@brownearnest

KENYATTA UNIVERSITY

MATH 421: PARTIAL DIFFERENTIAL EQUATIONS 1

WRITTEN BY: Dr. Isaac Chepkwony

VETTED BY: Dr. Amos Magua

Copyright ©Kenyatta University, 2020

All Rights Reserved
Published By: KENYATTA UNIVERSITY PRESS

1
|Purchased By: Brown Earnest|
Contents
Purpose of the Module ................................................................................................................................. 5
Module Learning Outcomes.......................................................................................................................... 5
Course Description ........................................................................................................................................ 5
LESSON 1 ....................................................................................................................................................... 6
CONCEPTS OF PARTIAL DIFFERENTIAL EQUATIONS ..................................................................................... 6
LESSON 2 ..................................................................................................................................................... 12
Formation of Partial Differential Equations ................................................................................................ 12
LESSON THREE............................................................................................................................................. 18
SOLUTION OF FIRST ORDER QUASILINEAR PDEs BY THE LAGRANGE’S METHOD OF TYPE ONE AND TWO18
LESSON FOUR .............................................................................................................................................. 23
SOLUTION OF FIRST ORDER QUASILINEAR PDEs BY THE LAGRANGE’S METHOD OF TYPES THREE AND
FOUR ........................................................................................................................................................... 23
LESSON FIVE ................................................................................................................................................ 28
INTEGRAL SURFACES PASSING THROUGH GIVEN CURVES AND SURFACES ORTHOGONAL TO A GIVEN
SYSTEM OF SURFACES................................................................................................................................. 28
LESSON SIX .................................................................................................................................................. 35
THE METHOD OF CANONICAL FORMS ........................................................................................................ 35
LESSON SEVEN ............................................................................................................................................ 40
THE METHOD OF SEPARATION OF VARIABLES ........................................................................................... 40
LESSON EIGHT ............................................................................................................................................. 45
Non-linear Partial Differential Equations of Order One.............................................................................. 45
LESSON NINE ............................................................................................................................................... 56
NON-LINEAR PARTIAL DIFFERENTIAL EQUATIONS OF TYPES THREE AND FOUR ........................................ 56
LESSON TEN................................................................................................................................................. 65
SOLUTION OF NON-LINEAR PDEs USING TRANSFORMATIONS .................................................................. 65
LESSON ELEVEN........................................................................................................................................... 74
CHARPIT’S METHOD OF SOLVING FIRST ORDER NON-LINEAR PARTIAL DIFFERENTIAL EQUATIONS ......... 74

3
Purpose of the Module
The aim of this course is to introduce the students to the theory of first order partial differential
equations (PDE) and give competence in solving them via analytical methods.

Module Learning Outcomes

By the end of this module the learner will be able to:
i) Classify any first order PDE into linear, semi-linear, quasilinear and nonlinearPDE;
ii) Form PDE by eliminating arbitrary constants and arbitrary functions in given relations;
iii) SolvePDE in Lagrange’s form using auxiliary methods;
iv) Solve first order PDE through the use of characteristics;
v) Find a specific solution of a first order Cauchy problem;
vi) Identify first order non-linearPDE as belonging to some special types and hence solve
them;

vii) Use Charpit’s and Jacobi methods to solve first order non-linearPDEs.

Course Description
Partial differential equations of the first order, first degree. Solutions using Lagrange’s system of
linear equations. Linear, semi-linear and quasi-linear partial differential equations of the first
order. Integral surface passing through a given curve (Cauchy problem). Use of the methods of
Cauchy, Charpit and Jacobi in solving non-linear partial differential equations of the first order.

5
LESSON 1

CONCEPTS OF PARTIAL DIFFERENTIAL EQUATIONS

1.1 Introduction
In this lesson we will study first order partial differential equations. Partial differential equations,
just like ordinary differential equations, are functional equations. A functional equation is an
equation where the unknown is a function.

1.2 Learning Outcomes

By the end of this lesson the learner will be able to:

1.2.1 Define a partial differential equation , order and degree of a PDE;

1.2.2 Classify first order partial differential equations

1.3 PARTIAL DIFFERENTIAL EQUATIONS

Partial differential equations, just like ordinary differential equations, are functional equations.

A functional equation is an equation where the unknown is a function.

Definition 1. A differential equation that contains a dependent variable of more than one
independent variables and one or more partial derivatives is called a partial differential
equation. In general, it may be written in the form

𝑓(𝑥, 𝑦, … … . . , 𝑢, 𝑢𝑥 , 𝑢𝑦 , … … . . , 𝑢𝑥𝑥 , 𝑢𝑥𝑦 , 𝑢𝑦𝑦 , … … . ) = 0, (1.1)

involving several independent variables 𝑥, 𝑦, … ., an unknown function 𝑢 of these variables, and

the partial derivatives 𝑢𝑥 , 𝑢𝑦 , … . . , 𝑢𝑥𝑥 , 𝑢𝑥𝑦 , 𝑢𝑦𝑦 , … … … … of the function. Subscripts on
𝜕𝑢 𝜕2𝑢
dependent variables denote differentiations e.g. 𝑢𝑥 = 𝜕𝑥 , 𝑢𝑥𝑦 = 𝜕𝑥𝜕𝑦.

Here equation (1.1) is considered in a suitable domain 𝐷 of the 𝑛 −dimensional space 𝑅 𝑛 in the
independent variables 𝑥, 𝑦, … … …. We seek functions 𝑢 = 𝑢(𝑥, 𝑦, … . ) which satisfy equation
(1.1) identically in 𝐷. Such functions, if they exist, are called solutions of equation (1.1). From
these many solutions we attempt to select a particular one by introducing additional conditions.

For instance

6
 𝑢𝑢𝑥𝑦 + 𝑢𝑥 = 𝑦,

𝑢𝑥𝑥 + 2𝑦𝑢𝑥𝑦 + 3𝑥𝑢𝑦𝑦 = 4𝑠𝑖𝑛𝑥,

2
(𝑢𝑥 )2 + (𝑢𝑦 ) = 1,

𝑢𝑥𝑥 − 𝑢𝑦𝑦 (∗)

are partial differential equations. The functions 𝑢(𝑥, 𝑦) = (𝑥 + 𝑦)3 and 𝑢(𝑥, 𝑦) = sin(𝑥 − 𝑦),
are solutions of (∗), as can easily be verified.

Definition 2

The order of a partial differential equation is the order of the highest-ordered partial derivative
appearing in the equation. For example, considering 𝑢 as the dependent variable and 𝑥, 𝑦 as
independent variables,
𝜕𝑢 𝜕𝑢
1. 𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = 𝑢 or 𝑥𝑝 + 𝑦𝑞 = 𝑢 is of order one and
𝜕2𝑢 𝜕2𝑢 𝜕2𝑢
2. + 3 𝜕𝑥𝜕𝑦 + 𝜕𝑦 2 = 0 or 𝑟 + 3𝑠 + 𝑡 = 0 is of order two.
𝜕𝑥 2

𝜕𝑢 𝜕𝑢 𝜕2𝑢 𝜕2𝑢 𝜕2𝑢

We have used the standard notation: 𝑝 = 𝜕𝑥 , 𝑞 = 𝜕𝑦 , 𝑟 = 𝜕𝑥 2 , 𝑠 = 𝜕𝑥𝜕𝑦 , 𝑡 = 𝜕𝑦 2

The degree of a PDE is the degree of the highest order derivative present in the PDE after
clearing the fractional powers.

For example
𝜕𝑢 𝜕𝑢
+ 2 𝜕𝑦 = 𝑢 (1)
𝜕𝑥

𝜕𝑢 𝜕2𝑢 𝜕2𝑢
= 𝑐 2 (𝜕𝑥 2 + 𝜕𝑦 2 ) (2)
𝜕𝑡

3
𝜕2𝑢 𝜕𝑢 4
= (𝜕𝑦) (3)
𝜕𝑥𝜕𝑦

Equations (1) and (2) are of first degree whereas equation (3) is of 4th degree.

Classification of first order PDEs

Partial differential equations can be classified in at least three ways. They are

1. Order of the PDE

2. Linear, Semi-linear, Quasi-linear and fully non-linear
7
 3. Scalar equation, system of equations

NOTE

a) A PDE of order 𝑚 is called quasi-linear if it is linear in the derivative of order 𝑚 with

coefficients that depend on the independent variables and derivatives of the unknown
function of order strictly less than 𝑚.
b) A quasilinear PDE where the coefficients of the derivatives of order 𝑚 are functions of
the independent variables alone is called a semi-linear PDE.
c) A PDE which is linear in the unknown function and all its derivatives with coefficients
depending on the independent variables alone is called a linear PDE.
d) A PDE which is not quasi-linear is called a fully non-linear PDE.
We have the following picture
Linear PDE⋤ Semi-linear PDE ⋤ Quasi-linear PDE⋤ Fully non-linear PDE

Remark
a) A single first order quasi-linear PDE must be of the form
𝑎(𝑥, 𝑦, 𝑢)𝑢𝑥 + 𝑏(𝑥, 𝑦, 𝑢)𝑢𝑦 = 𝑐(𝑥, 𝑦, 𝑢)
b) A single quasi-linear PDE where 𝑎, 𝑏 are functions of 𝑥 and 𝑦 alone is a semi-linear PDE.
c) A single semi-linear PDE where 𝑐(𝑥, 𝑦, 𝑢) = 𝑐0 (𝑥, 𝑦)𝑢 + 𝑐1 (𝑥, 𝑦) is a linear PDE.
d) Linear PDEs can further be classified into two: homogeneous and nonhomogeneous.
Every linear PDE can be written in the form 𝐿[𝑢] = 𝑓, where 𝑢 → 𝐿[𝑢] is a linear map,
and 𝑓 is a function of independent variables only. This linear PDE is said to be
homogeneous if 𝑓 is the zero function; otherwise it is called a non-homogeneous linear
PDE.
Examples
i. 𝑥𝑢𝑥 + 𝑦𝑢𝑦 − 2𝑢 = 0 is a linear homogenous PDE
ii. 𝑦𝑢𝑥 + 𝑢𝑦 − 𝑥𝑦 = 0 is a linear non-homogeneous PDE
iii. 𝑥𝑢𝑥 + 𝑦𝑢𝑦 − 𝑢2 − 𝑥 2 = 0 is a semi-linear PDE
iv. (𝑦 2 − 𝑢2 )𝑢𝑥 − 𝑥𝑦𝑢𝑦 = 𝑥𝑢 is a quasilinear
v. 𝑢𝑥2 + 𝑢𝑦 = 𝑥𝑦 is a fully non-linear PDE
𝑢𝑥 + 𝑢𝑦 = 2
vi. { is a system of first order PDEs
2𝑢𝑥 − 𝑦𝑢𝑦 = 0

8
Properly posed problems in the sense of Hadamard
A mathematical problem is said to be well-posed or properly-posed in the sense of
Hadamard if the following three conditions are satisfied:
1. The problem should admit at least one solution.
2. The problem should admit at most one solution.
3. The solution should dependent continuously on the data in the problem.

9
E-tivity1.2: Concepts of partial differential equations

Numbering and Pacing and 1.2.1 – 1.2.2

Sequencing
Title Partial Differential Equations

Purpose To introduce you to the language and concepts of partial

differential equations.
Brief summary of overall task Watch the PDEs by MATHITUPCANADA
and then attempt the given questions.
Spark

Solutions of PDEs may look like

this.
Individual contribution • Watch the video on PDEs
• Classify the following PDEs
a) 𝑥𝑢𝑥 + 2𝑢𝑦 = 𝑢2
b) 𝑢𝑢𝑥 + 𝑢2 𝑢𝑦3 = 2
c) 𝑢𝑥 + (1 + 𝑢)𝑢𝑦 = 0

Interaction begins • Post your answers on the discussion forum 1.2.3

• Read what your colleagues have posted.
• In a sentence or two, comment on what your
colleagues have posted.

E-moderator interventions • Focussing group discussion

• Encouraging quiet ones to contribute
• Providing feedback/ teaching points
• Summarising key points
• Closing the discussion
Schedule and time This activity should take one hour
Next Formation of PDEs

10
1.4 Assessment
1. We know the following classification of PDEs
Linear PDE⋤ Semi-linear PDE ⋤ Quasi-linear PDE⋤ Fully non-linear PDE
Each of the above inclusions is a strict inclusion. Justify the statement by giving
examples.
2. Give at least three examples of fifth order PDE belonging to each of the above (i.e
exer.1.) classes.
3. Classify the following equations by all the three ways of classification.
𝜕𝑢 2 𝜕2 𝑢
i. (𝜕𝑦) + 𝜕𝑥 2 = 1
𝜕𝑢 2
ii. 𝑠𝑖𝑛 (1 + 𝜕𝑥 ) + 𝑢3 = 𝑠𝑖𝑛𝑦
𝜕2𝑢 𝜕2𝑢
iii. + =0
𝜕𝑥 2 𝜕𝑦 2
𝜕2 𝑢 𝜕2 𝑢
+ 2
iv. 𝑒 𝜕𝑥2 𝜕𝑦 =1
𝜕2𝑢 𝜕2𝑢
v. − 𝜕𝑥 2 = 𝑐𝑜𝑠𝑢
𝜕𝑡 2

1.5 References
1. Tyn Myint-U and Lokenath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.
2. Frank, Ayres (1952). Theory and problems of Differential Equations. McGRAW-HILL
Publishers.

11
 LESSON 2

Formation of Partial Differential Equations

2.1 Introduction
In this lesson we will learn how to form partial differential equations. The PDEs are formed either by
elimination of arbitrary constants or by elimination of arbitrary functions from a relation with
one dependent variable and two or more independent variables.

2.2 Learning Outcomes

By the end of this lesson the learner will be able to:

2.2.1 Form PDEs by elimination of arbitrary constants;

2.2.2 Form PDEs by elimination of arbitrary constants

2.3 Formation of partial differential equations

A. When a PDE is formed by eliminating arbitrary constants:

1. If the number of arbitrary constants are more than the number of independent variables in
the given relations, the PDE obtained will be of a second or higher order.
2. If the number of arbitrary constants equals the number of independent variables in the
given relation then the PDE obtained will be of first order.

B. When PDE is formed by eliminating arbitrary functions:

When 𝑛 is the number of arbitrary functions, we may get several PDEs but out of which
generally one with two least order is selected.
𝜕4𝑢
e.g 𝑢 = 𝑥𝑓(𝑦) + 𝑦𝑔(𝑥) involves two arbitrary constants, 𝑓 and 𝑔. Here =0
𝜕𝑥 2 𝜕𝑦 2
and 𝑥𝑦𝑠 = 𝑥𝑝 + 𝑦𝑞 − 𝑧 are two possible PDEs obtained by the arbitrary functions.
However, 2nd equation being lower in order to the 1st is the desired PDE.

Elimination of arbitrary constants

Consider 𝑢 to be a function of two independent variables 𝑥 and 𝑦 defined by
𝑔(𝑥, 𝑦, 𝑢, 𝑎, 𝑏) = 0(1)
in which 𝑎 and 𝑏 are two arbitrary constants. By differentiating (1) partially with respect to 𝑥
and 𝑦, we obtain
𝜕𝑔 𝜕𝑔 𝜕𝑢 𝜕𝑔 𝜕𝑔
+ = +𝑝 = 0(2)
𝜕𝑥 𝜕𝑢 𝜕𝑥 𝜕𝑥 𝜕𝑢
𝜕𝑔 𝜕𝑔 𝜕𝑢 𝜕𝑔 𝜕𝑔
And + 𝜕𝑦
= + 𝑞 = 0. (3)
𝜕𝑢 𝜕𝑦 𝜕𝑦 𝜕𝑢

12
In general, the arbitrary constants may be eliminated from (1), (2), (3) yielding a partial
differential equation of order one 𝑓(𝑥, 𝑦, 𝑢, 𝑝, 𝑞) = 0.

Example 2.1
Form a PDE by eliminating the arbitrary constants 𝑎 and 𝑏 from
i. 𝑢 = 𝑎𝑥 + 𝑏𝑦 + 𝑎𝑏
ii. 𝑢 = 𝑎𝑥 + 𝑎2 𝑦 2 + 𝑏
iii. 𝑢 = (𝑥 2 + 𝑎)(𝑦 2 + 𝑏)
iv. 𝑢 = (𝑥 − 𝑎)2 + (𝑦 − 𝑏)2
v. 𝑢 = 𝑎𝑥 + (1 − 𝑎)𝑦 + 𝑏

Solutions
𝜕𝑢 𝜕𝑢
i. 𝑝 = 𝜕𝑥 = 𝑎, 𝑞 = 𝜕𝑦 = 𝑏
Substituting for 𝑎 and 𝑏 in 𝑢 = 𝑎𝑥 + 𝑏𝑦 + 𝑎𝑏 we get 𝑢 = 𝑝𝑥 + 𝑞𝑦 + 𝑝𝑞, a linear PDE
of first order.
𝜕𝑢 𝜕𝑢
ii. 𝑝 = 𝜕𝑥 = 𝑎, 𝑞 = 𝜕𝑦 = 2𝑎2 𝑦 Eliminating 𝑎 from these results, we get
𝑞 = 2𝑎2 𝑦 = 2𝑝2 𝑦 which is the required non-linear PDE of first order.
𝜕𝑢 𝜕𝑢 𝑝
iii. 𝑝 = 𝜕𝑥 = 2𝑥(𝑦 2 + 𝑏), 𝑞 = 𝜕𝑦 = 2𝑦(𝑥 2 + 𝑎). Therefore 𝑦 2 + 𝑏 = 2𝑥 and
𝑞 𝑝 𝑞 𝑝𝑞
𝑥 2 + 𝑎 = 2𝑦. Substituting these in 𝑢 = (𝑥 2 + 𝑎)(𝑦 2 + 𝑏), we get 𝑢 = (2𝑥) (2𝑦) = 4𝑥𝑦
or 4𝑥𝑦 = 𝑝𝑞, a non-linear first order PDE.
𝜕𝑢 𝜕𝑢 𝑝 𝑞 𝑝2 𝑞2
iv. 𝑝 = 𝜕𝑥 = 2(𝑥 − 𝑎), 𝑞 = 𝜕𝑦 = 2(𝑦 − 𝑏) ⇒ 𝑥 −𝑎 = 2, 𝑦−𝑏 = ⇒𝑢= +
2 4 4
2 2
or 4𝑢 = 𝑝 + 𝑞 , a non-linear first order PDE
𝜕𝑢 𝜕𝑢
v. 𝑝 = 𝜕𝑥 = 𝑎, 𝑞 = 𝜕𝑦 = 1 − 𝑎⇒𝑝 + 𝑞 = 1

Example 2.2

𝑥2 𝑦2 𝑢2
Form a PDE by eliminating the arbitrary constants 𝑎, 𝑏 and 𝑐 from the relation + 𝑏2 + 𝑐 2 = 1
𝑎2
Solution

Note that 𝑎, 𝑏 and 𝑐 are arbitrary constants and 𝑢 is the dependent variable, depending

𝑥2 𝑦2 𝑢2
on 𝑥 and 𝑦. We can write the given relation as 𝑓(𝑥, 𝑦, 𝑢) = (𝑎2 + 𝑏2 + 𝑐 2 − 1) = 0(1)

𝜕𝑓 𝜕𝑓 𝜕𝑢
Then differentiating (1) partially w.r.t 𝑥 and 𝑦 respectively, we have 𝜕𝑥 + 𝜕𝑢 𝜕𝑥 = 0 and

13
𝜕𝑓 𝜕𝑓 𝜕𝑢 2𝑥 2𝑢 𝜕𝑢 2𝑦 2𝑢 𝜕𝑢
+ 𝜕𝑢 𝜕𝑦 = 0⇒𝑎2 + 𝑐 2 𝜕𝑥 = 0 and 𝑏2 + 𝑐 2 𝜕𝑦 = 0 or 𝑐 2 𝑥 + 𝑎2 𝑢𝑝 = 0 (2)
𝜕𝑦

and 𝑐 2 𝑦 + 𝑏 2 𝑢𝑞 = 0 (3)

𝜕𝑢 2 𝜕2𝑢
Differentiating (2) w.r.t 𝑥 we get 𝑐 2 + 𝑎2 (𝜕𝑥 ) + 𝑎2 𝑢 𝜕𝑥 2 = 0

𝑐2 𝑢 𝜕𝑢
On substituting 𝑎2 = − 𝑥 𝜕𝑥 from (2) in the above equation, we get

𝑢 𝜕𝑢 𝜕𝑢 2 𝜕 2𝑢
− +( ) +𝑧 2 =0
𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑥
𝜕2 𝑢 𝜕𝑢 2 𝜕𝑢
Or 𝑥𝑢 𝜕𝑥 2 + 𝑥 (𝜕𝑥 ) − 𝑢 𝜕𝑥 = 0(4)

𝜕𝑢 2 𝜕2𝑢
Similarly differentiating (3) w.r.t y we get 𝑐 2 + 𝑏 2 (𝜕𝑦) + 𝑏 2 𝑢 𝜕𝑦 2 = 0 and

𝑐2 𝑢 𝜕𝑢
substituting 𝑏2 = − 𝑦 𝜕𝑦 from (3) in the above equation, we get

𝜕2𝑢 𝜕𝑢 2 𝜕𝑢
𝑦𝑢 𝜕𝑦 2 + 𝑦 (𝜕𝑦) − 𝑢 𝜕𝑦 = 0 (5)

Thus equations (4) and (5) are partial differential equations of first degree and second order.

