Chapter 2

First-Order Partial Differential Equations

BY

Prof. D. C. Sanyal

Retired Professor Of Mathematics

University Of Kalyani
West Bengal, India
E-mail : dcs [email protected]

1
Module-3: Nonlinear First-Order Partial Differential
Equations

1. Introduction
In this module, we discuss nonlinear partial differential equations of first-order of
the form

F(x, y, z, p, q) = 0 (1)

∂z ∂z
where x, y are independent variables, z = z(x, y) and p = ∂x
, q= ∂y
, the function F
being nonlinear. We have already seen in previous study that the partial differential
equation arising from the two-parameter system of surfaces

f (x, y, z, a, b) = 0 (2)

is of this form. We shall see later that the converse is also true.
Let us define first different types of solutions involved for the equation (1) and
consider a geometrical approach, introduced by Cauchy, to find the solutions.

2. Definitions of Various Types of Solutions

Any envelope of the system (2) touches at each of its points a member of the system
and, therefore, it has the same set of values (x, y, z, p, q) as the particular surface. Thus,
it is also a solution of the partial differential equation (1). Hence, we are led to the
following three types of integrals of the equation (1).

(a) Complete integral

The two-parameter system of surfaces of the form (2) is called a complete integral
of the equation (1).

(b) General integral

Suppose there exists a relation between the parameters a and b of the form b =
ψ(a), ψ being arbitrary. Then the one-parameter subsystem f (x, y, z, a, ψ(a)) = 0 of
the system (2) forms its envelope and is called the general integral of (1).

2
Chapter 2 First-Order Partial Differential Equations

(c) Singular integral

If the envelope of the two-parameter system of surfaces (2) exists, then it is also a
solution of the equation (1) and is known as the singular integral.

Example 1: Verify that z = ax + by + a + b − ab is a complete integral of the partial

differential equation z = px + qy + p + q − pq, a and b being arbitrary constants.
Show also that the envelope of all planes corresponding to complete integrals pro-
vided singular integral of the differential equation and determine a general integral
by finding the envelope of these planes that pass through the origin.

Solution: Let f (x, y, z, a, b) = z − (ax + by + a + b − ab) = 0 so that p = ∂z

∂x
= a, q = ∂z
∂y
= b and
hence f (x, y, z, a, b) = 0 is a complete integral of the equation z = px + qy + p + q − pq.
Now, the envelope of the two-parameter system f (x, y, z, a, b) = 0 is obtained by
∂f ∂f
eliminating a and b from the equations f = 0, ∂a
= 0, ∂b
= 0, i.e. from

z = ax + by + a + b − ab, −(x + 1) + b = 0, −(y + 1) + a = 0.

Elimination of a and b from these equations gives z = (x+1)(y +1) which is the singular
integral.
To find the general integral, we suppose that there exists a relation of the form
b = ψ(a) between the parameters a and b. So, we consider the one-parameter system

f (x, y, z, a, ψ(a)) = z − {ax + ψ(a)y + a + ψ(a) − aψ(a)} = 0.

∂f { }
Thus = − x + ψ ′ (a)y + 1 + ψ ′ (a) − ψ(a) − aψ ′ (a) = 0
∂a
Since the envelope passes through the origin, so from equation f = 0, we get
a 1
−a − ψ(a) + aψ(a) = 0 ⇒ ψ(a) = ⇒ ψ ′ (a) = −
a−1 (a − 1)2
√
∂f y
Then the equation ∂a = 0 gives x − 1
(a−1)2
y +1− 1
(a−1)2
− a
a−1 + a
(a−1)2
=0⇒a= x +1
√ √
y+ x
so that ψ(a) = √y .
Thus, from the equation f = 0, we get by substituting the values of a and ψ(a)
√ √ √ √ √ √ √ √ √
y+ x y+ x y+ x y y+ x
z= √ ·x+ √ ·y + √ −√ · √
y y y x y
√
⇒ z = 2 xy + x + y, i.e. (x + y − z)2 = 4xy.

This is the required general integral.

3
 Chapter 2 First-Order Partial Differential Equations

3. Cauchy’s Method of Characteristics

We now introduce a method, due to Cauchy, based on geometrical idea, to solve

nonlinear partial differential of the form (1).
Suppose the plane through the point P (x0 , y0 , z0 ) with its normal parallel to the di-
rection n̂ with direction ratios (p0 , q0 , −1) be uniquely specified by the set of numbers
(x0 , y0 , z0 , p0 , q0 ). Conversely, any set of five real numbers defines a plane in three-
dimensional space. A plane element (x0 , y0 , z0 , p0 , q0 ) satisfying the equation (1) is
called an integral element.
We rewrite equation (1) in the form

q = G(x, y, z, p) (3)

and keep x, y, z fixed but vary p so that we obtain a set of plane elements {x0 , y0 , z0 , p,
G(x0 , y0 , z0 , p)} depending on the single parameter p only and passes through the point
P . Thus the planar elements envelope a cone with P as vertex, called the elementary
cone of the equation (1) at the point P .

