FIRST ORDER NON LINEAR PARTIAL DIFFERENTIAL EQUATIONS

COMPATIBLE SYSTEM OF PARTIAL DIFFERENTIAL

EQUATIONS OR ORDER ONE
Consider the following partial differential equation of order one
f ( x , y , z , p ,q )=0
and g( x , y , z , p , q )=0
This system of equations is said to be compatible if every solution of
one is also a solution of the other.
Condition for Compatibility:
The given equation are f ( x , y , z , p ,q )=0 …(1)
and g( x , y , z , p , q )=0 …(2)

∂(f ,g) | ∂f ∂f ¿|¿

J= =¿ ∂ p ∂q ¿¿
∂( p,q)
Let ¿ …(3)
Here, J denotes Jacobian of f and g.
 It is possible to solve (1) and (2) for p and q.
Let the solution be f =φ ( x , y , z ), q=ψ ( x , y , z )
The condition that (1) and (2) should be compatible is that
dz= p dx +q dy or φ dx +ψ dy−dz=0
should be integrable.
Now, is integrable if P dx + Q dy + R dz = 0 is integrable if

P( ∂Q ∂ R
−
∂z ∂ y ) (
+R
∂R ∂P
−
∂x ∂z ) (
+R
∂ P ∂Q
−
∂ y ∂x
=0 )
Therefore, (4) is integrable if
φ ( ∂ψ
∂z ) (
−0 +ψ 0−
∂φ
∂z )+(−1) ( ∂ φ ∂ψ
−
∂ y ∂z )
=0

∂ψ ∂ ψ ∂φ ∂φ
+φ = +ψ
or ∂x ∂z ∂ y ∂z …(5)
Differentiating (1) partially w.r.t. x and z after having substituted the
value of p and q, we get
∂ f ∂ f ∂ φ ∂ f ∂ψ
+ . + . =0
∂x ∂ p ∂ x ∂q ∂x …(6)
 ∂ f ∂ f ∂ φ ∂ f ∂ψ
+ . + . =0
and ∂z ∂ p ∂z ∂q ∂ z …(7)

Multiplying (7) by f and adding to (6), we have

∂f
∂x
+φ
∂f ∂f ∂φ
+ (
∂ z ∂q ∂ x
+φ + ) (
∂φ ∂ f ∂ψ
∂ z ∂q ∂x
+ψ
∂ψ
∂z
=0 ) …(8)
Similarly, differentiating (2) w.r.t. x and z, we get
∂g
∂x
+φ
∂g ∂ g ∂φ
+ (
∂ z ∂q ∂ x
+φ + ) (
∂φ ∂ g ∂ψ
∂ z ∂q ∂ x
+ψ
∂ψ
∂z
=0 ) …(9)
∂φ ∂ φ ∂g
+φ
Eliminating ∂ x ∂ z from (8) and (9) by multiplying (8) by ∂ p and
∂f
(9) by ∂ p and then subtracting, we have

(
∂f ∂g ∂g ∂f
. − .
∂x ∂ p ∂x ∂ p ) (
+φ
∂ f ∂ g ∂ g ∂f
. − . +)( . − .
∂z ∂ p ∂z ∂ p ∂q ∂ p ∂q ∂ p ∂ x )(
∂ f ∂ g ∂ g ∂ f ∂ ψ ∂ψ
+φ
∂z )
=0

∂(f , g )
+φ − (
∂(f , g) ∂( f , g ) ∂ ψ
or ∂(x , p ) ∂( z , p ) ∂( p , q ) ∂ x
+φ
∂ψ
∂z )
=0
or J
(∂ψ
∂ x
+φ
∂ψ
∂ z
=)∂( f , g )
∂( x , p )
+φ
∂( f , g )
∂( z , p)

or
∂ψ
∂x
+φ
∂ ψ 1 ∂ (f , g)
=
∂ z J ∂( x , p){+φ
∂( f , g )
∂( z , p ) }
…(10)
Similarly, differentiating (1) and (2) w.r.t. y and z and proceeding in
the similar manner as above, we get
∂φ
∂y
+ψ =
{
∂ φ 1 ∂(f , g)
∂ z J ∂( y , q)
+ψ
∂( f , g )
∂( z ,q ) }…(11)
Substituting in (5) from (10) and (11) and replacing f, y by p, q
respectively, we get

{
1 ∂( f , g)
J ∂( y , p )
+p
} {
∂( f , g ) 1 ∂( f , g)
=
∂( z , q ) J ∂ ( y ,q )
+q
∂( f , g )
∂( z , q ) }
∂( f , g ) ∂( f , g ) ∂(f , g ) ∂( f , g )
+p + +q =0
or ∂( y , p) ∂( z , q) ∂( y , q ) ∂(z , q ) or [f, g] = 0
which is the required condition of compatibility.
Special Case:
 The first order partial differential equations p = P(x, y) and
∂ P ∂Q
=
q = Q(x, y) are compatible if and only if ∂ y ∂ x .
Proof. The given partial differential equations are
∂z
p = P (x, y) or ∂ x = P (x, y) …(1)
∂z
and q = Q (x, y) or ∂ y = Q (x, y) …(2)
Since P and Q are functions of two variables x and y, therefore P dx +
∂ P ∂Q
=
Q dy is an exact differential if and only if ∂ y ∂ x .
∂ P ∂Q
=
This implies dz = P dx + Q dy is integrable if and only if ∂ y ∂ x .
Now, (1) and (2) are compatible if and only if dz = P dx + Q dy is
integrable.
∂ P ∂Q
=
Hence (1) and (2) are compatible if and only if ∂ y ∂ x .
 ∂ P ∂Q
=
In other words, if ∂ y ∂ x , then (1) and (2) being compatible, have a
∂ P ∂Q
≠
common solution. If ∂ y ∂ x then (1) and (2) are not compatible and
hence these equations have no solution.
SOLVED EXAMPLES:
Example 1. Solve that the equations
xp – yp = x, x2p + q = xz are compatible and find their solution.
Solution. Let f(x, y, z, p, q) = xp – yq – x = 0 …(1)
and g(x, y, z, p, q) = x2p – q – xz = 0 …(2)