Formation of PDEs by elimination of arbitrary functions

Let 𝑈 = 𝑈(𝑥, 𝑦, 𝑢) and 𝑉 = 𝑉(𝑥, 𝑦, 𝑢) be the dependent variables of the variables 𝑥, 𝑦, 𝑢 and 𝑢 is
a function of 𝑥 and 𝑦. Let
𝜑(𝑈, 𝑉) = 0 (1)

be an arbitrary relation between them. Partially differentiating w.r.t 𝑥 and 𝑦, we obtain

𝜕𝜑 𝜕𝑈 𝜕𝑈 𝜕𝜑 𝜕𝑉 𝜕𝑉
( +𝑝 )+ ( +𝑝 )=0 (2)
𝜕𝑈 𝜕𝑥 𝜕𝑢 𝜕𝑉 𝜕𝑥 𝜕𝑢
And
𝜕𝜑 𝜕𝑈 𝜕𝑈 𝜕𝜑 𝜕𝑉 𝜕𝑉
( + 𝑞 𝜕𝑢 ) + 𝜕𝑉 (𝜕𝑦 + 𝑞 𝜕𝑢) = 0 (3)
𝜕𝑈 𝜕𝑦
𝜕𝜑 𝜕𝜑
Eliminating 𝜕𝑈 and from (2) and (3), we have
𝜕𝑉
𝜕𝑈 𝜕𝑈 𝜕𝑉 𝜕𝑉
+ 𝑝 𝜕𝑢 + 𝑝 𝜕𝑢 𝜕𝑈 𝜕𝑈 𝜕𝑉 𝜕𝑉 𝜕𝑈 𝜕𝑈 𝜕𝑉 𝜕𝑉
𝜕𝑥 𝜕𝑥
|𝜕𝑈 𝜕𝑈 𝜕𝑉 𝜕𝑉|=( 𝜕𝑥 + 𝑝 𝜕𝑢 ) (𝜕𝑦 + 𝑞 𝜕𝑢) − ( 𝜕𝑦 + 𝑞 𝜕𝑢 ) (𝜕𝑥 + 𝑝 𝜕𝑢)
𝜕𝑦
+ 𝑞 𝜕𝑢 𝜕𝑦
+ 𝑞 𝜕𝑢

14
 𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉
= − + 𝑝 ( 𝜕𝑢 𝜕𝑦 − 𝜕𝑦 𝜕𝑢) + 𝑞 ( 𝜕𝑥 𝜕𝑢 − 𝜕𝑢 𝜕𝑥 ) = 0
𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑥
𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉
Writing 𝜆𝑃 = − 𝜕𝑢 𝜕𝑦 , 𝜆𝑄 = − 𝜕𝑥 𝜕𝑢 , 𝜆𝑅 = − 𝜕𝑦 𝜕𝑥 , this takes the form
𝜕𝑦 𝜕𝑢 𝜕𝑢 𝜕𝑥 𝜕𝑥 𝜕𝑦
𝑃𝑝 + 𝑄𝑞 = 𝑅, a partial differential equation linear in 𝑝 and 𝑞 and free of the arbitrary function
𝜑(𝑈, 𝑉)

Example 2.3
Find the differential equation arising from
1. 𝑓(𝑥 2 + 𝑦 2 + 𝑢2 , 𝑢2 − 2𝑥𝑦) = 0
2. 𝑥 + 𝑦 + 𝑢 = 𝑓(𝑥 2 + 𝑦 2 + 𝑢2 )
3. 𝑢 = 𝑓(𝑥 2 − 𝑦 2 )
Solution

1. We write the functional relation in the form 𝜑(𝑈, 𝑉) = 0 with 𝑈 = 𝑥 2 + 𝑦 2 + 𝑢2 and

𝑉 = 𝑢2 − 2𝑥𝑦. Differentiating partially w.r.t 𝑥 and 𝑦, we have
𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
(𝑥 + 𝑝𝑢) + (−𝑦 + 𝑝𝑢) = 0⇒ (𝑥 + 𝑝𝑢)= - (−𝑦 + 𝑝𝑢)(𝑖)
𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉
𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
(𝑦 + 𝑞𝑢) + (−𝑥 + 𝑞𝑢) = 0⇒ (𝑦 + 𝑞𝑢)= - (−𝑥 + 𝑞𝑢) (𝑖𝑖)
𝜕𝑈 𝜕𝑉 𝜕𝑈 𝜕𝑉
𝜕𝑓 𝜕𝑓
Eliminating 𝜕𝑈 and 𝜕𝑉 from the two equations by dividing (𝑖) by (𝑖𝑖), we get

𝑦 2 − 𝑥 2 + 𝑢(𝑦 + 𝑥)(𝑞 − 𝑝) = 0, a non-linear first order PDE.

2. We write the functional relation in the form 𝜑(𝑈, 𝑉) = 0 with 𝑈 = 𝑥 2 + 𝑦 2 + 𝑢2 and

𝑉 = 𝑥 + 𝑦 + 𝑢. Differentiating partially w.r.t 𝑥 and 𝑦, we have
𝜕𝑢 𝜕𝑓 𝜕𝑢 𝜕𝑓
1 + 𝜕𝑥 = 𝜕𝑈 (2𝑥 + 2𝑢 𝜕𝑥 )⇒1 + 𝑝 = 𝜕𝑈 (2𝑥 + 2𝑝𝑢)(𝑖)
And
𝜕𝑢 𝜕𝑓 𝜕𝑢 𝜕𝑓
1 + 𝜕𝑦 = 𝜕𝑈 (2𝑦 + 2𝑢 𝜕𝑦)⇒1 + 𝑞 = 𝜕𝑈 (2𝑦 + 2𝑞𝑢)(𝑖𝑖)
𝜕𝑓
Eliminating 𝜕𝑈 by dividing (𝑖) by (𝑖𝑖), we get
𝑥 − 𝑦 = (𝑦 − 𝑢)𝑝 + (𝑢 − 𝑥)𝑞, a non-linear first order PDE.
3. We write the functional relation in the form 𝜑(𝑈, 𝑉) = 0 with 𝑈 = 𝑥 2 − 𝑦 2 and 𝑉 = 𝑢.
Differentiating 𝑢 = 𝑓(𝑥 2 − 𝑦 2 ) partially w.r.t 𝑥 and 𝑦, we have
𝜕𝑢 𝜕𝑓 𝜕𝑓
= 𝜕𝑈 (2𝑥)⇒𝑝 = 2𝑥 𝜕𝑈 (𝑖)
𝜕𝑥
𝜕𝑢 𝜕𝑓 𝜕𝑓
= 𝜕𝑈 (−2𝑦)⇒𝑞 = −2𝑦 𝜕𝑈 (𝑖𝑖)
𝜕𝑦
𝜕𝑓
Eliminating by dividing (𝑖) by (𝑖𝑖), we get
𝜕𝑈
𝑦𝑝 + 𝑥𝑞 = 0, a linear first order PDE.

15
E-tivity 2.2: Formation of PDEs

Numbering and Pacing and 2.2.1 – 2.2.3

Sequencing
Title Formation of PDEs

Purpose To expose you to formation of PDEs through

elimination of arbitrary constants and arbitrary
functions..
Brief summary of overall task Watch the video formation of PDEs by MATHSOLVES
ZONE
and then attempt the questions given.
Spark
FORMATION OF PDEs

Individual contribution • Watch the video on formation of PDEs

• Form the PDEs from
a) 𝑧 = 𝑓(𝑥 − 𝑦)

b) (𝑥 − 𝑎)2 +(𝑦 − 𝑏)2 +𝑧 2 = 1

Interaction begins • Post your answers on the discussion forum 2.2.3

• Read what your colleagues have posted.
• In a sentence or two, comment on what two of
your colleagues have posted.

E-moderator interventions • Focussing group discussion

• Encouragingquiet ones to contribute
• Providing feedback/ teaching points
• Summarising key points
• Closing the discussion
Schedule and time This activity should take one hour
Next Solution of quasilinear PDEs

2.4 Assessment

16
1. Form PDEs by eliminating arbitrary constants from the following relations
(note that 𝑧 = 𝑧(𝑥, 𝑦))

a) 𝑧 = 𝑎𝑥 + 𝑏𝑦 + 𝑎2 + 𝑏 2
b) 𝑧 = 𝑎𝑥𝑦 + 𝑏
c) 𝑧 = 𝑎(𝑥 + 𝑦)2 + 𝑏
d) (𝑥 − 𝑎)2 + (𝑦 − 𝑏)2 + 𝑧 2 = 1
e) 𝑧 = 𝑥𝑦 + 𝑦√𝑥 2 − 𝑎2 + 𝑏
2
f) 𝑧 = 𝑎𝑒 −𝑏 𝑡 𝑐𝑜𝑠𝑏𝑥
g) 𝑎𝑥 2 + 𝑏𝑦 2 + 𝑐𝑧 2 = 1
2. Form PDEs by eliminating arbitrary functions
a) 𝑓(𝑥 + 𝑦 + 𝑧, 𝑥 2 + 𝑦 2 − 𝑧 2 ) = 0
b) 𝑓(𝑥𝑦 + 𝑧 2 , 𝑥 + 𝑦 + 𝑧) = 0
𝑦
c) 𝑧 = 𝑓(𝑥 )
2.5 References
1. Tyn Myin-U and Lokeath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.
2. Frank Ayres (1952). Shaum’s outline of theory and problems of differential equations.
McGRAW-HILL publishers

17
 LESSON THREE

SOLUTION OF FIRST ORDER QUASILINEAR PDEs BY THE LAGRANGE’S

METHOD OF TYPE ONE AND TWO
3.1 Introduction
In this lesson we will study method of solutions of quasilinear PDES. We will specifically
consider the method of Lagrange auxiliary equations of types 1 and 2.

3.2 Learning Outcomes

By the end of this lesson the learner will be able to:

3.2.1 Solve quasilinear PDEs using the Lagrange’s method, rule1;

3.2.2 Solve quasilinear PDEs using the Lagrange’s method, rule 2

3.3 Lagrange’s Equations

An equation of the form

𝜕𝑢 𝜕𝑢
𝑃 +𝑄 = 𝑅 or 𝑃𝑝 + 𝑄𝑞 = 𝑅(∗)
𝜕𝑥 𝜕𝑦

Is called Lagrange’s equation.

Theorem

The general solution of the linear partial differential equation 𝑃𝑝 + 𝑄𝑞 = 𝑅 is given by

𝜙(𝑈, 𝑉) = 0 where ϕ is an arbitrary function and 𝑈(𝑥, 𝑦, 𝑢) = 𝑐1 and 𝑉(𝑥, 𝑦, 𝑢) = 𝑐2 form a
solution of the equations
𝑑𝑥 𝑑𝑦 𝑑𝑢
= = (Lagrange’s auxiliary system, also called subsidiary system of
𝑃 𝑄 𝑅

𝑃𝑝 + 𝑄𝑞 = 𝑅).

Working rule for solving 𝑷𝒑 + 𝑸𝒒 = 𝑹 by Lagrange’s method

Step 1. Put the given linear PDE of first order in the standard form 𝑃𝑝 + 𝑄𝑞 = 𝑅(1)

18
 𝑑𝑥 𝑑𝑦 𝑑𝑢
Step 2.Write down Lagrange’s auxiliary system, namely; = = (2)
𝑃 𝑄 𝑅

Step 3.Solve (2) to get 𝑈(𝑥, 𝑦, 𝑢) = 𝑐1 and 𝑉(𝑥, 𝑦, 𝑢) = 𝑐2 taken as two independent solutions
of (2).

Step 4. The general solution (or integral of (1) is then written in one of the following three
equivalent forms:

𝜙(𝑈, 𝑉) = 0, 𝑈 = 𝜙(𝑉) or 𝑉 = 𝜙(𝑈).

𝑑𝑥 𝑑𝑦 𝑑𝑢
Complete solutions. If 𝑈 = 𝑎 and 𝑉 = 𝑏 are two independent solutions of = = and if
𝑃 𝑄 𝑅
𝛼 and 𝛽 are arbitrary constants, 𝑈 = 𝛼𝑉 + 𝛽 is called the complete solution of 𝑃𝑝 + 𝑄𝑞 = 𝑅.

𝒅𝒙 𝒅𝒚 𝒅𝒖
Type 1. The PDE whose auxiliary system, 𝑷 = = is such that one of the variables is
𝑸 𝑹
either absent or cancels out from any two fractions of the given equations is said to be of type 1.
The general solution can be obtained by grouping two fractions.

Example 3.1

Solve 𝑦𝑢𝑝 + 𝑥𝑢𝑞 = 𝑥𝑦

Solution

Here 𝑃 = 𝑦𝑢, 𝑄 = 𝑥𝑢 and 𝑅 = 𝑥𝑦

𝑑𝑥 𝑑𝑦 𝑑𝑢
The Lagrange’s auxiliary system for the given PDE are = 𝑥𝑢 =
𝑦𝑢 𝑥𝑦

𝑑𝑥 𝑑𝑦
Grouping the first and second fractions, we have = 𝑥𝑢⇒𝑥𝑑𝑥 − 𝑦𝑑𝑦 = 0, now by integrating
𝑦𝑢
𝑥2 𝑦2
term by term we get − = 𝑎 or 𝑥 2 − 𝑦 2 = 𝐶1 . Similarly grouping the first and third
2 2
𝑑𝑥 𝑑𝑢
fractions, we have = ⇒𝑥𝑑𝑥 − 𝑢𝑑𝑢 = 0. Integrating term by term we get
𝑦𝑢 𝑥𝑦

𝑥 2 − 𝑢2 = 𝐶2 . Hence the required general solution is 𝜙(𝑥 2 − 𝑦 2 , 𝑥 2 − 𝑢2 ) = 0

Example 3.2

Solve 𝑦 2 𝑝 − 𝑥𝑦𝑞 = 𝑥(𝑢 − 2𝑦)

Solution

Here, 𝑃 = 𝑦 2 , 𝑄 = −𝑥𝑦, 𝑅 = 𝑥(𝑢 − 2𝑦)

19
 𝑑𝑥 𝑑𝑦 𝑑𝑢
Lagrange’s system are = −𝑥𝑦 = 𝑥(𝑢−2𝑦). Grouping the first two fractions, we have
𝑦2

𝑥𝑑𝑥 + 𝑦𝑑𝑦 = 0 ⇒ 𝑥 2 + 𝑦 2 = 𝐶1 . Now grouping second and third fractions, we have

𝑑𝑦 𝑑𝑢 𝑑𝑢 (𝑢−2𝑦) 𝑑𝑢 𝑢
= 𝑥(𝑢−2𝑦)⇒ =− ⇒ + 𝑦 = 2 which is linear in 𝑢 and 𝑦. Its integrating
−𝑥𝑦 𝑑𝑦 𝑦 𝑑𝑦
𝑑𝑢 𝑑
factor is 𝑒 𝑙𝑛𝑦 = 𝑦. Hence 𝑦 𝑑𝑦 + 𝑢 = 2𝑦 ⇒ 𝑑𝑦 (𝑦𝑢) = 2𝑦 ⇒ 𝑦𝑢 = 𝑦 2 +𝐶2

⇒ 𝑦𝑢 − 𝑦 2 = 𝐶2 . Hence the required general solution is 𝜙(𝑥 2 + 𝑦 2 , 𝑦𝑢 − 𝑦 2 ) = 0.

Type 2. Suppose one integral is known by the method of type 1 and suppose also that
another integral cannot be obtained by using the method of type 1. Then if one integral known to
us is used to find another integral, then the corresponding PDE is said to be of type 2.

Example 3.3
𝜕𝑢 𝜕𝑢
Solve + 𝜕𝑦 = 𝑥 + 𝑦 + 𝑢
𝜕𝑥

Solution

The Lagrange auxiliary equations are

𝑑𝑥 𝑑𝑦 𝑑𝑢
= =
1 1 𝑥+𝑦+𝑢
𝑑𝑥 𝑑𝑦
Grouping the first two fractions: = ⇒ 𝑑𝑥 − 𝑑𝑦 = 0 ⇒ 𝑥 − 𝑦 = 𝐶1 so that
1 1

𝑥 = 𝑦 + 𝐶1 .
𝑑𝑦 𝑑𝑢 𝑑𝑦 𝑑𝑢 𝑑𝑢
Grouping the last two fractions: = 𝑥+𝑦+𝑢 ⇒ = 2𝑦+𝐶 or 𝑑𝑦 = 𝑢 + 2𝑦 + 𝐶1 ⇒
1 1 1 +𝑢

𝑑𝑢
− 𝑢 = 2𝑦 + 𝐶1 which is linear in 𝑢 and 𝑦. Its integrating factor is 𝑒 −𝑦 and hence
𝑑𝑦

𝑑𝑢 𝑑
𝑒 −𝑦 (𝑑𝑦 − 𝑢) = (2𝑦 + 𝐶1 )𝑒 −𝑦 ⇒𝑑𝑦 (𝑢𝑒 −𝑦 ) = (2𝑦 + 𝐶1 )𝑒 −𝑦 ⇒ 𝑢𝑒 −𝑦 = (2𝑦 + 𝐶1 )(−𝑒 −𝑦 ) −
2(𝑒 −𝑦 ) + 𝐶2 by chain rule.

⇒ 𝑒 −𝑦 (𝑢 + 𝑥 + 𝑦 + 2) = 𝐶2 and hence the required general solution is

ϕ(𝑥 + 𝑦, 𝑒 −𝑦 (𝑢 + 𝑥 + 𝑦 + 2)) = 0

Example 3.4

Solve 𝑥𝑢(𝑢2 + 𝑥𝑦)𝑝 − 𝑦𝑢(𝑢2 + 𝑥𝑦)𝑞 = 𝑥 4

20
Solution
𝑑𝑥 𝑑𝑦 𝑑𝑢
The Lagrange auxiliary equations are = −𝑦𝑢(𝑢2 +𝑥𝑦) =
𝑥𝑢(𝑢2 +𝑥𝑦) 𝑥4

𝑑𝑥 𝑑𝑦 𝑑𝑥 𝑑𝑦
Grouping the first 2 fractions we get = −𝑦⇒ 𝑥 + = 0 ⇒ 𝑙𝑛𝑥 + 𝑙𝑛𝑦 = 𝑙𝑛𝐶1
𝑥 𝑦

⇒ 𝑥𝑦 = 𝐶1 (1)

Using (1), the first and third fractions give

𝑑𝑥 𝑑𝑢
2
= 4 ⇒ 𝑥 3 𝑑𝑥 = 𝑢(𝑢2 + 𝐶2 )𝑑𝑢
𝑥𝑢(𝑢 + 𝐶2 ) 𝑥

⇒ 𝑥 4 − 𝑢4 − 2𝐶1 𝑢2 = 𝐶2 ⇒ 𝑥 4 − 𝑢4 − 2𝑥𝑦𝑢2 = 𝐶2 (2)

Hence the required general solution is ϕ(𝑥𝑦, 𝑥 4 − 𝑢4 − 2𝑥𝑦𝑢2 ) = 0

E-tivity 3.2: Solutions of quasilinear PDEs using Lagrange auxiliary equations

Numbering, pacing and sequencing 3.2.1-3.2.2

Title Solutions of quasilinear PDEsusing Lagrange auxiliary

equations- Types 1 and 2.