Fig. 1

Now consider a surface S given by the equation

z = g(x, y) (4)

where the function g(x, y) and its first partial derivatives with respect to x and y are as-
sumed to be continuous in a region Ω of the xy-plane. Then the tangent plane at every
{ }
point of S determines a plane element of the type x0 , y0 , g(x0 , y0 ), gx (x0 , y0 ), gy (x0 , y0 )

4
Chapter 2 First-Order Partial Differential Equations

which is called the tangent element of the surface S at the point (x0 , y0 , g(x0 , y0 )). Thus,
we have the result:

Theorem 1: A necessary and sufficient condition for a surface to be an integral surface

of a partial differential equation is that at each point its tangent element should touch the
elementary cone.

Cauchy characteristic equations

Consider a curve Γ with parametric equations x = x(t), y = y(t), z = z(t) lying on the
surface S given by equation (4) so that

z = g{x(t), y(t)}, ∀ t ∈ I, (5)

where I is the given interval. Then, if P0 is a point on Γ determined by the parameter

t0 . the direction ratios of the tangent line P0 P1 are {x′ (t0 ), y ′ (t0 ), z′ (t0 )}, where x′ (t0 ) =
( )
dx
dt etc. This direction is perpendicular to the direction (p0 , q0 , −1) provided
t=t0

p0 x′ (t0 ) + q0 y ′ (t0 ) + (−1)z′ (t0 ) = 0, i.e z′ (t0 ) = p0 x′ (t0 ) + q0 y ′ (t0 ).

Fig. 2

Thus, any set {x(t), y(t), z(t), p(t), q(t)} of five real functions satisfying the condition

z′ (t) = p(t)x′ (t) + q(t)y ′ (t) (6)

defines a strip at the point (x, y, z) of the curve Γ . If such a strip is an integral element
of the equation (1), viz. F(x, y, z, p, q) = 0, then it is called an integral strip of the equa-
tion. Thus the set of functions {x(t), y(t), z(t), p(t), q(t)} is an integral strip of this partial

5
 Chapter 2 First-Order Partial Differential Equations

differential equation (1), if they satisfy the condition (6) and the condition

F {x(t), y(t), z(t), p(t), q(t)} = 0 ∀ t ∈ I. (7)

If at each point the curve Γ touches a generator of the elementary cone, then the cor-
responding strip is called a characteristic strip. The point (x + dx, y + dy, z + dz) lies on
the tangent plane to the elementary cone if

dz = pdx + qdy (8)

where p, q satisfy the equation (1). Now differentiating (1) and (8) with respect to p,
we get respectively

∂F ∂F ∂q dq
+ = 0 and0 = dx + dy. (9)
∂p ∂q ∂p dp

Using relations (8) and (9), we have

dx dy dz
= = . (10)
Fp Fq pFp + qFq

Thus, x′ (t), y ′ (t), z′ (t) are proportional to Fp , Fq , pFp + qFq respectively along a charac-
teristic strip. We choose the parameter t such that

x′ (t) = Fp , y ′ (t) = Fq , z′ (t) = pFp + qFq .

Now along a characteristic strip p = p(t), a function of t and, therefore,

∂p ′ ∂p ′ ∂p ∂F ∂p ∂F ∂p ∂F ∂q ∂F
p′ (t) = x (t) + y (t) = + = + ,
∂x ∂y ∂x ∂p ∂y ∂q ∂x ∂p ∂x ∂q
∂p ∂2 z ∂2 z ∂q
where we used the result ∂y
= ∂y∂x
= ∂x∂y
= ∂x
. Also, differentiation of (7) partially
with respect to x gives

∂F ∂F ∂F ∂p ∂F ∂q
+p + + =0
∂x ∂z ∂p ∂x ∂q ∂x
∂F ∂F
i.e. +p + p′ (t) = 0
∂x ∂z

so that on a characteristics strip

p′ (t) = −(Fx + pFz ).

Similarly, we have q′ (t) = −(Fy + qFz ).

6
Chapter 2 First-Order Partial Differential Equations

Thus, we have the following system of five ordinary differential equations for the
determination of the characteristic strip:

x′ (t) = Fp , y ′ (t) = Fq , z′ (t) = pFp + qFq ,

(11)
p′ (t) = −(Fx + pFz ), q′ (t) = −(Fy + qFz ).

