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

44 /
50

Example 14 contd..
Consider a Cauchy problem u = 0 on x y = 0
x = , y = , u = 0
From (58) and (60) we get 2 2 = C1 and 2 = C 2 which
gives C1 = 2C2 . Therefore, the solution of the Cauchy
problem is obtained, when we take h(, ) = 2.
This gives, taking only the suitable one,
o
1 np
2
u=
(x y) + 1 1
(69)
2
We note that the solution of the Cauchy problem is
determined uniquely at all points in the (x, y)-plane.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

45 /
50

Exercise
1. Show that all the characteristic curves of the partial
differential equation
(2x + u)ux + (2y + u)uy = u
through the point (1,1) are given by the same straight
line x y = 0
2. Discuss the solution of the differential equation
uux + uy = 0, y > 0, < x <
with Cauchy data
(
2 x2
u(x, 0) =
0
A Model Lession FD PDE Part 1

P. Prasad

for |x|
for |x| > .
Department of Mathematics

(70)
46 /
50

Exercise contd..
3. Find the solution of the differential equation

m 
1 u ux mM uy = 0
r
satisfying
My
u(0, y) =
y
where m, r, , M are constants, in a neighourhood of
the point x = 0, y = 0.
4. Find the general integral of the equation
(2x y)y 2 ux + 8(y 2x)x2 uy = 2(4x2 + y 2 )u
and deduce the solution of the Cauchy problem when
1
the u(x, 0) = 2x
on a portion of the x-axis.
A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

47 /
50

R. Courant and D. Hilbert. Methods of Mathematical Physics,


vol 2: Partial Differential Equations. Interscience Publishers, New
York, 1962.

L. C. Evans. Partial Differential Equations. Graduate Studies in


Mathematics, Vol 19, American Mathematical Society, 1999.

F. John. Partial Differential Equations. Springer-Verlag, New


York, 1982.

P. Prasad. (1997) Nonlinearity, Conservation Law and Shocks,


Part I: Genuine Nonlinearity and Discontinuous Solutions,
RESONANCE- Journal of Science Education by Indian Academy
of Sciences, Bangalore, Vol-2, No.2, 8-18.
P. Prasad. (1997) Nonlinearity, Conservation Law and Shocks,
Part II: Stability Consideration and Examples, RESONANCEJournal of Science Education by Indian Academy of Sciences,
Bangalore, Vol-2, No.7, 8-19.

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

48 /
50

P. Prasad. Nonlinear Hyperbolic Waves in Multi-dimensions.


Monographs and Surveys in Pure and Applied Mathematics,
Chapman and Hall/CRC, 121, 2001.

P. Prasad. A theory of first order PDE through propagation of


discontinuities. Ramanujan Mathematical Society News Letter,
2000, 10, 89-103; see also the webpage:

P. Prasad and R. Ravindran. Partial Differential Equations.


Wiley Eastern Ltd, New Delhi, 1985.

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

49 /
50

Thank You!

A Model Lession FD PDE Part 1

P. Prasad

Department of Mathematics

50 /
50

You might also like