Purpose To expose you to the methodLagrange auxiliary

equations of types 1 and 2.

Brief summary of overall task Watch video on the methods of solving quasilinear
PDEsand then solve given problems.
Spark
LAGRANGE AUXILIARY
EQUATIONS OF TYPES 1
AND 2.
Individual contribution • Watch videos on the methods of solving
quasilinear PDEs
• Answer the question;
Use the appropriate method from the two
methods to solve the PDE
1. 𝑥𝑢𝑥 + 𝑦𝑞 = 𝑢
2. 𝑢(𝑢2 + 𝑥𝑦)(𝑢𝑝 − 𝑦𝑞) = 𝑥 4
Interaction begins • Post your answers on the discussion forum 3.2.3
• Read what your colleagues have posted.
• In a sentence or two, comment on what two of

21
 your colleagues have posted.

E-moderator interventions • Focussing group discussion

• Encouraging quiet ones to contribute
• Providing feedback/ teaching points
• Summarising key points
• Closing the discussion
Schedule and time This activity should take 1 hour
Next end

3.4 Assessment
Find the general solution of each of the following using the appropriate method

a) 𝑢𝑢𝑥 + 2𝑢𝑢𝑦 = −𝑥
b) (1 + 𝑥 2 )𝑢𝑥 + 𝑢𝑦 = 𝑢
𝑦2𝑢
c) 𝑢𝑥 + 𝑥𝑢𝑢𝑦 = 𝑦 2
𝑥
d) 𝑥𝑦𝑝 + 𝑦 2 𝑞 = 𝑥𝑦𝑢 − 2𝑥 2
e) 𝑝 + 3𝑞 = 5𝑢 + tan (𝑦 − 3𝑥)
f) 𝑝 − 2𝑞 = 3𝑥 2 sin (𝑦 + 2𝑥)
3.5 References
1. Tyn Myin-U and Lokeath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.

2. Frank Ayres (1952). Shaum’s outline of theory and problems of differential equations.

22
 LESSON FOUR

SOLUTION OF FIRST ORDER QUASILINEAR PDEs BY THE LAGRANGE’S

METHOD OF TYPES THREE AND FOUR
4.1 Introduction
In this lesson we will study two other methods of solutions of quasilinear PDES. We will
specifically consider the method of solutions of quasilinear PDEs of types 3 and 4.

4.2 Learning Outcomes

By the end of this lesson the learner will be able to:

4.2.1 Solve quasilinear PDEs using the Lagrange’s method, type 3;

4.2.2 Solve quasilinear PDEs using the Lagrange’s method, type 4

4.3 Quasilinear PDEs of Type 3 (Existence of multipliers 𝑷𝟏 , 𝑸𝟏 and 𝑹𝟏 such that

𝑷𝟏 𝑷 + 𝑸𝟏 𝑸 + 𝑹𝟏 𝑹 = 𝟎)

The class of quasilinear PDEs 𝑃𝑝 + 𝑄𝑞 = 𝑅, such that there exist 𝑃1 , 𝑄1 and 𝑅1 so that

𝑃1 𝑃 + 𝑄1 𝑄 + 𝑅1 𝑅 = 0 , is said to be of type 3.

Method of solution.
𝑑𝑥 𝑑𝑦 𝑑𝑢
Given the PDE 𝑃𝑝 + 𝑄𝑞 = 𝑅, the auxiliary system is 𝑃
= 𝑄
= 𝑅
and if

𝑃1 𝑃 + 𝑄1 𝑄 + 𝑅1 𝑅 = 0 then 𝑃1 𝑑𝑥 + 𝑄1 𝑑𝑦 + 𝑅1 𝑑𝑢 = 0 which can be integrated to give

𝑈1 (𝑥, 𝑦, 𝑢) = 𝐶1 . This method may be repeated to get another integral 𝑈2 (𝑥, 𝑦, 𝑢) = 𝐶2 .

𝑃1 , 𝑄1 and 𝑅1 are called multipliers. As a special case, these can be constants also.

Sometimes only one integral is possible by use of multipliers. In such cases second integral

should be obtained by using type 1 or type 2 methods as the case may be.

Example 4.1

Solve (𝑢2 − 2𝑦𝑢 − 𝑦 2 )𝑝 + (𝑥𝑦 + 𝑥𝑢)𝑞 = 𝑥𝑦 − 𝑥𝑢

23
Solution

Lagrange auxiliary equations for this PDE are

𝑑𝑥 𝑑𝑦 𝑑𝑢 𝑥𝑑𝑥 𝑑𝑦 𝑑𝑢
= 𝑥(𝑦+𝑢) = 𝑥(𝑦−𝑢) or 𝑢2 −2𝑦𝑢−𝑦 2 = (𝑦+𝑢) = (𝑦−𝑢)
𝑢2 −2𝑦𝑢−𝑦 2

Choose the multipliers as 1, 𝑦, 𝑢 so that 𝑢2 − 2𝑦𝑢 − 𝑦 2 + 𝑦(𝑦 + 𝑢) + 𝑢(𝑦 − 𝑢) = 0.

Therefore 𝑥𝑑𝑥 + 𝑦𝑑𝑦 + 𝑢𝑑𝑢 = 0. Integrating term by term we get 𝑥 2 + 𝑦 2 + 𝑢2 = 𝐶1 .

𝑑𝑦 𝑑𝑢
Grouping the last two fractions, we have (𝑦+𝑢) = (𝑦−𝑢) ⇒ (𝑦 − 𝑢)𝑑𝑦 − (𝑦 + 𝑢)𝑑𝑢 = 0

⇒ 𝑦𝑑𝑦 − 𝑢𝑑𝑢 − (𝑢𝑑𝑦 + 𝑦𝑑𝑢) = 0 ⇒ 𝑦𝑑𝑦 − 𝑢𝑑𝑢 − 𝑑(𝑢𝑦) = 0. Now integrating term by

term we get 𝑦 2 − 𝑢2 − 2𝑢𝑦 = 𝐶2 . The required general solution is

𝑥 2 + 𝑦 2 + 𝑢2 = 𝜙(𝑦 2 − 𝑢2 − 2𝑢𝑦).

Example 4.2

Solve 𝑢(𝑥𝑝 − 𝑦𝑞) = 𝑦 2 − 𝑥 2

Solution

Lagrange auxiliary equations are

𝑑𝑥 𝑑𝑦 𝑑𝑢
= −𝑦𝑢 = 𝑦 2−𝑥 2 . Choosing 𝑥, 𝑦, 𝑢 as multipliers, we have
𝑥𝑢

𝑥𝑑𝑥+𝑦𝑑𝑦+𝑢𝑑𝑢 𝑥𝑑𝑥+𝑦𝑑𝑦+𝑢𝑑𝑢
= ⇒ 𝑥𝑑𝑥 + 𝑦𝑑𝑦 + 𝑢𝑑𝑢 = 0. Now integrating term by
𝑢𝑥 2 −𝑢𝑦 2 +𝑢𝑦 2 −𝑢𝑥 2 0
term we get 𝑥 + 𝑦 2 + 𝑢 = 𝐶1.
2 2

Grouping the first and the second fractions, we have

𝑑𝑥 𝑑𝑦 𝑑𝑥 𝑑𝑦
= −𝑦𝑢 ⇒ + = 0 and now by integrating term by term, we get
𝑥𝑢 𝑥 𝑦

𝑙𝑛𝑥 + 𝑙𝑛𝑦 = 𝑙𝑛𝐶2 ⇒ 𝑥𝑦 = 𝐶2 . Hence the required general solution is

𝑥 2 + 𝑦 2 + 𝑢2 = 𝜙(𝑥𝑦).

4.2.1 Quasilinear PDEs of type 4

Consider the Lagrange auxiliary system

24
𝑑𝑥 𝑑𝑦 𝑑𝑢
= = . Let 𝑃1 , 𝑄1 and 𝑅1 be functions of 𝑥, 𝑦, 𝑧. Then by a well-known principle of
𝑃 𝑄 𝑅
𝑑𝑥 𝑑𝑦 𝑑𝑢 𝑃1 𝑑𝑥+𝑄1 𝑑𝑦+𝑅1 𝑑𝑢
algebra, = = =
𝑃 𝑄 𝑅 𝑃1 𝑃+𝑄1 𝑄+𝑅1 𝑅

𝑃1 𝑑𝑥+𝑄1 𝑑𝑦+𝑅1 𝑑𝑢
Suppose the numerator of (∗)
𝑃1 𝑃+𝑄1 𝑄+𝑅1 𝑅

is an exact differential of the denominator of (∗). Then (∗) can be combined with a
𝑑𝑥 𝑑𝑦 𝑑𝑢
suitable fraction of = = to give an integral. However, in some problems, another
𝑃 𝑄 𝑅
set of multipliers 𝑃2 , 𝑄2 and 𝑅2 are so chosen that the fraction
𝑃2 𝑑𝑥+𝑄2 𝑑𝑦+𝑅2 𝑑𝑢
(∗∗)
𝑃2 𝑃+𝑄2 𝑄+𝑅2 𝑅

is such that its numerator is exact differential of denominator. Fractions (∗) and (∗∗) are
then combined to give an integral. This method may also be repeated in some problems to
get another integral.

A quasilinear PDE that admits the above is said to be of type 4.

Example 4.3

Find the general solution to the PDE 𝑥𝑢𝑝 + 𝑦𝑢𝑞 = 𝑥𝑦

Solution
𝑑𝑥 𝑑𝑦 𝑑𝑢
Lagrange auxiliary equations are = 𝑦𝑢 =
𝑥𝑢 𝑥𝑦

𝑑𝑥 𝑑𝑦 𝑥
Taking the first two fractions and cancelling 𝑢, we get − = 0 so that = 𝐶1 .
𝑥 𝑦 𝑦

1 1 𝑑𝑥 𝑑𝑦 𝑑𝑢 𝑑𝑥⁄ + 𝑑𝑦⁄ 𝑦𝑑𝑥+𝑥𝑑𝑦

𝑥 𝑦
Taking 𝑥 , 𝑦 , 0 as multipliers, = 𝑦𝑢 = = =
𝑥𝑢 𝑥𝑦 𝑢+𝑢 2𝑥𝑦𝑢

𝑦𝑑𝑥+𝑥𝑑𝑦 𝑑𝑢
⇒ = 𝑥𝑦 ⇒ 𝑑(𝑥𝑦) − 2𝑢𝑑𝑢 = 0 ⇒ 𝑥𝑦 − 𝑢2 = 𝐶2. Hence the required general
2𝑥𝑦𝑢
𝑥
solution is 𝜙 (𝑥𝑦 − 𝑢2 , 𝑦) = 0.

Example 4.4

Solve (𝑥 2 − 𝑦 2 − 𝑢2 )𝑝 + 2𝑥𝑦𝑞 − 2𝑥𝑢 = 0

Solution

First put the PDE in Lagrange’s form i.e. (𝑥 2 − 𝑦 2 − 𝑢2 )𝑝 + 2𝑥𝑦𝑞 = 2𝑥𝑢

25
 𝑑𝑥 𝑑𝑦 𝑑𝑢
Lagrange auxiliary system is = 2𝑥𝑦 = 2𝑥𝑢
𝑥 2 −𝑦 2 −𝑢2

𝑑𝑦 𝑑𝑢 𝑑𝑦 𝑑𝑢 𝑦
Taking the last two fractions = 2𝑥𝑢 ⇒ = ⇒ 𝑙𝑛𝑦 − 𝑙𝑛𝑢 = 𝑙𝑛𝐶1 ⇒ = 𝐶1
2𝑥𝑦 𝑦 𝑢 𝑢

𝑑𝑥 𝑑𝑦 𝑑𝑢 𝑥𝑑𝑥+𝑦𝑑𝑦+𝑢𝑑𝑢
Choosing 𝑥, 𝑦, 𝑢 as multipliers = 2𝑥𝑦 = 2𝑥𝑢 = 𝑥 3−𝑥𝑦 2−𝑥𝑢2 +2𝑥𝑦 2 +2𝑥𝑢2
𝑥 2 −𝑦 2 −𝑢2

𝑥𝑑𝑥+𝑦𝑑𝑦+𝑢𝑑𝑢
= .
𝑥(𝑥 2 +𝑦 2 +𝑢2 )

𝑑𝑢 𝑥𝑑𝑥+𝑦𝑑𝑦+𝑢𝑑𝑢 𝑑𝑢 2𝑥𝑑𝑥+2𝑦𝑑𝑦+2𝑢𝑑𝑢
We can take the combination 2𝑥𝑢 = ⇒ = and on
𝑥(𝑥 2 +𝑦 2 +𝑢2 ) 𝑢 (𝑥 2 +𝑦 2 +𝑢2 )
𝑥 2 +𝑦 2 +𝑢2
integrating we get 𝑙𝑛(𝑥 2 + 𝑦 2 + 𝑢2 ) = 𝑙𝑛𝑢 + 𝑙𝑛𝐶2 ⇒ = 𝐶2
𝑢

𝑥 2 +𝑦 2 +𝑢2 𝑦
The required general solution is 𝜙 ( , 𝑢) = 0.
𝑢

E-tivity 4.2: The method of multipliers ( PDEs of types 3 and 4)

Numbering, pacing and sequencing 4.2.1 -4.2.2

Title The method of multipliers

Purpose To expose you to the method method of multipliers of

solving quasilinear PDEs.

Brief summary of overall task Watch video on Multipliersand then solve given
problems.
Spark
PDEs of Types 3 and 4

Individual contribution • Watch videos on the methods of solving

quasilinear PDEs
• Answer the question;
Use the methodof multipliers to solve the PDE
(𝑥 + 1)𝑢𝑦 + (1 + 𝑦)𝑢𝑥 = 𝑧

Interaction begins • Post your answers on the discussion forum 4.2.3

• Read what your colleagues have posted.
• In a sentence or two, comment on what two of
your colleagues have posted.

26
E-moderator interventions • Focussing group discussion
• Encouraging quiet ones to contribute
• Providing feedback/ teaching points
• Summarising key points
• Closing the discussion
Schedule and time This activity should take 1 hour
Next Integral surfaces passing through a curve and surfaces
orthogonal to a system of surfaces

4.5 Assessment
Solve
a) 𝑢(𝑥 + 𝑦)𝑝 + 𝑢(𝑥 − 𝑦)𝑞 = 𝑥 2 + 𝑦 2
b) 𝑥(𝑦 2 − 𝑢2 )𝑝 + 𝑦(𝑢2 − 𝑥 2 )𝑞 = 𝑢(𝑥 2 − 𝑦 2 )
c) 𝑥(𝑦 − 𝑢) + 𝑦(𝑢 − 𝑥)𝑞 = 𝑢(𝑥 − 𝑦)
d) 𝑦𝑝 − 𝑥𝑞 = 2𝑥 − 3𝑦
e) 𝑦 2 (𝑥 − 𝑦)𝑝 + 𝑥 2 (𝑦 − 𝑥)𝑞 = 𝑢(𝑥 2 + 𝑦 2 )
f) (𝑦 + 𝑢)𝑝 + (𝑢 + 𝑥)𝑞 = 𝑥 + 𝑦
g) (1 + 𝑦)𝑝 + (1 + 𝑥)𝑞 = 𝑢

4.6 References
1. Tyn Myin-U and Lokeath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.
2. Frank Ayres (1952). Shaum’s outline of theory and problems of differential equations.
McGRAW-HILL publishers

27
 LESSON FIVE

INTEGRAL SURFACES PASSING THROUGH GIVEN CURVES AND SURFACES

ORTHOGONAL TO A GIVEN SYSTEM OF SURFACES

5.1 Introduction
In this lesson we will study integral surfaces passing through given curves and those orthogonal
to a given system of surfaces.

5.2 Learning Outcomes

By the end of this lesson the learner will be able to:

5.2.1 Solve quasilinear PDEs subject to some constraints;

5.2.2 Obtain the equation of surfaces which intersect the surfaces of a given system
orthogonally.

5.3 Integral surfaces passing through a given curve

We discuss two methods of using general solutions to get integral surfaces that pass
through given curves.

Method I

Let 𝑃𝑝 + 𝑄𝑞 = 𝑅 be the given equation. Let its auxiliary equations give the following two
independent solutions 𝑈(𝑥, 𝑦, 𝑢) = 𝐶1 and 𝑉(𝑥, 𝑦, 𝑢) = 𝐶2 (1)

Suppose we wish to obtain the integral surface which passes through the curve whose
equation in parametric form is given by 𝑥 = 𝑥(𝑡), 𝑦 = 𝑦(𝑡), 𝑢 = 𝑢(𝑡) where 𝑡 is a
parameter. Then (1) may be expressed as

𝑈(𝑥(𝑡), 𝑦(𝑡), 𝑢(𝑡)) = 𝐶1 and 𝑉(𝑥(𝑡), 𝑦(𝑡), 𝑢(𝑡)) = 𝐶2 (2)

We eliminate single parameter 𝑡 from the equations (2) and get a relation involving 𝐶1 and
𝐶2 . Finally, we replace 𝐶1 and 𝐶2 with the help of (1) and obtain the required integral
surface.

28
Example 5.1

Find the integral surface of the linear partial differential equation

𝑥(𝑦 2 + 𝑢)𝑝 − 𝑦(𝑥 2 + 𝑢)𝑞 = (𝑥 2 − 𝑦 2 )𝑢 which contains the straight line 𝑥 + 𝑦 = 0, 𝑢 = 1.