Equations (11) are known as Cauchy’s characteristic equations of the partial differential
equation F(x, y, z, p, q) = 0.

Theorem 2: The function F(x, y, z, p, q) is constant along every characteristic strip of the
partial differential equation F(x, y, z, p, q) = 0.

Proof: Along a characteristic strip

[ }]
d {
F x(t), y(t), z(t), p(t), q(t)
dt
= Fx x′ (t) + Fy y ′ (t) + Fz z′ (t) + Fp p′ (t) + Fq q′ (t)

= Fx Fp + Fy Fq + Fz (pFp + qFq ) − Fp (Fx + pFz ) − Fq (Fy + qFz )

=0

so that F(x, y, z, p, q) is constant along characteristic strip.

Corollary: If a characteristic strip contains at least one integral element of F(x, y, z, p, q) =

0, then it is an integral strip of this equation.

We are now in a position to solve Cauchy problem stated earlier.

Let us find the solution of the equation (1), viz. F(x, y, z, p, q) = 0 such that the integral
surface passes through the curve Γ having parametric equations

x = ϕ(ξ), y = ψ(ξ), z = χ(ξ). (12)

Then, in the solution

x = x(x0 , y0 , z0 , p0 , q0 , t0 , t) etc (13)

of the characteristic equations (11), we may take the initial values of x, y, z as x0 = ϕ(ξ),
y0 = ψ(ξ), z0 = χ(ξ) and so the corresponding values of p0 , q0 are given by the relations

χ′ (ξ) = p0 ϕ′ (ξ) + q0 ψ ′ (ξ) and F{ϕ(ξ), ψ(ξ), χ(ξ), p0 , q0 } = 0

7
 Chapter 2 First-Order Partial Differential Equations

Substituting these values of x0 , y0 , z0 , p0 , q0 and the appropriate value of t0 in equa-

tion (13), x, y and z can be expressed in terms of two parameters ξ and t in the form

x = X(ξ, t), y = Y (ξ, t) z = Z(ξ, t).

Elimination of ξ and t amongst these relations leads to an equation of the form

Θ(x, y, z) = 0 (14)

This is the equation of the integral surface of the equation F(x, y, z, p, q) = 0 through
the curve Γ .

Example 2: Find the characteristics of the equation pq = z and determine the integral
surface which passes through the parabola x = 0, y 2 = z.

Solution: Here F(x, y, z, p, q) = pq − z = 0 and so the characteristic equations are

x′ (t) = Fp = q, y ′ (t) = Fq = p, z′ (t) = pFp + qFq = 2pq,

p′ (t) = −(Fx + pFz ) = p, q′ (t) = −(Fy + qFz ) = q

Since the equation of the given curve is x = 0, y 2 = z, so we choose the initial values as
x0 = 0, y0 = ξ z0 = ξ 2 .
From the equation z0′ = p0 x0′ + q0 y0′ , we get q0 = 2ξ and the given equation pq = z
provides p0 = ξ/2.
Now, the equations x′ (t) = q and q′ (t) = q on integration, give x = q + c1 , while the
equations y ′ (t) = p and p′ (t) = p, on integration, lead to y = p + c2 , where c1 and c2
are constants. Using the initial conditions, we get 0 = 2ξ + c1 and ξ = ξ/2 + c2 so that
c1 = −2ξ and c2 = ξ/2. Hence, it follows that

x = q − 2ξ, y = p + ξ/2.

Again, the equations p′ (t) = p and q′ (t) = q, on integration, give p = c3 et and q = c4 et ,

where c3 and c4 are constants. Using the initial conditions, we have ξ/2 = c3 , 2ξ = c4 .
Thus p = ξ2 et , q = 2ξet . Hence,
ξ t 4y + x 1
x = 2ξ(et − 1), y = (e + 1) ⇒ et = , ξ = (4y − x).
2 4y − x 4
Also, the characteristic equation z′ (t) = 2pq = 2ξ 2 e2t gives on integration z = ξ 2 e2t ,
where we have used the initial condition z0 = ξ 2 at t = 0. Thus
(4y − x)2 (4y + x)2
z= · , i.e 16z = (x + 4y)2
16 (4y − x)2

8
Chapter 2 First-Order Partial Differential Equations

Hence the characteristics of the given partial differential equation are

ξ t
x = 2ξ(et − 1), y = (e + 1), z = ξ 2 e2t
2

and the equation of the required integral surface is 16z = (x + 4y)2