Solution

Lagrange auxiliary equations are

𝑑𝑥 𝑑𝑦 𝑑𝑢
= =
𝑥(𝑦 2 + 𝑢) −𝑦(𝑥 2 + 𝑢) (𝑥 2 − 𝑦 2 )𝑢
𝑥𝑑𝑥+𝑦𝑑𝑦−𝑑𝑢 𝑥𝑑𝑥+𝑦𝑑𝑦−𝑑𝑢
Choosing the multipliers as 𝑥, 𝑦, −1 we have 𝑥 2 𝑦 2+𝑥 2𝑢−𝑦 2𝑥 2 −𝑦 2𝑢−𝑥 2𝑢+𝑦 2𝑢 = 0

⇒ 𝑥𝑑𝑥 + 𝑦𝑑𝑦 − 𝑑𝑢 = 0 ⇒ 𝑥 2 + 𝑦 2 − 2𝑢 = 𝐶1 (1)

1 1 1 1 1 1
1 1 1 𝑑𝑥+ 𝑑𝑦+ 𝑑𝑢 𝑑𝑥+ 𝑑𝑦+ 𝑑𝑢
𝑥 𝑦 𝑢 𝑥 𝑦 𝑢
Next choose multipliers 𝑥 , 𝑦 , 𝑧 then we have =
𝑦 2 +𝑢−𝑥 2 −𝑢+𝑥 2 −𝑦 2 0

1 1 1
⇒ 𝑑𝑥 + 𝑦 𝑑𝑦 + 𝑢 𝑑𝑢 = 0 ⇒ 𝑙𝑛𝑥 + 𝑙𝑛𝑦 + 𝑙𝑛𝑢 = 𝑙𝑛𝐶2 ⇒ 𝑥𝑦𝑢 = 𝐶2 (2)
𝑥

Now, taking 𝑡 as a parameter, the given equation of the straight line 𝑥 + 𝑦 = 0, 𝑢 = 1

Can be put in parametric form 𝑥 = 𝑡, 𝑦 = −𝑡, 𝑢 = 1 (3)

Using (3), (2) may be re-written as −𝑡 2 = 𝐶2 and (1) as 2𝑡 2 − 2 = 𝐶1 which give

2(−𝐶2 ) − 2 = 𝐶1 ⇒ 2𝐶2 + 𝐶1 + 2 = 0 (4)

Putting the values of 𝐶1 and 𝐶2 from (1) and (2) in (4), the required integral surface is

2𝑥𝑦𝑢 + 𝑥 2 + 𝑦 2 − 2𝑢 + 2 = 0

Method II

Let 𝑃𝑝 + 𝑄𝑞 = 𝑅 be the given equation. Let the Lagrange’s auxiliary equations give the
following two independent integrals 𝑈(𝑥, 𝑦, 𝑢) = 𝐶1 and 𝑉(𝑥, 𝑦, 𝑢) = 𝐶2 (1)

Suppose we wish to obtain the integral surface which passing through the curve which is
determined by the following two equations 𝜗(𝑥, 𝑦, 𝑢) = 0 and 𝜑(𝑥, 𝑦, 𝑢) = 0 (2)

29
We eliminate 𝑥, 𝑦, 𝑢 from four equations of (1) and (2) and obtain a relation between 𝐶1

and 𝐶2 . Finally replace 𝐶1 by 𝑈(𝑥, 𝑦, 𝑢) and 𝐶2 by 𝑉(𝑥, 𝑦, 𝑢) in that relation and obtain the
required integral surface.

Example 5.3

Find the integral surface of the partial differential equation (𝑥 − 𝑦)𝑝 + (𝑦 − 𝑥 − 𝑢)𝑞 = 𝑢

through the circle 𝑢 = 1, 𝑥 2 + 𝑦 2 = 1.

Solution

Lagrange’s auxiliary equations are

𝑑𝑥 𝑑𝑦 𝑑𝑢
= =
𝑥−𝑦 𝑦−𝑥−𝑢 𝑢
𝑑𝑥 𝑑𝑦 𝑑𝑢 𝑑𝑥+𝑑𝑦+𝑑𝑢
Choosing 1, 1, 1 as the multipliers, we have = 𝑦−𝑥−𝑢 = = ⇒
𝑥−𝑦 𝑢 0

𝑑𝑥 + 𝑑𝑦 + 𝑑𝑢 = 0 so that 𝑥 + 𝑦 + 𝑢 = 𝐶1 (1)
𝑑𝑦 𝑑𝑢 2𝑑𝑦 2𝑑𝑢
Taking the last two fractions and using (1), we get 𝑦−(𝐶 = or 2𝑦−𝐶 − = 0.
1 −𝑦) 𝑢 1 𝑢

Integrating it, we have 𝑙𝑛(2𝑦 − 𝐶1 ) − 2𝑙𝑛𝑢 = 𝑙𝑛𝐶2 or (2𝑦 − 𝐶1 )𝑢2 = 𝐶2 or

(2𝑦−𝑥−𝑦−𝑢)
= 𝐶2 or (𝑦 − 𝑥 − 𝑢) = 𝐶2. (2)
𝑢2

The given curve is 𝑢 = 1, 𝑥 2 + 𝑦 2 = 1

Putting 𝑢 = 1 in (1) and (2), we get 𝑥 + 𝑦 = 𝐶1 − 1 and 𝑦 − 𝑥 = 𝐶2 + 1

But 2(𝑥 2 + 𝑦 2 ) = (𝑥 + 𝑦)2 + (𝑦 − 𝑥)2 ⇒ 2(1) = (𝐶1 − 1)2 + (𝐶2 + 1)2 or

(𝑥 + 𝑦 + 𝑢 − 1)2 + (𝑦 − 𝑥 − 𝑢 + 1)2 = 2 is the required integral surface.

Example 5.4

Find the equation of the integral surface of the differential

(𝑥 2 − 𝑦𝑢)𝑝 + (𝑦 2 − 𝑥𝑢)𝑞 = 𝑢2 − 𝑥𝑦 which passes through the line 𝑥 = 1, 𝑦 = 0

Solution

Lagrange’s auxiliary equations are

30
 𝑑𝑥 𝑑𝑦 𝑑𝑢
= 2 = 2
𝑥2 − 𝑦𝑢 𝑦 − 𝑥𝑢 𝑢 − 𝑥𝑦

Choosing 1, 1, 0 and 0, 1, −1 as multipliers, so that

𝑑𝑥 𝑑𝑦 𝑑𝑢 𝑑𝑥−𝑑𝑦 𝑑𝑦−𝑑𝑢
= 𝑦 2−𝑥𝑢 = 𝑢2 −𝑥𝑦=𝑥 2 −𝑦 2+𝑢(𝑦−𝑢) = 𝑦 2−𝑢2 +𝑥(𝑦−𝑢)
𝑥 2 −𝑦𝑢

Grouping the last two fractions we get

𝑑𝑥−𝑑𝑦 𝑑𝑦−𝑑𝑢 𝑑𝑥−𝑑𝑦 𝑑𝑦−𝑑𝑢 𝑑𝑥−𝑑𝑦 𝑑𝑦−𝑑𝑢
= 𝑦 2−𝑢2+𝑥(𝑦−𝑢) ⇒ = (𝑦−𝑢)(𝑥+𝑦+𝑢) ⇒ =
𝑥 2 −𝑦 2 +𝑢(𝑦−𝑢) (𝑥−𝑦)(𝑥+𝑦+𝑢) 𝑥−𝑦 𝑦−𝑢

𝑥−𝑦
On integrating, we have ln(𝑥 − 𝑦) − ln(𝑦 − 𝑢) = 𝑙𝑛𝐶1 or = 𝐶1 (1)
𝑦−𝑢

𝑑𝑥 𝑑𝑦 𝑑𝑢 𝑥𝑑𝑥+𝑦𝑑𝑦+𝑢𝑑𝑢
Choosing 𝑥, 𝑦, 𝑧 as multipliers , = 𝑦 2−𝑥𝑢 = 𝑢2 −𝑥𝑦 = 𝑥 3 −𝑥𝑦𝑢+𝑦 3−𝑥𝑦𝑢+𝑢3−𝑥𝑦𝑢
𝑥 2 −𝑦𝑢

𝑥𝑑𝑥+𝑦𝑑𝑦+𝑢𝑑𝑢 𝑥𝑑𝑥+𝑦𝑑𝑦+𝑢𝑑𝑢
= 𝑥 3 +𝑦 3+𝑢3−3𝑥𝑦𝑢 = (𝑥+𝑦+𝑢)(𝑥 2+𝑦 2+𝑢2−𝑥𝑦−𝑦𝑢−𝑢𝑥) (2)

Again choosing 1, 1, 1 as multipliers,

𝑑𝑥 𝑑𝑦 𝑑𝑢 𝑑𝑥+𝑑𝑦+𝑑𝑢
= 𝑦 2−𝑥𝑢 = 𝑢2 −𝑥𝑦=(𝑥 2+𝑦 2+𝑢2 −𝑥𝑦−𝑦𝑢−𝑢𝑥) (3)
𝑥 2 −𝑦𝑢

𝑥𝑑𝑥+𝑦𝑑𝑦+𝑢𝑑𝑢 𝑑𝑥+𝑑𝑦+𝑑𝑢
From (2) and (3), (𝑥+𝑦+𝑢)(𝑥2 +𝑦 2 +𝑢2 −𝑥𝑦−𝑦𝑢−𝑢𝑥)
= 𝑥 2 +𝑦 2+𝑢2−𝑥𝑦−𝑦𝑢−𝑢𝑥

𝑥𝑑𝑥+𝑦𝑑𝑦+𝑢𝑑𝑢
⇒ (𝑥+𝑦+𝑢)
= 𝑑𝑥 + 𝑑𝑦 + 𝑑𝑢 or (𝑥 + 𝑦 + 𝑢)(𝑑𝑥 + 𝑑𝑦 + 𝑑𝑢) − 𝑥𝑑𝑥 − 𝑦𝑑𝑦 − 𝑢𝑑𝑢 = 0

1 1
Integrating, we have 2 (𝑥 + 𝑦 + 𝑢)2 − 2 (𝑥 2 + 𝑦 2 + 𝑢2 ) = 𝑥𝑦 + 𝑦𝑢 + 𝑢𝑥 = 𝐶2 (4)

1
The given curve is represented by 𝑥 = 1, 𝑦 = 0 and therefore (1) becomes − 𝑢 = 𝐶1 and
𝑥−𝑦
(4) becomes 𝑢 = 𝐶2 so that 𝐶1 𝐶2 = −1 or 𝐶1 𝐶2 + 1 = 0 or (𝑥𝑦 + 𝑦𝑢 + 𝑢𝑥) + 1 = 0
𝑦−𝑢

Or (𝑥 − 𝑦)(𝑥𝑦 + 𝑦𝑢 + 𝑢𝑥) + 𝑦 − 𝑢 = 0 is the required integral surface.

5.4 Surfaces orthogonal to a given system of surfaces

Let 𝑓(𝑥, 𝑦, 𝑢) = 𝐶 (∗)

31
represent a system of surfaces where 𝐶 is a parameter. Suppose we wish to obtain a system
of surfaces which cut each of (∗) at right angles. Then the direction ratios of the normal at
𝜕𝑓 𝜕𝑓 𝜕𝑓
the point (𝑥, 𝑦, 𝑢) to (∗) which passes through that point are (𝜕𝑥 , , ).
𝜕𝑦 𝜕𝑢

Let the surface 𝑢 = 𝜙(𝑥, 𝑦) (∗∗)

Cuts each surface of (∗) at right angles. Then the normal at (𝑥, 𝑦, 𝑢) to (∗∗) has direction
𝜕𝑢 𝜕𝑢
ratios (𝜕𝑥 , , −1) i.e. (𝑝, 𝑞, −1). Since normal at (𝑥, 𝑦, 𝑢) to (∗) and (∗∗) are at right
𝜕𝑦
angles, we have
𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
𝑝 𝜕𝑥 + 𝑞 𝜕𝑦 − 𝜕𝑢 = 0 or 𝑝 𝜕𝑥 + 𝑞 𝜕𝑦 = 𝜕𝑢 (∗∗∗)

which is of the form 𝑃𝑝 + 𝑄𝑞 = 𝑅.

Conversely, we easily verify that any solution of (∗∗∗) is orthogonal to every surface of (∗).

Example 5.5

Find the surface which intersects the surfaces of the system 𝑢(𝑥 + 𝑦) = 𝑐(3𝑢 + 1)
orthogonally and which passes through the circle 𝑥 2 + 𝑦 2 = 1, 𝑢 = 1.

Solution
𝑢(𝑥+𝑦)
The given system of surfaces can be written as 𝑓(𝑥, 𝑦, 𝑢) ≡ =𝐶 (1)
3𝑢+1

𝜕𝑓 𝑢 𝜕𝑓 𝑢
Therefore = 3𝑢+1 , = 3𝑢+1.
𝜕𝑥 𝜕𝑦

𝜕𝑓 1(3𝑢+1−3𝑢 𝑥+𝑦
= (𝑥 + 𝑦) (3𝑢+1)2
= (3𝑢+1)2
𝜕𝑢

𝜕𝑓 𝜕𝑓 𝜕𝑓 𝑢 𝑢 𝑥+𝑦
The required orthogonal surface is solution of 𝑝 𝜕𝑥 + 𝑞 𝜕𝑦 = 𝜕𝑢 i.e. 𝑝 + 3𝑢+1 𝑞 = (3𝑢+1)2
3𝑢+1

or 𝑢(3𝑢 + 1)𝑝 + 𝑢(3𝑢 + 1)𝑞 = 𝑥 + 𝑦 (2)

𝑑𝑥 𝑑𝑦 𝑑𝑢
Lagrange’s auxiliary equations for (2) are = 𝑢(3𝑢+1) = 𝑥+𝑦 (3)
𝑢(3𝑢+1)

Taking the first two fractions of (3), we get 𝑑𝑥 − 𝑑𝑦 = 0 so that 𝑥 − 𝑦 = 𝐶1 (4)

𝑥𝑑𝑥+𝑦𝑑𝑦−𝑢(3𝑢+1)𝑑𝑢
Choosing 𝑥, 𝑦, −𝑢(3𝑢 + 1) as multipliers, each fraction of (3)= 0

Therefore, 𝑥𝑑𝑥 + 𝑦𝑑𝑦 − 3𝑧 2 𝑑𝑧 − 𝑧𝑑𝑧 = 0

Integrating, we get 𝑥 2 + 𝑦 2 − 2𝑧 3 − 𝑧 2 = 𝐶2 (5)

32
Hence any surface which is orthogonal to (1) has an equation of the form

𝑥 2 + 𝑦 2 − 2𝑧 3 − 𝑧 2 = 𝜙(𝑥 − 𝑦) (6)

In order to get the required surface passing through the circle 𝑥 2 + 𝑦 2 = 1, 𝑢 = 1 we must
choose 𝜙(𝑥 − 𝑦) = −2. Thus the required particular surface is 𝑥 2 + 𝑦 2 − 2𝑧 3 − 𝑧 2 = −2.

E-tivity 5.2: Integral surfaces passing through given curves and surfaces orthogonal to a
given system of surfaces.

Numbering, pacing and sequencing 5.2.1- 5.2.2

Title Surfaces passing through given curves and surfaces

orthogonal to a given system of surfaces.

Purpose To expose you to the methods of obtaining surfaces that

pass through given curves and also determining
orthogonal surfaces to a given system of surfaces.

Brief summary of overall task Watch video on surfaceand then solve given problems.

Spark
Surfaces intersected by
curves and those orthogonal
to them.
Individual contribution • Watch videos on the methods of solving
quasilinear PDEs
• Answer the question;
Use the two methods to solve the PDE
𝑥𝑢𝑥 + 𝑦𝑢𝑦 = 𝑢 ; 𝑥 + 𝑦 = 1 on 𝑦𝑢 = 1

Interaction begins • Post your answers on the discussion forum 5.2.3

• Read what your colleagues have posted.
• In a sentence or two, comment on what two of
your colleagues have posted.

33
E-moderator interventions • Focussing group discussion
• Encouraging quiet ones to contribute
• Providing feedback/ teaching points
• Summarising key points
• Closing the discussion
Schedule and time This activity should take 1 hour
Next Method of canonical forms
5.5 Assessment
1. Find the particular integrals of the following PDEs to represent surfaces passing through
the given curves
a) 𝑝 + 𝑞 = 1; 𝑥 = 0, 𝑦 2 = 𝑢
b) 𝑥𝑝 + 𝑦𝑞 = 𝑢; 𝑥 + 𝑦 = 1, 𝑦𝑢 = 1
c) (𝑦 − 𝑢)𝑝 + (𝑢 − 𝑥)𝑞 = 𝑥 − 𝑦; 𝑢 = 0, 𝑦 = 2𝑥
d) 𝑥(𝑦 − 𝑢)𝑝 + 𝑦(𝑢 − 𝑥)𝑞 = 𝑢(𝑥 − 𝑦); 𝑥 = 𝑦 = 𝑢
e) 𝑦𝑝 − 2𝑥𝑦𝑞 = 2𝑥𝑢; 𝑥 = 𝑡, 𝑦 = 𝑡 2 , 𝑢 = 𝑡 3
2. Find the equation of the system of surfaces which cut orthogonally the cones of the
system 𝑥 2 + 𝑦 2 + 𝑢2 = 𝑐𝑥𝑦
3. Find the surface which is orthogonal to the one parameter system 𝑢 = 𝑐𝑥𝑦(𝑥 2 + 𝑦 2 ) and
which passes through the hyperbola 𝑥 2 − 𝑦 2 = 𝑎2 , 𝑢 = 0.
5.6 References
1. Tyn Myin-U and Lokeath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.
2. Frank Ayres (1952). Shaum’s outline of theory and problems of differential equations.
McGRAW-HILL publishers

34
 LESSON SIX

THE METHOD OF CANONICAL FORMS

6.1 Introduction.
In this lesson we will study the method of canonical form as a means of solving quasilinear
PDEs.
6.2 Learning Outcomes
By the end of this lesson the learner will be able to:

6.2.1 Find a characteristic to a given PDE;

6.2.2 Solve a quasilinear PDE using the method of canonical forms.

6.3 The method of canonical forms

It is often convenient to transform the more general first order linear PDE
𝑎(𝑥, 𝑦)𝑢𝑥 + 𝑏(𝑥, 𝑦)𝑢𝑦 + 𝑐(𝑥, 𝑦)𝑢 = 𝑑(𝑥, 𝑦) (1)
Into a canonical (or standard) form which can be easily integrated to find the general solution of
(1). We use the characteristics of this equation, which are found from the differential equation
𝑑𝑦 𝑏(𝑥,𝑦)
for characteristics 𝑑𝑥 = 𝑎(𝑥,𝑦), to introduce the new transformations by equations
𝜉 = 𝜉(𝑥, 𝑦), 𝜂 = 𝜂(𝑥, 𝑦) (2)
Where ξ and η are once continuously differentiable and their Jacobian
𝐽(𝑥, 𝑦) = 𝜉𝑥 𝜂𝑦 − 𝜉𝑦 𝜂𝑥 (3)
Is nonzero in a domain of interest so that 𝑥 and 𝑦 can be determined uniquely from the system of
equations (3). Thus by chain rule,
𝑢𝑥 = 𝑢𝜉 𝜉𝑥 + 𝑢𝜂 𝜂𝑥 , 𝑢𝑦 = 𝑢𝜉 𝜉𝑦 + 𝑢𝜂 𝜂𝑦
We substitute these partial derivatives (3) into (1) to obtain the equation
𝐴𝑢𝜉 + 𝐵𝑢𝜂 + 𝐶𝑢 = 𝑑 (4)
Where 𝐴 = 𝑎𝜉𝑥 + 𝑏𝜉𝑦 , 𝐵 = 𝑎𝜂𝑥 + 𝑏𝜂𝑦 (5)
From (5) we see that 𝐵 = 0 if η is a solution of the first order equation

35
𝑎𝜂𝑥 + 𝑏𝜂𝑦 = 0 (6)
In practice, it is convenient to choose 𝜉 = 𝜉(𝑥, 𝑦) and 𝜂(𝑥, 𝑦) = 𝑦 or 𝜉 = 𝑥 and 𝜂 = 𝜂(𝑥, 𝑦) so
that 𝐽 ≠ 0.

Example 6.1

Reduce the equation 𝑢𝑥 − 𝑢𝑦 = 𝑢 to canonical form and obtain the general solution.

Solution
𝑑𝑦
The differential equation for characteristic is 𝑑𝑥 = −1⇒ The characteristic curve is ξ(𝑥, 𝑦) =
𝑥 + 𝑦 = 𝐶1 . Choose η= 𝑦 = 𝐶2, Where 𝐶1 and 𝐶2 are constants.

𝑢𝑥 = 𝑢𝜉 𝜉𝑥 + 𝑢𝜂 𝜂𝑥 = 𝑢𝜉 and 𝑢𝑦 = 𝑢𝜉 𝜉𝑦 + 𝑢𝜂 𝜂𝑦 = 𝑢𝜉 + 𝑢𝜂

Therefore, 𝑢𝑥 − 𝑢𝑦 = 𝑢𝜉 − 𝑢𝜉 − 𝑢𝜂 ⇒ 𝑢𝑥 − 𝑢𝑦 = − 𝑢𝜂 = 𝑢, i.e the canonical form is

𝑢𝜂 = −𝑢, which is a first order ode in 𝜂. Solving this we get the general solution as

𝑢(𝑥, 𝑦) = 𝐹(𝑥 + 𝑦)𝑒 −𝑦 , where 𝐹 is an arbitrary function.

Example 6.2

Reduce the equation 𝑦𝑢𝑥 + 𝑢𝑦 = 𝑥 to canonical form and obtain the general solution.

Solution
𝑑𝑦 1
The differential equation for characteristic is 𝑑𝑥 = 𝑦⇒ The characteristic curve is

𝑦2
ξ(𝑥, 𝑦) = − 𝑥 = 𝐶1 . Choose η= 𝑦 = 𝐶2 , where 𝐶1 and 𝐶2 are constants.
2

𝑦2 η2 η2
Note that 𝑥 = −𝜉 = − 𝜉 i.e 𝑥 = −𝜉
2 2 2

𝑢𝑥 = 𝑢𝜉 𝜉𝑥 + 𝑢𝜂 𝜂𝑥 = −𝑢𝜉 and 𝑢𝑦 = 𝑢𝜉 𝜉𝑦 + 𝑢𝜂 𝜂𝑦 = 𝑦𝑢𝜉 + 𝑢𝜂

η2
Therefore, 𝑦𝑢𝑥 + 𝑢𝑦 = −𝜂𝑢𝜉 + 𝑦𝑢𝜉 + 𝑢𝜂 ⇒ 𝑦𝑢𝑥 + 𝑢𝑦 = 𝑢𝜂 = 𝑥 = − 𝜉, i.e the
2
η2 𝜂3
canonical form is 𝑢𝜂 = − 𝜉, which is a first order ode in 𝜂. Integrating, we get 𝑢(𝜉, 𝜂) = −
2 6
𝑦3 𝑦3 𝑦2
𝜂𝜉 + 𝑓(𝜉) and therefore the required general solution is 𝑢(𝑥, 𝑦) = − + 𝑥𝑦 + 𝑓 ( 2 − 𝑥)
6 2

1 y2 1 y2
= 𝑥𝑦 − 3 𝑦 3 + 𝑓( 2 − 𝑥) 1.e 𝑢(𝑥, 𝑦) = 𝑥𝑦 − 3 𝑦 3 + 𝑓( 2 − 𝑥)

where 𝑓 is an arbitrary function.

6.4 Cauchy Problems
We determine a specific solution by formulating an Initial Value Problem or a Cauchy Problem.

36
Theorem (The Cauchy Problem for a First- Order Partial Differential Equation)

Suppose that 𝐶 is a given curve in the (𝑥, 𝑦) −plane with its parametric equations

𝑥 = 𝑥0 (𝑡), 𝑦 = 𝑦0 (𝑡) (1)

Where 𝑡 belongs to an interval 𝐼 ⊂ 𝑅, and the derivatives 𝑥0 ′(𝑡) and 𝑦0 ′(𝑡) are piecewise
continuous functions, such that (𝑥0 ′)2 + (𝑦0 ′)2 ≠ 0. Also suppose that 𝑢 = 𝑢0 (𝑡) is a given
function on the curve 𝐶. Then, there exists a solution 𝑢 = 𝑢(𝑥, 𝑦) of the equation
𝐹(𝑥, 𝑦, 𝑢, 𝑢𝑥 , 𝑢𝑦 ) = 0 (2)

in a domain 𝐷 of 𝑅 2 containing the curve C for all 𝑡 ∈ 𝐼, and the solution 𝑢(𝑥, 𝑦) satisfies the
initial data, that is

𝑢(𝑥0 (𝑡), 𝑦0 (𝑡)) = 𝑢0 (𝑡) (3)

For all values of 𝑡 ∈ 𝐼.

Example 6.3

Obtain the solution of the linear equation 𝑢𝑥 − 𝑢𝑦 = 1 with Cauchy data 𝑢(𝑥, 0) = 𝑥 2 .

Solution
The differential equation for characteristic is
𝑑𝑦
= −1⇒ the equation of characteristic is 𝑦 + 𝑥 = 𝐶1. Take the transformations as
𝑑𝑥

ξ(𝑥, 𝑦) = 𝑥 + 𝑦 and choose η= 𝑦.

𝑢𝑥 = 𝑢𝜉 𝜉𝑥 + 𝑢𝜂 𝜂𝑥 = 𝑢𝜉 and 𝑢𝑦 = 𝑢𝜉 𝜉𝑦 + 𝑢𝜂 𝜂𝑦 = 𝑢𝜉 + 𝑢𝜂

Therefore 𝑢𝑥 − 𝑢𝑦 = 𝑢𝜉 − 𝑢𝜉 − 𝑢𝜂 = −𝑢𝜂 , that is −𝑢𝜂 = 1 ⇒ 𝑢(𝜉, 𝜂) = −η + f(𝜉)

⇒ 𝑢(𝑥, 𝑦) = 𝑓(𝑥 + 𝑦) − 𝑦. Apply the Cauchy data, we get 𝑢(𝑥, 0) = 𝑥 2 = 𝑓(𝑥).

Hence 𝑢(𝑥, 𝑦) = (𝑥 + 𝑦)2 − 𝑦.

Example 6.4
Obtain the solution of the equation

𝑥𝑢𝑥 + 𝑦𝑢𝑦 = 𝑥𝑒 −𝑢

With the data 𝑢 = 0 on 𝑦 = 𝑥 2 .

Solution
𝑑𝑦 𝑦 𝑦
The differential equation for characteristic is 𝑑𝑥
= 𝑥 ⇒ the equation of characteristic is 𝑥
= 𝐶1 .
𝑦
Take the transformations as ξ(𝑥, 𝑦) = 𝑥 and choose η= 𝑦.

37
 𝑦 1
𝑢𝑥 = 𝑢𝜉 𝜉𝑥 + 𝑢𝜂 𝜂𝑥 = − 𝑥 2 𝑢𝜉 and 𝑢𝑦 = 𝑢𝜉 𝜉𝑦 + 𝑢𝜂 𝜂𝑦 = 𝑥 𝑢𝜉 + 𝑢𝜂
𝑦 𝑦 𝜂 1
Therefore 𝑥𝑢𝑥 + 𝑦𝑢𝑦 = − 𝑥 𝑢𝜉 + 𝑥 𝑢𝜉 + 𝑦𝑢𝜂 = −𝜂𝑢𝜂 , that is −𝜂𝑢𝜂 = 𝜉 𝑒 −𝑢 ⇒ 𝑢𝜂 = − 𝜉 𝑒 −𝑢

𝜂 𝑦
⇒𝑢(ξ, η) = ln (− 𝜉 + 𝑔(ξ))⇒𝑢(𝑥, 𝑦) = ln (𝑔 (𝑥 ) − 𝑥). Now apply the Cauchy data

⇒ 0 = 𝑙𝑛(𝑔(𝑥) − 𝑥) ⇒ 𝑔(𝑥) − 𝑥 = 1 ⇒ 𝑔(𝑥) = 𝑥 + 1 and hence the required solution is

𝑦
𝑢(𝑥, 𝑦) = ln ( + 1 − 𝑥)
𝑥

.E-tivity 6.2: The method of canonical forms

Numbering, pacing and sequencing 6.2.1 – 6.2.2

Title The method of canonical forms

Purpose To expose you to the method of canonical forms for

solving linear PDEs.

Brief summary of overall task Watch video on canonical formand then solve given
problems.
Spark
QUASILINEAR PDEs

Individual contribution • Watch videos on the methods of solving

quasilinear PDEs
• Answer the question;
Use the two methods to solve the PDE
𝑥𝑢𝑥 + 𝑦𝑢𝑦 = 2𝑥𝑦, with 𝑢 = 2 on 𝑦 = 𝑥 2

Interaction begins • Post your answers on the discussion forum 3.2.3

• Read what your colleagues have posted.
• In a sentence or two, comment on what two of
your colleagues have posted.
E-moderator interventions • Focussing group discussion
• Encouraging quiet ones to contribute
• Providing feedback/ teaching points
• Summarising key points

38
 • Closing the discussion

Schedule and time This activity should take 1 hour

Next Method of separation of variables

6.4 Assessment
1. Obtain the canonical form and find the general solution of each of the following
equations
a) 𝑢𝑥 + 𝑦𝑢𝑦 = 0
b) 𝑦𝑢𝑦 − 𝑥𝑢𝑥 = 1
c) (1 + 𝑥 2 )𝑢𝑥 + 𝑢𝑦 = 0
d) 𝑢𝑥 + 2𝑥𝑦 2 𝑢𝑦 = 0
2. Find the solution of the following Cauchy problems:
a) 3𝑢𝑥 + 2𝑢𝑦 = 0, 𝑢(𝑥, 0) = 𝑠𝑖𝑛𝑥
2
b) 𝑦𝑢𝑥 + 𝑥𝑢𝑦 = 0, 𝑢(0, 𝑦) = 𝑒 −𝑦
c) 𝑥𝑢𝑥 + 𝑦𝑢𝑦 = 2𝑥𝑦, 𝑢 = 2 𝑜𝑛 𝑦 = 𝑥 2
d) 𝑢𝑥 + 𝑥𝑢𝑦 = 0, 𝑢(0, 𝑦) = 𝑠𝑖𝑛𝑦

6.5 References
1 Tyn Myin-U and Lokeath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.
2 Frank Ayres (1952). Shaum’s outline of theory and problems of differential equations.
McGRAW-HILL publishers

39
 LESSON SEVEN

THE METHOD OF SEPARATION OF VARIABLES

7.1 Introduction
In this lesson we study the method of separation of variables as a tool for solving first order
PDEs. The method of separation of variables is a systematic method of solving PDEs. Its
essential feature is to transform the PDEs by a set of ODEs. The required solution of the PDE is
then exposed as a product 𝑢(𝑥, 𝑦) = 𝑋(𝑥)𝑌(𝑦) ≠ 0 or as the sum 𝑢(𝑥, 𝑦) = 𝑋(𝑥) + 𝑌(𝑦) where
𝑋(𝑥) and 𝑌(𝑦) are functions of 𝑥 and 𝑦 respectively.
7.2 Learning Outcomes
By the end of this lesson the learner will be able to:
7.2.1 Solve first order PDEs by the method of separation of variables, where the solution
will be given as a product 𝑢(𝑥, 𝑦) = 𝑋(𝑥)𝑌(𝑦) ≠ 0;
7.2.2 Solve first order PDEs by the method of separation of variables, where the solution
will be given as a sum 𝑢(𝑥, 𝑦) = 𝑋(𝑥) + 𝑌(𝑦)
7.3 The separation of variables where the solution of the PDE will be given as a product
of separated functions i.e. 𝒖(𝒙, 𝒚) = 𝑿(𝒙)𝒀(𝒚).
Usually, the first order PDE can be solved by separation of variables without the need for Fourier
series. We illustrate this method by the following examples.
Example 7.1

40
Solve the initial value problem

𝑢𝑥 + 2𝑢𝑦 = 0, 𝑢(0, 𝑦) = 4𝑒 −2𝑦

Solution

We seek for a separable solution 𝑢(𝑥, 𝑦) = 𝑋(𝑥)𝑌(𝑦) ≠ 0 and substitute into the equation to
obtain 𝑋(𝑥)′ 𝑌(𝑥) + 2𝑋(𝑥)𝑌(𝑦)′ = 0. This can be expressed in the form
𝑋(𝑥)′ 𝑌(𝑦)′
=− (1)
2𝑋(𝑥) 𝑌(𝑦)

Since the left hand side of this equation is a function of 𝑥 only and the right hand side is a
function of 𝑦 only, it follows that (1) can be true if both sides are equal to the same constant
value 𝜆 which is called an arbitrary separation constant. Consequently, (1) gives two ODEs

𝑋 ′ (𝑥) − 2𝜆𝑋(𝑥) = 0, 𝑌 ′ (𝑦) + 𝜆𝑌(𝑦) = 0 (2)

These equations have solutions given, respectively by

𝑋(𝑥) = 𝐴𝑒 2𝜆𝑥 and 𝑌(𝑦) = 𝐵𝑒 −𝜆𝑦 (3)

Where A and B are arbitrary integrating constants.

Consequently, the general solution is given by 𝑢(𝑥, 𝑦) = 𝐴𝐵𝑒 (2𝜆𝑥−𝜆𝑦) = 𝐶𝑒 (2𝜆𝑥−𝜆𝑦) where

𝐶 = 𝐴𝐵 is an arbitrary constant. Now, using the given condition, we have

4𝑒 −2𝑦 = 𝑢(0, 𝑦) = 𝐶𝑒 −𝜆𝑦 and hence, we deduce that 𝐶 = 4 and 𝜆 = 2. Therefore, the final
solution is 𝑢(𝑥, 𝑦) = 4𝑒 4𝑥−2𝑦 .
Example 7.2

Use the method of separation of variables 𝑢(𝑥, 𝑦) = 𝑋(𝑥)𝑌(𝑦) ≠ 0 to solve the equation
2 𝑥2
𝑦 2 (𝑢𝑥 )2 + 𝑥 2 (𝑢𝑦 ) = (𝑥𝑦𝑢)2 , 𝑢(𝑥, 0) = 3𝑒 4

Solution
2
𝑢𝑥 = 𝑋 ′ (𝑥)𝑌(𝑦), 𝑢𝑦 = 𝑋(𝑥)𝑌′(𝑦) and so 𝑦 2 (𝑢𝑥 )2 + 𝑥 2 (𝑢𝑦 ) = (𝑥𝑦𝑢)2 ⇒
2 2
′ ′
2 ′ (𝑥)𝑌(𝑦))2 2 (𝑋(𝑥)𝑌′(𝑦))2 1 (𝑋 (𝑥)) 1 (𝑌 (𝑦))
𝑦 (𝑋 +𝑥 = (𝑥𝑦𝑋(𝑥)𝑌(𝑦))2 ⇒ + = 1 and
𝑥 2 (𝑋(𝑥))2 𝑦 2 (𝑌(𝑦))2
hence
2 2
′ ′
1 (𝑋 (𝑥)) 1 (𝑌 (𝑦))
= 1 − 𝑦2 2 = 𝜆2 , where 𝜆2 is a separation constant. Thus
𝑥 2 (𝑋(𝑥))2 (𝑌(𝑦))

𝜆 2
1 𝑋′(𝑥) 1 𝑌′(𝑦)
= 𝜆 and = √1 − 𝜆2. Solving these ODEs, we find 𝑋(𝑥) = 𝐴𝑒 2𝑥 and
𝑥 𝑋(𝑥) 𝑦 𝑌(𝑦)

41
 1 2
𝑌(𝑦) = 𝐵𝑒 2𝑦√1−𝜆 , where 𝐴 and 𝐵 are arbitrary constants. Thus the general solution is
1 2 +𝑦√1−𝜆2 )
𝑢(𝑥, 𝑦) = 𝐶𝑒 2(𝜆𝑥 , where 𝐶 = 𝐴𝐵 is an arbitrary constant. Now apply the given
1 𝑥2
(𝜆𝑥 2 ) 1
condition, 𝑢(𝑥, 0) = 𝐶𝑒 2 = 3𝑒 4 , therefore 𝐶 = 3 and 𝜆 = 2. The required solution is
1 1 2 𝑦 1 2 +√3
𝑢(𝑥, 𝑦) = 3𝑒 2(2𝑥 + √3)
2 = 3𝑒 4(𝑥 𝑦)
.
7.4 The separation of variables where the solution of the PDE will be given as a sum of
separated functions i.e. 𝒖(𝒙, 𝒚) = 𝑿(𝒙) + 𝒀(𝒚).
We demonstrate how this method works with the help of the following example.
Example 7.3

Use the method of separation of variables 𝑢(𝑥, 𝑦) = 𝑓(𝑥) + 𝑔(𝑦) to solve the equation

𝑢𝑥 2 + 𝑢𝑦 2 = 1

Solution
2 2
𝑢𝑥 = 𝑓 ′ (𝑥), 𝑢𝑦 = 𝑔′(𝑦) and therefore 𝑢𝑥 2 + 𝑢𝑦 2 = 1 ⇒ (𝑓 ′ (𝑥)) = 1 − (𝑔′ (𝑦)) = 𝜆2
where 𝜆2 is a separation constant. Thus we obtain 𝑓 ′ (𝑥) = 𝜆 and 𝑔′ (𝑦) = √1 − 𝜆2.

Solving these ODEs, we find 𝑓(𝑥) = 𝜆𝑥 + 𝐴 and 𝑔(𝑦) = 𝑦√1 − 𝜆2 + 𝐵, where 𝐴 and 𝐵 are
constants of integration. Finally the solution is given by 𝑢(𝑥, 𝑦) = 𝜆𝑥 + 𝑦√1 − 𝜆2 + 𝐶, where

𝐶 = 𝐴 + 𝐵 is an arbitrary constant.
Example 7.4

Use 𝑣 = ln 𝑢 and 𝑣 = 𝑓(𝑥) + 𝑔(𝑥) to solve the equation 𝑥 2 𝑢𝑥 2 + 𝑦 2 𝑢𝑦 2 = 𝑢2

Solution
1 1
Since 𝑣 = ln 𝑢, 𝑣𝑥 = 𝑢 𝑢𝑥 and 𝑣𝑦 = 𝑢 𝑢𝑦 and hence we have 𝑥 2 {𝑓 ′ (𝑥)}2 + 𝑦 2 {𝑔′(𝑦)}2 = 1

Or 𝑥 2 {𝑓 ′ (𝑥)}2 = 1 − 𝑦 2 {𝑔′ (𝑦)}2 = 𝜆2 , where 𝜆2 is a separation constant. Thus we obtain

𝜆 1
𝑓 ′ (𝑥) = 𝑥 and 𝑔′ (𝑦) = 𝑦 √1 − 𝜆2. Integrating these equations gives

𝑓(𝑥) = 𝜆 ln 𝑥 + 𝐴 and 𝑔(𝑦) = √1 − 𝜆2 ln 𝑦 + 𝐵. Hence the required general solution is

2
𝑢(𝑥, 𝑦) = 𝐶𝑥 𝜆 𝑦 √1−𝜆 , where 𝐶 is an integrating constant.

E-tivity 7.2: The method of separation of variables

42
Numbering, pacing and sequencing 7.2.1 -7.2.2

Title The method of separation of variables

Purpose To expose you to the method of separation of variables

for solving first order PDEs.

Brief summary of overall task Watch video on the separation of variablesand then
solve given problems.
Spark
Separation of variables

Individual contribution • Watch videos on the methods of solving

quasilinear PDEs
• Answer the question;
Use the two methods to solve the PDE
𝑥𝑢𝑥 + 𝑦𝑢𝑦 = 2𝑥𝑦, with 𝑢 = 2 on 𝑦 = 𝑥 2

Interaction begins • Post your answers on the discussion forum 3.2.3

• Read what your colleagues have posted.
• In a sentence or two, comment on what two of
your colleagues have posted.
E-moderator interventions • Focussing group discussion
• Encouraging quiet ones to contribute
• Providing feedback/ teaching points
• Summarising key points
• Closing the discussion
Schedule and time This activity should take 1 hour
Next Type I and II nonlinear PDEs

7.5 Assessment

1. Apply the method of separation of variables 𝑢(𝑥, 𝑦) = 𝑓(𝑥)𝑔(𝑦) to solve

a) 𝑢𝑥 + 𝑢 = 𝑢𝑦 , 𝑢(𝑥, 0) = 4𝑒 −3𝑥
b) 𝑢𝑥 𝑢𝑦 = 𝑢2
c) 𝑢𝑥 + 2𝑢𝑦 = 0, 𝑢(0, 𝑦) = 3𝑒 −2𝑦
2. Apply the method of separation of variables 𝑢(𝑥, 𝑦) = 𝑓(𝑥) + 𝑔(𝑦) to solve
a) 𝑢𝑥2 + 𝑢𝑦2 = 1
b) 𝑢𝑥2 + 𝑢𝑦2 = 𝑢

43
 c) 𝑦𝑢𝑥 + 𝑥𝑢𝑦 = 0, 𝑢(0, 𝑦) = 𝑦 2

7.6 References
1 Tyn Myin-U and Lokeath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.
2 Frank Ayres (1952). Shaum’s outline of theory and problems of differential equations.
McGRAW-HILL publishers

44
 LESSON EIGHT

Non-linear Partial Differential Equations of Order One

8.1 Introduction

In this lesson we discuss special procedures for handling the first two of four types of non-linear
PDEs. This requires attention to the structure of the nonlinear PDE for one to really master the
techniques.
8.2 Learning Outcomes
By the end of this lesson the learner will be able to:
8.2.1 Solve nonlinear PDEs of type 1.
8.2.2 Solve nonlinear PDEs of type 2

8.3 Complete and singular solutions.

Let the non-linear partial differential equation of order one

𝐹(𝑥, 𝑦, 𝑢, 𝑝, 𝑞) = 0 (1)

be derived from

𝑔(𝑥, 𝑦, 𝑢, 𝑎, 𝑏) = 0 (2)

by eliminating the arbitrary constants 𝑎and𝑏. Then (2) is called a (or the) complete solution of

1). This complete solution represents a two-parameter family of surfaces which mayor may not

have an envelope. To find the envelope (if one exists) we eliminate 𝑎 and 𝑏 from

𝜕𝑔 𝜕𝑔
𝑔 = 0, = 0, =0
𝜕𝑎 𝜕𝑏

If the eliminant

45
 a) λ(𝑥, 𝑦, 𝑢) = 0

b) satisfies 1), it is called the singular solution of 1); if

𝜆(𝑥, 𝑦, 𝑢) = 𝑔(𝑥 𝑦 𝑢). 𝜂(𝑥, 𝑦, 𝑢)

and if 𝜉 = 0 satisfies (1) while 𝜂 = 0 does not, ξ = 0 is the singular solution. As in the case of

ordinary differential equations , the singular solution may be obtained from the partial

differential equation by eliminating 𝑝 and 𝑞 from

𝜕𝑓 𝜕𝑓
𝑓 = 0, = 0, = 0
𝜕𝑝 𝜕𝑞

Example 8.1

It is readily verified that 𝑢 = 𝑎𝑥 + 𝑏𝑦 − (𝑎2 + 𝑏 2 ) is a complete solution of

𝑢 = 𝑝𝑥 + 𝑞𝑦 − (𝑝2 + 𝑞 2 ). Eliminating 𝑎 and 𝑏 from

𝜕𝑔 𝜕𝑔
𝑔 = 𝑢 − 𝑎𝑥 – 𝑏𝑦 + 𝑎2 + 𝑏 2 = 0, = −𝑥 + 2𝑎 = 0, = −𝑦 + 2𝑏 = 0,
𝜕𝑎 𝜕𝑏

1 1 1 1
we have 𝑢 = 𝑥2 + 𝑦2 − (𝑥 2 + 𝑦 2 ) = (𝑥 2 + 𝑦 2 ). This satisfies the differential
2 2 4 4

equation and is the singular solution. The complete solution represents a two-parameter family of

planes which envelope the paraboloid 𝑥 2 + 𝑦 2 = 4𝑢.

General solution

If, in the complete solution (2), one of the constants, say 𝑏, is replaced by 𝑎known function of

the other, say 𝑏 = 𝜗(𝑎), then

𝑔(𝑥, 𝑦, 𝑢, 𝑎, 𝜗(𝑎)) = 0

46
is a one-parameter family of the surfaces of 1). If this family has an envelope, its equation may

be found as usual by eliminating 𝑎 from

𝜕
𝑔(𝑥, 𝑦, 𝑢, 𝑎, 𝜗(𝑎)) = 0 and 𝑔(𝑥, 𝑦, 𝑢, 𝑎, 𝜗(𝑎))
𝜕𝑎

and determining that part of the result which satisfies 1).

Example 8.2

Set 𝑏 = 𝜗(𝑎) = 𝑎 in the complete solution of Example 8.1. The result of eliminating 𝑎 from

𝜕𝑔 1
𝑔 = 𝑢 − 𝑎(𝑥 + 𝑦) + 2𝑎2 = 0 and𝜕𝑎 = −(𝑥 + 𝑦) + 4𝑎 = 0is𝑢 = (𝑥 + 𝑦)2
8

which can be readily shown to satisfy the differential equation of Example 8.1. This is a pa-

rabolic cylinder with its elements parallel to the 𝑥𝑦 − plane.

The totality of solutions obtained by varying 𝜗(𝑎) is called the general solution of the

differential equation. Thus, from Example 8.2, 8𝑧 = (𝑥 + 𝑦)2 is included in the general

solution of the differential equation of Example 8.1.

When 𝑏 = 𝜗(𝑎), 𝜗 arbitrary, is used, the elimination of a between

𝜕𝑔
𝑔 = 0 and =0
𝜕𝑎

is not possible; hence, we are unable to express the general solution as a single equation,

involving an arbitrary function, as we were in the case of the linear equation.

Solutions.

Before considering a general method for obtaining a complete solution of 1), we give

47
special procedures for handling four types of equations.

𝐹(𝑥, 𝑦, 𝑢, 𝑝, 𝑞) = 0 (1)

be derived from

𝑔(𝑥, 𝑦, 𝑢, 𝑎, 𝑏) = 0 (2)

TYPEI: 𝑓(𝑝, 𝑞) = 0.

This PDE has a complete solution of the form 𝑢 = 𝑎𝑥 + ℎ(𝑎)𝑦 + 𝑐

b= ℎ(𝑎)

examples of type I pdes

p2 + q2 =1

pq + q2 =4

form 𝑢 = 𝑎𝑥 + 𝑏𝑦 + 𝑐

48
Example8.3

𝑝2 − 𝑞 2 = 1 is of type 1.

It follows that a complete solution is𝑢 = 𝑎𝑥 + ℎ(𝑎) 𝑦 + 𝑐 ,

where 𝑓(𝑎, ℎ (𝑎)) = 0, and 𝑎 and 𝑐 are arbitrary constants.

The equations for determining the singular solution are

𝑧 = 𝑎𝑥 + ℎ (𝑎) 𝑦 + 𝑐,

0 = 𝑥 + ℎ′ (𝑎)𝑦,

0 = 1.

Thus, there is no singular solution.

The general solution is obtained by putting 𝑐 = 𝜗(𝑎), 𝜗 arbitrary, and eliminating 𝑎 between

𝑢 = 𝑎𝑥 + ℎ ( 𝑎) 𝑦 + 𝜗 ( 𝑎 ) and 0 = 𝑥 + ℎ′ ( 𝑎) 𝑦 + 𝜗′(𝑎) (1)

The first equation of (1) for a stipulated function 𝜗 ( 𝑎 )represents a one-parameter family of

planes and its envelope (a part of the general solution) is a developable surface.

49
Example 8.4

Solve𝑝2 − 𝑞 2 = 1.

Solution 𝑎2 − 𝑏 2 − 1 = 0

𝑧 = 𝑎𝑥 + ℎ (𝑎) 𝑦 + 𝑐,

Here 𝑓(𝑝, 𝑞) = 𝑝2 − 𝑞 2 − 1 = 0, 𝑓(𝑎, ℎ(𝑎)) = 𝑎2 − [ℎ(𝑎)]2 − 1 = 0 and

1
ℎ(𝑎) = (𝑎2 − 1)2 .

1
A complete solution is 𝑧 = 𝑎𝑥 + (𝑎2 − 1)2 𝑦 + 𝑐.

1 1
0 = 𝑥 + (𝑎2 − 1)−2 (2𝑎)𝑋 𝑦
2

0=1

A neater form is obtained by putting 𝑎 = 𝑠𝑒𝑐 𝛼; then ℎ(𝑎) = 𝑡𝑎𝑛 𝛼 and we have

𝑢 = 𝑥 𝑠𝑒𝑐 𝛼 + 𝑦 𝑡𝑎𝑛 𝛼 + 𝑐.

50
If we set 𝑐 = 𝜗(𝛼) = 0, the result of eliminating 𝑎 from

𝑢 = 𝑥 𝑠𝑒𝑐 𝛼 + 𝑦 𝑡𝑎𝑛 𝛼. 0 = 𝑥 𝑡𝑎𝑛 𝛼 + 𝑦 𝑠𝑒𝑐 𝛼 or 0 = 𝑥 𝑠𝑖𝑛 𝛼 + 𝑦

is 𝑢2 = 𝑥 2 − 𝑦 2 .

This developable surface (cone) is a part of the general solution of the given differential

equation.

1
Note that we might have taken ℎ(𝛼) = −(𝑎2 − 1)2 and obtained as a complete solution

1
𝑢 = 𝑎𝑥 − (𝑎2 − 1)2 𝑦 + 𝑐.

Example 8.5

Solve 𝑝 2 + 𝑞 2 = 9.

Solution

A complete solution is 𝑧 = 𝑎𝑥 + 𝑏𝑦 + 𝑐, where 𝑎2 + 𝑏 2 = 9. The equations for

determining the singular solution are

𝑎
𝑧 = 𝑎𝑥 + √9 − 𝑎2 𝑦 + 𝑐, 0 = 𝑥 − √9−𝑎2
𝑦, 0 = 1. Thus, there is no singular solution.

51
Example 8.6

Solve 𝑝𝑞 + 𝑝 + 𝑞 = 0.

Solution

A complete solution is 𝑧 = 𝑎𝑥 + 𝑏𝑦 + 𝑐,

where 𝑎𝑏 + 𝑎 + 𝑏 = 0 or 𝑧 = 𝑎𝑥 − 𝑎/(𝑎 + 1) 𝑦 + 𝑐

There is no singular solution

TYPE II: 𝑢 = 𝑝𝑥 + 𝑞𝑦 + 𝑓(𝑝, 𝑞)

This PDE has the solution 𝑢 = 𝑎𝑥 + 𝑏𝑦 + 𝑓(𝑎, 𝑏)

Example 8.7

Find the complete solution of

1 1
𝑢 = 𝑝𝑥 + 𝑞𝑦 + 3𝑝3 𝑞 3

Solution

The complete solution is

𝑢 = 𝑎𝑥 + 𝑏𝑦 + 𝑓(𝑎, 𝑏).

This is known as the extended Clairaut type, for obvious reasons. This complete solution consists

of a two-parameter family of planes. The singular solution (if one exists) is a surface having the

complete solution as its tangent planes.

52
Example 8.8.

1 1
Find the singular solution of 𝑢 = 𝑝𝑥 + 𝑞𝑦 + 3𝑝3 𝑞 3

Solution

1 1
A complete solution is 𝑢 = 𝑎𝑥 + 𝑏𝑦 + 3𝑝3 𝑞 3

The derivatives with respect to 𝑎 and 𝑏are 𝑥 + 𝑎 −2/3 𝑏1/3 = 0 and 𝑦 + 𝑎1/3 𝑏 −2/3 = 0.

Then 𝑎𝑥 + 𝑏𝑦 = − 2𝑎1/3 𝑏1/3, 𝑥𝑦 = 𝑎−1/3 𝑏 −1/3

and, substituting in the complete solution, we obtain the singular solution

1
𝑢 = 𝑎1/3 𝑏1/3 = 𝑥𝑦 or 𝑥𝑦𝑢 = 1.

Example 8.9

Solve 𝑧 = 𝑝𝑥. + 𝑞𝑦 + 𝑝2 + 𝑝𝑞 + 𝑞 2

Solution

A complete solution is 𝑧 = 𝑎𝑥 + 𝑏𝑦 + 𝑎2 + 𝑎𝑏 + 𝑏 2

Differentiating the complete solution with respect to a and b, we have

0 = 𝑥 + 2𝑎 + 𝑏, 0 = 𝑦 + 𝑎 + 2𝑏.

(𝑦 − 2𝑥) (𝑥 − 2𝑦)
Solving to obtain 𝑎 = , 𝑏= and substituting in the complete solution, the
3 3

singular solution is 3𝑧 = 𝑥𝑦 − 𝑥 2 − 𝑦 2

53
Example 8.10

Solve 𝑧 = 𝑝𝑥 + 𝑞𝑦 + 𝑝2 𝑞 2

Solution

A complete solution is 𝑧 = 𝑎𝑥 + 𝑏𝑦 + 𝑎2 𝑏 2 . The equations obtained by differentiating with

3 𝑦2 3 𝑥2
respect to α and b are 0 = 𝑥 + 2𝑎𝑏 2 and 0 = 2𝑎2 𝑏. Then𝛼 = − √2𝑥 , 𝑏 = − √2𝑦 ,

3 𝑦2 3 𝑥2 3 𝑥2𝑦2 33
and the singular solution is 𝑧 = − 𝑥 √2𝑥 − 𝑦 √2𝑦 + √ = − 4 √4 𝑥 2/3 𝑦 2/3
16

54
Etivity 8.2:Nonlinear PDEs of types I and II
Numbering and pacing and 8.2.1 – 8.2.2
sequencing
Title Non-linear PDEs of types I and II
Purpose To help you to know how to solve non-linear PDEs of
type 1 and II.
Brief summary of overall task Watch the video on non-linear PDE and solve the given
questions.
Spark 𝑇𝑌𝑃𝐸 1 𝐴𝑁𝐷 𝐼𝐼 𝑁𝑂𝑁𝐿𝐼𝑁𝐸𝐴𝑅 𝑃𝐷𝐸𝑠
Individual contribution Watch the video on non-linear PDE.
Solve 𝑝2 + 𝑝 = 𝑞 2 and 𝑧 = 𝑥𝑝 + 𝑦𝑞 + 𝑝𝑞

Interaction begins • Post your answers on discussion forum 8.2.3

• Read what your colleagues have posted.
• In a sentence or two, comment on what two of
your colleagues have posted keeping netiquette in
mind

E-moderator interventions • Focussing group discussion

• Encouraging quiet ones to contribute
• Providing feedback/ teaching points

• Closing the discussion

Schedule and time This activity should take one hour.
Next Type III and IV nonlinear PDEs

8.4 Assessment

1. Find the complete solution of 𝑝2 + 𝑝 = 𝑞 2

2. Find the complete ( an singular solution if it exists) of 𝑦𝑝 − 𝑥 2 𝑞 2 = 𝑥 2 𝑦

8.5 References
1. Tyn Myin-U and Lokeath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.

55
 2. Frank Ayres (1952). Shaum’s outline of theory and problems of differential equations.
McGRAW-HILL publishers

LESSON NINE

NON-LINEAR PARTIAL DIFFERENTIAL EQUATIONS OF TYPES THREE AND FOUR

9.1 Introduction

In this lesson we discuss nonlinear PDEs of type three and four. The success of solving PDEs
that fall into these types of PDEs lies on recognizing the structure of the PDE.
9.2 Learning Outcomes
By the end of this lesson the learner will be able to:
9.2.1 Solve non-linear PDE of type thre;.
9.2.2 Solve nonlinear PDE of type four.

9.3 PDEs of Type III: 𝑓(𝑧, 𝑝, 𝑞) = 0. f(x,y,z,p,q)=0 sol g(x,y,z,a,b)

Let z= F(x + ay) =F(u) u =(x + ay)

Example 9.1

The PDE 𝑧 = 𝑝2 + 𝑞 2 . To solve this:

Assume 𝑧 = 𝐹(𝑥 + 𝑎𝑦) = 𝐹(𝑢), where a is the arbitrary constant. Then

𝜕𝑧 𝑑𝑧 𝜕𝑢 𝑑𝑧 𝑑𝑧 𝜕𝑢 𝑑𝑧
𝑝= = = and 𝑞 = = 𝑎
𝜕𝑥 𝑑𝑢 𝜕𝑥 𝑑𝑢 𝑑𝑢 𝜕𝑦 𝑑𝑢

When these are substituted in the given differential equation, we obtain an ordinary differential

equation of order one

𝑑𝑧 𝑑𝑧
ƒ(𝑧, ,𝑎 ) = 0
𝑑𝑢 𝑑𝑢

56
whose solution is the required complete solution.

Example 9.2

Solve 𝑧 = 𝑝2 + 𝑞 2 .

Solution

𝑑𝑧 𝑑𝑧
Put𝑧 = 𝐹(𝑥 + 𝑎𝑦) = 𝐹(𝑢). Then 𝑝 = ,𝑞 =𝑎 and the given equation may be reduced to
𝑑𝑢 𝑑𝑢

𝑑𝑧 2 𝑑𝑧 2
𝑧 = (𝑑𝑢) + 𝑎2 (𝑑𝑢) .

𝑑𝑧 2
Z = (𝑑𝑢) (1 + 𝑎2 )

Separation of variables and integration

𝑑𝑧 √𝑧 𝑑𝑧 1 1 1
Solving = √1+𝑎2 = or = √1+𝑎2
𝑑𝑢, we obtain 2√𝑧 = √1+𝑎2
𝑢 + 𝑘 = √1+𝑎2
(𝑢 + 𝑏).
𝑑𝑢 √𝑧

Squaring both sides and rearranging

Thus, a complete solution is 4(1 + 𝑎2 )𝑧 = (𝑥 + 𝑎𝑦 + 𝑏)2 , a family of parabolic cylinders.

Taking the derivatives with respect to 𝑎 and 𝑏, we have

8𝑎𝑧 − 2(𝑥 + 𝑎𝑦 + 𝑏)𝑦 = 0, The singular solution is 𝑧 = 0.

⇒𝑥 + 𝑎𝑦 + 𝑏 = 0.

57
Example 9.3

Solve 4(1 + 𝑧 3 ) = 9𝑧 4 𝑝𝑞

Solution

𝑑𝑧 𝑑𝑧
Assume 𝑧 = 𝐹(𝑥 + 𝑎𝑦) = 𝐹(𝑢). Then 𝑝 = , 𝑞= 𝛼 ,
𝑑𝑢 𝑑𝑢

𝑑𝑧 ∛𝑎 𝑧 2
and the given equation becomes 4(1 + 𝑧 3 ) = 9𝛼𝑧 4 (𝑑𝑢)2 or√1+𝑧 3 𝑑𝑧 = 2𝑑𝑢.

Integrating, √𝑎 (1 + 𝑧 3 ) = 𝑢 + 𝑏,

and a complete solution is 𝑎(1 + 𝑧 3 ) = (𝑥 + 𝑎𝑦 + 𝑏)2

Using the results of differentiating this with respect to a and b,

1 + 𝑧 3 = 2(𝑥 + 𝑎𝑦 + 𝑏)𝑦 and 0 = 2(𝑥 + 𝑎𝑦 + 𝑏).

the Singular solution is𝑧 3 + 1 = 0.

Example 9.4

Solve 𝑝(1 − 𝑞 2 ) = 𝑞(1 − 𝑧).

Solution

𝑑𝑧 𝑑𝑧
Assume 𝑧 = 𝐹(𝑥 + 𝑎𝑦) = 𝐹(𝑢). Then 𝑝 = , 𝑞 = 𝛼 𝑑𝑢 and the given equation becomes
𝑑𝑢

58
 𝑑𝑧 𝑑𝑧 𝑑𝑧 𝑑𝑧 𝑑𝑧
(𝑑𝑢) [1 − 𝑎2 (𝑑𝑢)2 ] = 𝑎 𝑑𝑢 (1 − 𝑧)or (𝑑𝑢) [1 − 𝑎 + 𝑎𝑧 − 𝑎2 (𝑑𝑢)2 ] = 0

𝑑𝑧 𝑑𝑧 2 𝑎𝑑𝑧
Then 𝑑𝑢 = 0and 𝑧 = 𝑐; or 1 − 𝑎 + 𝑎𝑧 − 𝑎2 (𝑑𝑢) = 0, = 𝑑𝑢and
√1−𝑎+𝑎𝑧

2√1 − 𝑎 + 𝑎𝑧 = 𝑢 + 𝑏 = 𝑥 + 𝑎𝑦 + 𝑏 or 4(1 − 𝑎 + 𝑎𝑧) = (𝑥 + 𝑎𝑦 + 𝑏)2

Each of 𝑧 = 𝑐 and 4(1 − 𝑎 + 𝑎𝑧) = (𝑥 + 𝑎𝑦 + 𝑏) is a solution; the latter is a complete

solution. Using it, the equations for obtaining the singular solution are

𝜕𝑔 𝜕𝑔
𝑔 = 4(1 − 𝑎 + 𝑎𝑧) − (𝑥 + 𝑎𝑦 + 𝑏)2 = 0, = 4(−1 + 𝑧) − 2𝑦(𝑥 + 𝑎𝑦 + 𝑏) = 0,
𝜕𝑎 𝜕𝑏

= −2(𝑥 + 𝑎𝑦 + 𝑏) = 0

there is no singular solution.

Example 9.5

Solve 1 + 𝑝2 = 𝑞𝑧.

Solution

𝑑𝑧 𝑑𝑧
Assume 𝑧 = 𝐹( 𝑥 + 𝑎𝑦) = 𝐹(𝑢). Then𝑝 = 𝑑𝑢 , 𝑞 = 𝑎 𝑑𝑢,, and the given equation becomes

𝑑𝑧 𝑑𝑧 𝑑𝑧 1
(𝑑𝑢)2 − 𝑎𝑧 𝑑𝑢 + 1 = 0or 𝑑𝑢 = 2 𝑑𝑢.
𝑎𝑧−√𝑎2 𝑧 2 − 4

Rationalizing the left member of the latter equation, we obtain

(𝑎𝑧 + √𝑎2 𝑧 2 − 4)𝑑𝑧 = 2 𝑑𝑢, whose solution is

59
 1 2 1 𝑎𝑧
𝑎𝑧 + [ √𝑎2 𝑧 2 − 4 − 2 𝑙𝑛 (𝑎𝑧 + √𝑎2 𝑧 2 − 4] = 2(𝑢 + 𝑏).
2 𝑎 2

A complete solution is then

𝑎2 𝑧 2 + √𝑎2 𝑧 2 − 4 − 41𝑛(𝑎𝑧 + √𝑎2 𝑧 2 − 4) = 4𝑎(𝑥 + 𝑎𝑦 + 𝑏).

Note that .𝑎2 𝑧 2 + 𝑎𝑧√𝑎2 𝑧 2 − 4 + 4𝑙𝑛(𝑎𝑧 + √𝑎2 𝑧 2 − 4) = 4𝑎(𝑥 + 𝑎𝑦 + 𝑏), obtained from

𝑑𝑧 1
= 2 𝑑𝑢, is also a complete solution.
𝑎𝑧+ √𝑎2 𝑧 2 − 4

TYPE IV: 𝑓1 (𝑥, 𝑝) = 𝑓2 (𝑦, 𝑞). separable

The PDE 𝑝 − 𝑥 2 = 𝑞 + 𝑦 2 is of type IV. To solve PDEs in this class, we set

𝑓1 (𝑥, 𝑝) = 𝑎, 𝑓2 (𝑦, 𝑞) = 𝑎 , where a is an arbitrary constant, and solve to obtain

and 𝑝 = 𝑓1 (𝑥, 𝑎) and 𝑞 = 𝑓2 (𝑦, 𝑎)

Since 𝑧 is a function .of 𝑥 and y,𝑑𝑧 = 𝑝𝑑𝑥 + 𝑞𝑑𝑦 = 𝐹1 (𝑥, 𝑎)𝑑𝑥 + 𝐹2 (𝑦, 𝑎)𝑑𝑦.

Thus, 𝑧 = ∫ 𝐹1 (𝑥, 𝑎) 𝑑𝑥 + ∫ 𝐹2 (𝑦, 𝑎) 𝑑𝑦 + 𝑏, (1) containing two arbitrary

constants, is the required complete solution.

Example 9.6

Solve 𝑝 − 𝑞 = 𝑥 2 + 𝑦 2

60
Solution

The PDE can be written as 𝑝 − 𝑥 2 = 𝑞 + 𝑦 2 . Setting𝑝 − 𝑥 2 = 𝑎, 𝑞 + 𝑦 2 = 𝑎, we obtain

𝑝 = 𝑎 + 𝑥2, 𝑞 = 𝑎 − 𝑦2

Integrating 𝑑𝑧 = 𝑝𝑑𝑥 + 𝑞𝑑𝑦 = (𝑎 + 𝑥 2 )𝑑𝑥 + (𝑎 − 𝑦 2 )𝑑𝑦, the required complete solution

𝑥3 𝑦3
is 𝑧 = 𝑎𝑥 + + 𝑎𝑦 − + 𝑏. There is no singular solution.
3 3

Example 9.7

Solve √𝑝 − √𝑞 + 3𝑥 = 0

Solution

Set √𝑝 + 3𝑥 = 𝑎 and √𝑞 = 𝑎. Then 𝑝 = (𝑎 − 3𝑥)2 and 𝑞 = 𝑎2. A complete solution is

𝑧 = ∫ 𝑝 𝑑𝑥 + ∫ 𝑞𝑑𝑦 + 𝑏 = ∫ (𝑎 − 3𝑥)2 𝑑𝑥 + 𝑎2 ∫ 𝑑𝑦 + 𝑏 or

1
𝑧 = − 9 (𝑎 − 3𝑥)3 + 𝑎2 𝑦 + 𝑏.There is no singular solution.

Example 9.8

Solve𝑞 = −𝑝𝑥 + 𝑝2 .

Solution

1
Set 𝑝2 − 𝑝𝑥 = 𝑎 and 𝑞 = 𝑎. Then 𝑝 = (𝑥 + √𝑥 2 + 4𝑎).
2

𝑝2 − 𝑝𝑥 − 𝑎 = 0

61
 1
A complete solution is 𝑧 = ∫ (𝑥 + √𝑥 2 + 4𝑎)𝑑𝑥 + 𝑎∫ 𝑑𝑦 + 𝑏.
2

1
Or 𝑧 = (𝑥 2 + 𝑥 √𝑥 2 + 4𝑎 + 𝛼 𝑙𝑛 (𝑥 + √𝑥 2 + 4𝑎) + 𝑎𝑦 + 𝑏
4

There is no singular solution.

Etivity 9.2: Non-linear PDEs of type III and IV

62
Numbering and pacing and 9.2.1 -9.2.2
sequencing
Title Nonlinear PDEs of type III and IV
Purpose To help you to know how to solve PDEs of type III and
IV.
Brief summary of overall task Watch the video on NONLINEAR and solve the given
questions.
Spark TYPE III AND IV NONLINEAR PDEs
Individual contribution a) Watch the video on NONLINEAR and solve
𝑝2 + 𝑞 2 = 4𝑧
𝑧 2 (𝑝2 + 𝑞 2 + 1) = 1
Interaction begins • Post your answers on discussion forum 9.2.3
• Read what your colleagues have posted.
• In a sentence or two, comment on what two of
your colleagues have posted.
E-moderator interventions • Focussing group discussion
• Encouraging quiet ones to contribute
• Providing feedback/ teaching points

• Closing the discussion

Schedule and time This activity should take one hour.
Next Method of Transformations
9.4 Assessment
1. Find the complete and singular solution of
𝑧 2 (𝑝2 𝑧 2 + 𝑞 2 ) = 1
2. Find the complete integral of 𝑝 − 3𝑥 2 = 𝑞 2 − 𝑦 (Type iv)
3. P(1+q) = qz

4. 1 + q2 = q(z-a)

5. Px + qy = 9q

6. q- p + x –y =0

7. q= xp + p2

8. 𝑝2 𝑦 (1 + 𝑥 2 ) = 𝑞𝑥 2

𝑝2
63
 9. 𝑝2 𝑦 (1 + 𝑥 2 ) = 𝑞𝑥 2

𝑎𝑥
𝑝=
√(1 + 𝑥 2 )

𝑞 𝑝2
= (1 + 𝑥 2 ) = 𝑎2
𝑦 𝑥2

q= 𝑎2 𝑦

dz =pdx + qdy

z = ∫ 𝑝𝑑𝑥 + ∫ 𝑞𝑑𝑦

9.5 References
1. Tyn Myin-U and Lokeath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.
2. Frank Ayres (1952). Shaum’s outline of theory and problems of differential equations.
McGRAW-HILL publishers

64
 LESSON TEN

SOLUTION OF NON-LINEAR PDEs USING TRANSFORMATIONS

10.1 Introduction

In this lesson we discuss the method of transformations for solving non-linear first order PDEs. It
may be possible, at times to find transformation of variables which will reduce a given equation
to one of the four types that we have discussed.
10.2 Learning Outcomes
By the end of this lesson the learner will be able to:
10.2.1 Recognize the right transformation to use for a given PDE;
10.2.2 Solve PDE using the transformation method.

10.3 Transformations

As in the case of ordinary differential equations, it is possible at times to find a transformation of

the variables which will reduce a given equation to one of the above four types.

The combination 𝑝𝑥, for example, suggests the transformation 𝑋 = 𝑙𝑛 𝑥, since then

𝜕𝑧 𝜕𝑧 𝑑𝑋 1 𝜕𝑧 𝜕𝑧
𝑝 = = 𝜕𝑋 𝑑𝑥 = and 𝑝𝑥 =
𝜕𝑥 𝑥 𝜕𝑋 𝜕𝑋

𝜕𝑧 𝜕𝑧 𝜕𝑧
Thus, 𝑞 = 𝑝𝑥 + 𝑝2 𝑥 2 becomes = + (𝜕𝑋)2 (Type 1)
𝜕𝑦 𝜕𝑋

Similarly, the combination 𝑞𝑦 suggests the transformation 𝑌 = 𝐼𝑛 𝑦.

𝑝𝑞
The appearance of 𝑧 ,𝑧 in an equation suggests the transformation Z = In z,

𝜕𝑧 𝜕𝑧 𝜕𝑍 𝜕𝑍 𝑝 𝜕𝑍 𝑞 𝜕𝑍
since then 𝑝 = = = 𝑧 and 𝑧 = 𝜕𝑥 ; similarly , 𝑧 = 𝜕𝑦
𝜕𝑥 𝜕𝑍 𝜕𝑥 𝜕𝑥

65
 𝑞 𝑝 2 𝜕𝑍 𝜕𝑍 2
Thus, 𝑧 = (𝑧 ) becomes𝜕𝑦 = ( 𝜕𝑥) , of type I

Example 10.1

Solve 4𝑥𝑦𝑧 = 𝑝𝑞 + 2𝑝𝑥 2 𝑦 + 2𝑞𝑥𝑦 2

1 1 1 1 1 1 1 1
𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧
4𝑋 2 𝑌 2 𝑧 = 2𝑋 2 𝜕𝑋 2𝑌 2 𝜕𝑌 + 2.2𝑋 2 𝜕𝑋 𝑋𝑌 2 + 2.2𝑌 2 𝜕𝑌 . 𝑋 2 𝑌

𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧
𝑧= + 𝑋+ .𝑌
𝜕𝑋 𝜕𝑌 𝜕𝑋 𝜕𝑌

Z = px + qy + f(p,q)

Z= PX +QY + PQ

Z = aX + bY +ab

Z= ax2 + by2 +ab complete solution

Solution

1 1 1 1
𝜕𝑧 𝜕𝑧 𝜕𝑋 𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑌 𝜕𝑧
Let 𝑥 = 𝑋 2 , 𝑦 = 𝑌 2 . Then 𝑝 = 𝜕𝑥 = 𝜕𝑋 𝜕𝑥 = 2𝑋 2 𝜕𝑋 and 𝑞 = 𝜕𝑥 = 𝜕𝑌 𝜕𝑦 = 2𝑌 2 𝜕𝑌.

𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧
Substituting in the given equation, we have 𝑧 = 𝑋 𝜕𝑋 + 𝑌 𝜕𝑌 + 𝜕𝑋 𝜕𝑌 of type II.

66
A complete solution is 𝑧 = 𝑎𝑋 + 𝑏𝑌 + 𝑎𝑏 or 𝑧 = 𝑎𝑥 2 + 𝑏𝑦 2 + 𝑎𝑏.

Z= −𝑦 2 𝑥 2 − 𝑥 2 𝑦 2 + 𝑥 2 𝑦 2 so that z =−𝑥 2 𝑦 2

Eliminating a and b from this and 0 = 𝑥 2 + 𝑏, 0 = 𝑦 2 + 𝑎, obtained by differentiating it with

respect to a and b, the singular solution is found to be 𝑧 + 𝑥 2 𝑦 2 = 0

Example 10.2

Solve 𝑝𝑞 = 𝑥 𝑚 𝑦 𝑛 𝑧 2𝑙

Solution

𝑝𝑧 −𝑙 𝑞𝑧 −𝑙
The given PDE can be written as 𝑥 𝑚 . = 1.
𝑦𝑛

The transformation

𝑧1−𝑙 𝑥 𝑚+1
𝑍= , 𝑋 = ,
1−𝑙 𝑚+1

𝑦 𝑛+1 𝜕𝑍 𝜕𝑍 𝜕𝑥 1 𝜕𝑍 𝜕𝑍 𝜕𝑦 1
𝑌= , = = 𝑧 −1 𝑝 𝑚 , = = 𝑧 −1 𝑞 𝑛
𝑛 + 1 𝜕𝑋 𝜕𝑥 𝜕𝑋 𝑥 𝜕𝑌 𝜕𝑦 𝜕𝑌 𝑦

𝜕𝑍 𝜕𝑍
Reduces the given differential equation to 𝜕𝑋 . 𝜕𝑌 = 1.

1
This equation is of Type I and its solution is 𝑍 = 𝑎𝑋 + 𝑌+𝑐
𝑎

𝑧 1−𝑙 𝑥 𝑚+1 𝑦 𝑛+1

A complete solution of the given equation is =𝑎 + 𝑎(𝑛+1) + 𝑐.
1−𝑙 𝑚+1

There is no singular solution.

𝑧 −1 𝑥3 𝑦4
Solve 𝑝𝑞 = 𝑥 2 𝑦 3 𝑧 4 has the solution =𝑎 + 𝑎(4) + 𝑐.
−1 3

67
Example 10.3

Solve 𝑥 2 𝑝2 + 𝑦 2 𝑞 2 = 𝑧.

(xp)2

Solution

1) The transformation

1 𝜕𝑍 𝜕𝑍 𝜕𝑥 1
𝑋 = 𝐼𝑛 𝑥, 𝑌 = 𝐼𝑛 𝑦, 𝑍 = 2𝑧 2 , = = 𝑝𝑥𝑧 −2 ,
𝜕𝑋 𝜕𝑥 𝜕𝑋

𝜕𝑍 𝜕𝑍 𝜕𝑦 1
= = 𝑝𝑦𝑧 −2
𝜕𝑌 𝜕𝑦 𝜕𝑌

reduces the given equation to

𝜕𝑍 𝜕𝑍 𝜕𝑍 𝜕𝑍
𝑧(𝜕𝑋)2 + 𝑧(𝜕𝑌)2 = z or (𝜕𝑋)2 + (𝜕𝑌)2= 1, of type 1

A complete solution is 𝑍 = 𝑎𝑋 + 𝑏𝑌 + 𝑐 or 4𝑧 = (𝑎 𝑙𝑛𝑥 + 𝑏𝐼𝑛 𝑦 + 𝑐)2 ,

where 𝑎2 + 𝑏 2 = 1.

The singular solution is 𝑧 = 0.

𝜕𝑧 𝑑𝑋 𝜕𝑥 1 𝜕𝑧 1 𝜕𝑧
2) The transformation 𝑋 = 𝑙𝑛 𝑥, 𝑌 = 𝑙𝑛 𝑦, 𝑝 = = = = , 𝑞=
𝜕𝑥 𝜕𝑋 𝑑𝑥 𝑥 𝜕𝑋 𝑦 𝜕𝑌

𝜕𝑧 2 𝜕𝑧 2
reduces the given differential equation to(𝜕𝑋) + (𝜕𝑌) = 𝑧 of type III

𝜕𝑧 𝑑𝑧 𝜕𝑢 𝑑𝑧 𝑑𝑧 𝑑𝑧
We set 𝑧 = 𝐹(𝑋 + 𝛼𝑌) = 𝐹(𝑢). then𝜕𝑋 = = , 𝑑𝑌 = 𝑎 𝑑𝑢 , and
𝑑𝑢 𝜕𝑋 𝑑𝑢

68
 𝑑𝑧 𝑑𝑧 𝑑𝑧
(𝑑𝑢)2 + 𝑎2 (𝑑𝑢)2 = 𝑧 or √1 + 𝑎2 = 𝑑𝑢
√𝑧

Integrating, 2√1 + 𝑎2 𝑧1/2 = 𝑢 + 𝑏 = 𝑋 + 𝑎𝑌 + 𝑏 = 𝑙𝑛 𝑥 + 𝑎 𝑙𝑛 𝑦 + 𝑏.

A complete solution is 4(1 + 𝑎2 )𝑧 = (𝑙𝑛 𝑥 + 𝑎 𝑙𝑛 𝑦 + 𝑏)2 . The singular solution is 𝑧 = 0.

Example 10.4

Solve4𝑥𝑦𝑧 = 𝑝𝑞 + 2𝑝𝑥 2 𝑦 + 2𝑞𝑥𝑦 2

Solution

1 1 1 1
𝜕𝑧 𝜕𝑧 𝜕𝑋 𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑌 𝜕𝑧
Let 𝑥 = 𝑋 2 , 𝑦 = 𝑌 2 . Then 𝑝 = 𝜕𝑥 = 𝜕𝑋 𝜕𝑥 = 2𝑋 2 𝜕𝑋 and 𝑞 = 𝜕𝑦 = 𝜕𝑌 𝜕𝑦 = 2𝑌 2 𝜕𝑌

𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧
Substituting in the given equation, we have 𝑧 = 𝑋 𝜕𝑋 + 𝑌 𝜕𝑌 + of type II.
𝜕𝑋 𝜕𝑌

A complete solution is 𝑍 = 𝑎𝑋 + 𝑏𝑌 + 𝑎𝑏 or 𝑧 = 𝑎𝑥 2 + 𝑏𝑦 2 + 𝑎𝑏.

Eliminating a and b from this and 0 = 𝑥 2 + 𝑏, 0 = 𝑦 2 + 𝑎. obtained by differentiating it

with respect to a and b, the singular solution is found to be 𝑧 + 𝑥 2 𝑦 2 = 0.

Example 10.5

Solve 𝑝2 𝑥 2 = 𝑧(𝑧 − 𝑞𝑦)

Solution

𝜕𝑧 𝜕𝑧 𝑑𝑋 1 𝜕𝑧 1 𝜕𝑧
The transformation 𝑌 = 𝐼𝑛 𝑦, 𝑋 = 𝐼𝑛 𝑥, 𝑝 = = = 𝑥 𝜕𝑋 , 𝑞 =
𝜕𝑥 𝜕𝑋 𝑑𝑥 𝑦 𝜕𝑌

𝜕𝑧 𝜕𝑧
reduces the given equation to A) (𝜕𝑋)2 = 𝑧(𝑧 − 𝜕𝑌), of Type III.

69
 𝜕𝑧 𝑑𝑧 𝜕𝑧 𝑑𝑧
We set 𝑧 = 𝐹(𝑋 + 𝑎𝑌) = 𝐹(𝑢). Then = 𝑑𝑢,𝜕𝑌 = 𝑎 𝑑𝑢 and A) becomes
𝜕𝑋

𝑑𝑧 2 𝑑𝑧
( ) = 𝑧 2 − 𝑎𝑧
𝑑𝑢 𝑑𝑢

𝑑𝑧 1 𝑑𝑧
Then = 2 𝑧(√𝑎2 + 4 − 𝑎), 2 = (√𝑎2 + 4 − 𝑎)𝑑𝑢 and 𝑙𝑛 𝑧 2 = (√𝑎2 + 4 − 𝑎)(𝑢 + 𝑏).
𝑑𝑢 𝑧

A complete solution is 𝑙𝑛 𝑧 2 = (√𝑎2 + 4 − 𝑎)(𝑙𝑛𝑥 + 𝑎𝑙𝑛𝑦 + 𝑏). There is no singular solution.

Example 10.5

Solve𝑝2 + 𝑞 2 = 𝑧 2 (𝑥 + 𝑦)

Solution

𝑝 2 𝑞 2
The given PDE can be written as (𝑧 ) + ( 𝑧 ) = 𝑥 + 𝑦.

𝜕𝑍 𝜕𝑍
The transformation 𝑧 = 𝑙𝑛 𝑧, 𝑝 = 𝑧 𝜕𝑥 , 𝑞 = 𝑧 reduces the given equation to
𝜕𝑦

𝜕𝑍 2 𝜕𝑍 2 𝜕𝑍 2 𝜕𝑍 2
(𝜕𝑥 ) + (𝜕𝑦) = 𝑥 + 𝑦 or (𝜕𝑥 ) − 𝑥 = 𝑦 − (𝜕𝑦) , of type IV.

𝜕𝑍 2 𝜕𝑍 2 𝜕𝑍 1
𝜕𝑍 1
Set (𝜕𝑥 ) − 𝑥 = 𝑎 = 𝑦 − (𝜕𝑦) . Then𝜕𝑥 = (𝑎 + 𝑥)2 and𝜕𝑦 = (𝑦 − 𝑎)2

1
A complete solution is 𝑍 = ∫ (𝛼 + 𝑥)2 𝑑𝑥 + ∫ (𝑦 − 𝑎)2 𝑑𝑦 + 𝑏

3 3
2 2
Or𝑙𝑛 𝑧 = (𝑎 + 𝑥)2 + (𝑦 − 𝑎)2 + 𝑏.
3 3

70
Etivity 10.2: The method of transformations

Numbering and pacing and 10.2.1 – 10.2.2

sequencing
Title The method of transformations.
Purpose To help you to know how to solve nonlinear PDEs using
the method of transformation.
Brief summary of overall task Watch the video on transformation and solve the given
questions.
Spark 𝑇𝑅𝐴𝑁𝑆𝐹𝑂𝑅𝑀𝐴𝑇𝐼𝑂𝑁 𝑀𝐸𝑇𝐻𝑂𝐷
Individual contribution b) Watch the video on transformation and solve
c) 𝑥 4 𝑝2 − 𝑦𝑧𝑞 − 𝑧 2 = 0

Interaction begins • Post your answers on discussion forum 10.2.3

• Read what your colleagues have posted.
• In a sentence or two, comment on what two of
your colleagues have posted.
E-moderator interventions • Focussing group discussion
• Encouraging quiet ones to contribute
• Providing feedback/ teaching points

• Closing the discussion

Schedule and time This activity should take one hour.
Next Charpit’s method

10.4 Assessment
1 1
1. Use the transformation 𝑋 = 𝑥 , 𝑌 = 𝑦 , 𝑍 = 𝑙𝑛𝑧 to solve 𝑥 4 𝑝2 + 𝑦 2 𝑧𝑞 − 2𝑧 2 = 0

2. Use an appropriate transformation to solve 𝑦𝑧𝑝2 = 𝑞

𝑝2
3. 𝑝2 + 𝑞 2 = 𝑧 2 (𝑥 + 𝑦) or 𝑧 2 +

𝑍 = ln 𝑧

71
(y-x) (qy-px) = (p-q)2 Let( x+y) =X and xy=Y

10.5 References
11 Tyn Myin-U and Lokeath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.
12 Frank Ayres (1952). Shaum’s outline of theory and problems of differential equations.
McGRAW-HILL publishers

COMPATIBILITY OF PDES

https://www.youtube.com/watch?v=3lvBbfWRLVU

https://www.youtube.com/watch?v=7RRpGyibh3A

make p the subject

make q the subject

dz = pdx + qdy separable

z=P + Q +constant

xp –yq = x

xp- y(xz- x2p) =x

p= (1+yz)/1+xy

72
x2p + q = xz

show that x2 –ay =p and q = y2-ax are compatible and hence

solve them.

dz = pdx + qdy separable

dz= (x2 –ay)dx + (y2-ax)dy

dZ = x2dx- aydx + y2dy – axdy

dz = x2dx + y2dy – a(ydx +xdy)

ad(xy)

z= x3/3 +y3/3 – axy +b

p = dz/dx = x2 –ay

q = y2 –ax

exercise

73
1.p= x2+y2 and q= x2-y2

2. p = 5x -7y and q= 6x +8y

[f,g] =0

LESSON ELEVEN

CHARPIT’S METHOD OF SOLVING FIRST ORDER NON-LINEAR PARTIAL

DIFFERENTIAL EQUATIONS
11.1 Introduction

In this lesson we discuss Charpit’s method for solving first order non-linear PDEs. This is the
general method of solving first order non-linear PDEs. Since the solution by this method is
generally complicated, this method is applied to solve equations which cannot be reduced to any
of the above standard forms.
11.2 Learning Outcomes
By the end of this lesson the learner will be able to:
11.2.1 Recognize when to use Charpit’s method to solve non-linear first order PDE;
11.2.2 Solve appropriate first order non-linear PDEs using Charpit’s method.

11.3 Charpit’s Method

Consider the non-linear partial differential equation

74
𝑓(𝑥, 𝑦, 𝑧, 𝑝, 𝑞) = 0. (1)

Since z is a function of x and y, it follows that

𝑑𝑧 = 𝑝 𝑑𝑥 + 𝑞 𝑑𝑦. (2*)

Let us assume 𝑝 = 𝑢(𝑥, 𝑦, 𝑧, 𝑎), where a is an arbitrary constant, substitute in

(1) and solve to obtain 𝑞 = 𝑣(𝑥, 𝑦, 𝑧, 𝑎). For these values of 𝑝 and 𝑞, (2*) becomes

𝑑𝑧 = 𝑢𝑑𝑥 + 𝑣𝑑𝑦. (3*)

Now if (3*) can be integrated, yielding

𝑔(𝑥, 𝑦, 𝑧, 𝑎, 𝑏) = 0, (4)

this is a complete solution of (1).

Example11.1

Solve 𝑝𝑞 + 𝑞𝑥 = 𝑦.

Solution

𝑦
Take 𝑝 = 𝑎 − 𝑥, substitute in 𝑝𝑞 + 𝑞𝑥 = 𝑦, and solve for 𝑞 = 𝑎.

𝑦
Substituting in 𝑑𝑧 = 𝑝 𝑑𝑥 + 𝑞𝑑𝑦, we, have 𝑑𝑧 = (𝑎 − 𝑥)𝑑𝑥 + (𝑎)𝑑𝑦, an integrable equa-

tion, with solution

1 1 𝑦2
𝑧 = 𝑎𝑥 − 𝑥2 + + 𝜅 or 2𝑎𝑧 = 2𝑎2 𝑥 − 𝑎𝑥 2 + 𝑦 2 + 𝑏
2 2 𝑎

75
Since the success of the above procedure depends upon our making a fortunate choice for p, it

cannot be suggested as a standard procedure. We turn now to a general method for solving

𝑓(𝑥, 𝑦, 𝑧, 𝑝, 𝑞) = 0. (1)

This consists in finding an equation

𝐹(𝑥, 𝑦, 𝑧, 𝑝, 𝑞) = 0 (2)

such that (1) and (2) may be solved for 𝑝 = 𝑃(𝑥, 𝑦, 𝑧) and 𝑞 = 𝑄(𝑥, 𝑦, 𝑧 ), (that is, such that

𝜕𝑓 𝜕𝑓
𝜕𝑝 𝜕𝑞
∆= |𝜕𝐹 𝜕𝐹
| ≠ 0, identically), (3)
𝜕𝑝 𝜕𝑞

and such that for these value of p and q the total differential equation

𝑑𝑧 = 𝑝 𝑑𝑥 + 𝑞 𝑑𝑦 = 𝑝(𝑥, 𝑦, 𝑧)𝑑𝑥 + 𝑞(𝑥, 𝑦, 𝑧)𝑑𝑦 (4)

𝜕𝑄 𝜕𝑃 𝜕𝑃 𝜕𝑄 𝜕𝑞 𝜕𝑝
is integrable, that is, 𝑃 = − 𝑄 𝜕𝑧 − + = − = 0
𝜕𝑧 𝜕𝑦 𝜕𝑋 𝜕𝑥 𝜕𝑦

Differentiating (1) and (2) partially with respect to x and y, we find

𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑝 𝜕𝑓 𝜕𝑞
+ 𝑝 𝜕𝑧 + + = 0 (5)
𝜕𝑥 𝜕𝑝 𝜕𝑥 𝜕𝑞 𝜕𝑥

𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑝 𝜕𝑓 𝜕𝑞
+ 𝑝 𝜕𝑧 + + = 0 (6)
𝜕𝑦 𝜕𝑝 𝜕𝑦 𝜕𝑞 𝜕𝑦

𝜕𝐹 𝜕𝐹 𝜕𝐹 𝜕𝑝 𝜕𝐹 𝜕𝑞
+ 𝑝 𝜕𝑧 + + = 0 (7)
𝜕𝑞 𝜕𝑝 𝜕𝑥 𝜕𝑞 𝜕𝑥

𝜕𝐹 𝜕𝐹 𝜕𝐹 𝜕𝑝 𝜕𝐹 𝜕𝑞
+ 𝑝 𝜕𝑧 + + = 0 (8)
𝜕𝑦 𝜕𝑝 𝜕𝑦 𝜕𝑞 𝜕𝑦

76
 𝜕𝐹 𝜕𝐹 𝜕𝑓 𝜕𝑓
Multiplying (5) by𝜕𝑝 , (6) by , (7) by − 𝜕𝑝 , (8) by − and adding, we obtain (noting
𝜕𝑞 𝜕𝑞

𝜕𝑝 𝜕𝑞
that 𝜕𝑦 = 𝜕𝑥 )

𝜕𝑓 𝜕𝑓 𝜕𝐹 𝜕𝑓 𝜕𝑓 𝜕𝐹 𝜕𝑓 𝜕𝑝 𝜕𝑓 𝜕𝐹 𝜕𝑓 𝜕𝑓 𝜕𝐹
( + 𝑝 ) + ( + 𝑞 ) − − − (𝑝 + 𝑞 )
𝜕𝑥 𝜕𝑧 𝜕𝑝 𝜕𝑦 𝜕𝑧 𝜕𝑞 𝜕𝑝 𝜕𝑥 𝜕𝑞 𝜕𝑦 𝜕𝑝 𝜕𝑞 𝜕𝑧

= 0

This is a linear partial differential equation in F, considered as a function of the independent

variables 𝑥, 𝑦, 𝑧, 𝑝, 𝑞. The auxiliary system is

𝑑𝑝 𝑑𝑞 𝑑𝑥 𝑑𝑦 𝑑𝑧 𝑑𝐹
𝜕𝑓 𝜕𝑓 = (𝜕𝑓/𝜕𝑦+ 𝑞 𝜕𝑓/𝜕𝑧)
= 𝜕𝑓 = 𝜕𝑓 = 𝜕𝑓 𝜕𝑓 = . (9)
( +𝑝 ) − − − (𝑝 + 𝑞 ) 0
𝜕𝑥 𝜕𝑧 𝜕𝑝 𝜕𝑞 𝜕𝑝 𝜕𝑞

Thus, we may take for (1) any solution of this system which involves 𝑝 or q, or both, which

contains an arbitrary constant, and for which (2) holds.

Example 11.2

Solve 𝑞 = − 𝑥𝑝 + 𝑝2

Solution

𝜕𝑓 𝜕𝑓 𝜕𝑓
Here ƒ = 𝑝2 − 𝑥𝑝 − 𝑞so that𝜕𝑥 = −𝑝 𝜕𝑦 = 0, =0
𝜕𝑧

𝜕𝑓 𝜕𝑓
= 2𝑝 − 𝑥, = −1
𝜕𝑝 𝜕𝑞

𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
+ 𝑝 𝜕𝑧 = −𝑝, + 𝑞 𝜕𝑧 = 0, − (𝑝 𝜕𝑝 + 𝑞 𝜕𝑞) = −2𝑝2 + 𝑥𝑝 + 𝑞.
𝜕𝑥 𝜕𝑦

77
 𝑑𝑝 𝑑𝑞 𝑑𝑥 𝑑𝑦 𝑑𝑧
The auxiliary system is = = −2𝑝+𝑥 = = −2𝑝2+𝑥𝑝+𝑞
−𝑝 0 1

𝑑𝑝 𝑑𝑦
From −𝑝 = we have 𝑙𝑛 𝑝 = −𝑦 + 𝑙𝑛 𝛼 or 𝑝 = 𝛼𝑒 −𝑦
1

Using the given differential equation, 𝑞 = −𝑥𝑝 + 𝑝2 = −𝛼𝑒 −𝑦 + 𝛼 2 𝑒 −2𝑦

Then 𝑑𝑧 = 𝑝 𝑑𝑥 + 𝑞𝑑𝑦 becomes 𝑑𝑧 = 𝑎𝑒 −𝑦 𝑑𝑥 + (−𝑎𝑥𝑒 −𝑦 + 𝛼 2 𝑒 −2𝑦 )𝑑𝑦.

Integrating,

1
𝑧 = 𝑎𝑥𝑒 −𝑦 − 𝑎2 𝑒 −2𝑦 + 𝑏
2

There is no singular solution.

Example 11.3

Solve 16𝑝2 𝑧 2 + 9𝑞 2 𝑧 2 + 4𝑧 2 − 4 = 0

Solution

Let 𝑓(𝑥, 𝑦, 𝑧, 𝑝, 𝑞) = 16𝑝2 𝑧 2 + 9𝑞 2 𝑧 2 + 4𝑧 2 − 4

𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
Then = 0 = 𝜕𝑦 , = 32𝑝2 𝑧 + 18𝑞 2 𝑧 + 8𝑧, 𝜕𝑝 = 32𝑝𝑧 2 , 𝜕𝑞 = 18𝑝𝑧 2 and the auxiliary
𝜕𝑥 𝜕𝑧

𝑑𝑝 𝑑𝑞 𝑑𝑥 𝑑𝑦 𝑑𝑧
system 𝜕𝑓 𝜕𝑓 = 𝜕𝑓 𝜕𝑓 = 𝜕𝑓 = 𝜕𝑓 = 𝜕𝑓 𝜕𝑓 is
+𝑝 +𝑞 − − −(𝑝 + q
𝜕𝑥 𝜕𝑧 𝜕𝑦 𝜕𝑧 𝜕𝑝 𝜕𝑞 𝜕𝑝 𝜕𝑞

Using the multipliers 4𝑧, 0, 𝐼, 0, 4𝑝, we find

4𝑧(32𝑝3 𝑧 + 𝑙8𝑝𝑞 2 𝑧 + 8𝑝𝑧) + 1(_32𝑝𝑧 2 ) + 4𝑝(−32𝑝2 𝑧 2 – 𝑙8𝑞 2 𝑧 2 ) = 0

and so 𝑑𝑥 + 4𝑝 𝑑𝑧 + 4𝑧 𝑑𝑝 = 0.

Then 𝑥 + 4𝑝𝑧 = 𝑎 and

78
Substituting for p in the given differential equation, we find

2 1
(𝑥 − 𝑎)2 + 9𝑞 2 𝑧 2 + 4𝑧 2 − 4 = 0. Using the root 𝑞 = 3𝑧 √1 − 𝑧 2 − 4 (𝑥 − 𝑎)2,

1
𝑥−𝑎 2 1 3[𝑧 𝑑𝑧+ (𝑥−𝑎)𝑑𝑥]
𝑑𝑧 = 𝑝 𝑑𝑥 + 𝑞 𝑑𝑦 = − 𝑑𝑥 + 3𝑧 √1 − 𝑧 2 − 4 (𝑥 − 𝑎)2 𝑑𝑦 or 𝑑𝑦 = 4
1
4𝑧 2√1−𝑧 2 − (𝑥−𝑎)2
4

3 1 (𝑥−𝑎)2 (𝑦−𝑏)2
Then 𝑦 − 𝑏 = − 2 √√1 − 𝑧 2 − 4 (𝑥 − 𝑎)2 𝑜𝑟 + + 𝑧2 = 1 is complete
4 9/4

solution. This is a family of ellipsoids with centers on the 𝑥𝑦 − plane. The semi-axes of the

ellipsoids are 2 units parallel to the 𝑥 −axis, 3/2 units parallel to the 𝑦 −axis, and 1 unit parallel

to the 𝑧 −axis. The singular solution consists of the parallel planes 𝑧 = ±1.

Another complete solution may be found by noting that the equation is of Type III. Using

𝑑𝑧 𝑑𝑧
𝐹(𝑥 + 𝑎𝑦) = 𝐹(𝑢) and setting 𝑝 = and 𝑞 = 𝛼 , the given equation-
𝑑𝑢 𝑑𝑢

becomes

𝑑𝑧 2 𝑑𝑧 2 𝑧 𝑑𝑧 2
16𝑧 ( ) + 9𝑎 𝑧 ( ) + 4𝑧 2 − 4 = 0 𝑜𝑟
2 2 2
= 𝑑𝑢
𝑑𝑢 𝑑𝑢 √1 − 𝑧 2 √16 + 9𝑎2

2 2
-√1 − 𝑧 2 = √16+9𝑎2 (𝑢 + 𝑏) = √16+9𝑎2
(x+αy+b).

This complete solution (16 + 9𝑎2 ) (1 − 𝑧 2 ) = 4 (𝑥 + 𝑎𝑦 + 6)2 represents a family of

elliptic cylinders with elements parallel to the 𝑥𝑦- plane. The major axis of a cross section lies in

the 𝑥𝑦- plane and the minor axis is 2 units parallel to the 𝑧 −axis.

79
Etivity 11.2: Charpit’s method

80
Numbering and pacing and 11.2.1 – 11.2.2
sequencing
Title Charpit’s Method
Purpose To help you to know how to solve ODEs that are
variable separable and those that can be reduced to
variable separable form.
Brief summary of overall task Watch the video on Charpit’s by Khan Academy and
solve the given questions.
Spark 𝐶𝐻𝐴𝑅𝑃𝐼𝑇 ′ 𝑆 𝑀𝐸𝑇𝐻𝑂𝐷
Individual contribution d) Watch the video on Charpits and solve
e) (𝑝2 + 𝑞 2 )𝑥 = 𝑧𝑝

Interaction begins • Post your answers on discussion forum 11.2.3

• Read what your colleagues have posted.
• In a sentence or two, comment on what two of
your colleagues have posted.
E-moderator interventions • Focussing group discussion
• Encouraging quiet ones to contribute
• Providing feedback/ teaching points

• Closing the discussion

Schedule and time This activity should take one hour.Type equation here.
Next End

10.4 Assessment
Apply Charpit’s method to find the complete integral (and singular solution, if it exists) of the
following equations
a) 𝑞 = (𝑧 + 𝑝𝑥)2
b) 𝑧 2 (𝑝2 𝑧 2 + 𝑞 2 ) = 1
c) 𝑧 = 𝑝𝑞
d) 𝑞 = 3𝑝2
e) 2𝑥𝑧𝑝 + 𝑥 2 = 𝑧 2 𝑞 2 + 𝑧 2 12mks
f) 2𝑥𝑧 − 𝑝𝑥 2 − 2𝑞𝑥𝑦 + 𝑝𝑞 = 0
10.5 References
1. Tyn Myin-U and Lokeath Debnath (1992). Linear Partial Differential Equations for
Scientists and Engineers, 4th Ed. Birkhauser publishers.
2. Frank Ayres (1952). Shaum’s outline of theory and problems of differential equations.
McGRAW-HILL publishers

81