∂(f , g ) ∂f ∂f
|
=¿ ∂ x ¿|¿ ¿ ¿
∂(x , p ) ∂p
\ ¿
∂ ( f , g ) |∂ f ∂f
¿|¿ ¿ ¿
=¿ ∂ z ∂p
∂ (z , p )
¿
 ∂( f ,g) | ∂ f ∂ f ¿|¿
=¿ ∂ y ∂ q ¿¿
∂( y,q)
¿
∂(f , g ) ∂f ∂f
|
=¿ ∂ z ∂ q ¿|¿ ¿¿
∂( z, q )
¿
∂(f , g ) ∂( f , g ) ∂( f , g ) ∂( f , g )
[ f , g ]= +p + +q
∂( x , p ) ∂( z , p) ∂( y , q ) ∂( z , q )
= (p – 1) x2 – x (2xp – z) + px2 – q + q (–xy)
= p x2 – x2 – 2x2p + xz + px2 – q –xyq
= – x2 + xz – q – xyq
= – x2 + x2p – xyq [∵ From (2), xz – q = x2p]
= x [– x + xp – yq]
= x.(0) = 0 [By (1)]
Since [f, g] = 0, therefore (1) and (2) are compatible.
To find common solution of (1) and (2):
 Let us solve (1) and (2) for p and q.
Multiplying (2) by y and adding to (1), we get
xp + x2yp – x (1+yz) = 0
1+ yz
p=
⇒ 1+ xy
From (2), we have q = xz – x2p
=xz −x 2
1+ yz
1+ xy ( ) [Using the value of p]
x ( z−x )
=
1+xy
Substituting the values of p and q in dz = p dx + q dy, we get

dz= ( 1+ yz
1+ xy ) (
dx+
1+ xy )
x ( z −x )
dy

or (1 + xy) dz = (1 + yz) dx + x (z – x) dy
= z (y dx + x dy) + dx – x2 dy
= z d(xy) + dx – x2 dy
or (1 + xy) dz – z d (xy) = dx – x2dy
 (1+xy )dz−z d ( xy ) dx−x 2 dy
2
=
or (1+ xy ) (1+ xy )2

( )
1
2
dx−dy
(
−d y +
1
)
d ( z
= ) x
=
x
1+ xy
( ) ( )
2 2
1 1
y+ y+
or x x
Integrating, we have
z 1
= +c
1+ xy 1
y+
x , where c is an arbitrary constant
z x1
= +c
or 1+ xy 1+ xy
or z – x = c (1 + xy),
which is the required common solution of (1) and (2).

Example 2. Show that the equation z = px + qy is compatible with

any equation f (x, y, z, p, q) = 0 which is homogeneous in x, y, z.
Solution. The given partial differential equation is z = px + qy
 Let g (x, y, z, p, q) = px + qy – z = 0 …(1)
Also, given another partial differential equation homogeneous in x,y,z as
f (x, y, z, p, q) = 0 …(2)
If (2) is of degree n, then by Euler’s theorem on homogenous
function, we have

x ( ) ( ) ( )
∂f
∂x
+y
∂f
∂y
+z
∂f
∂z
=n . f

x ( )+ y ( ) ( )
∂f ∂f ∂f
+ z =0
⇒ ∂x ∂y ∂z …(3) [Using (2)]
We know that (1) and (2) compatible if
∂( f , g ) ∂( f , g ) ∂( f , g ) ∂( f , g )
[ f , g ]= +p + +q =0
∂( x , p ) ∂( z , p) ∂( y , q ) ∂( z , q ) …(4)

∂(f , g ) | ∂ f ∂ f ¿|¿
=¿ ∂ x ∂ p ¿ ¿
∂(x , p )
Now, ¿
 ∂( f , g ) | ∂ f ∂f ¿|¿
=¿ ∂ z ∂ p ¿ ¿
∂(z , p )
¿
∂( f , g ) | ∂ f ∂ f ¿|¿
=¿ ∂ y ∂ q ¿ ¿
∂( y , q )
¿
∂(f , g ) | ∂ f ∂ f ¿|¿
=¿ ∂ z ∂ q ¿ ¿
∂( z, q )
¿
Substituting these in (4), we get
∂f ∂f
[ f , g ]=x − p + p
∂x ∂p ( ∂f
∂z )
−p
∂f
∂p
∂f
∂y
∂f
∂q ( ∂f ∂f
+ y −q +q y +
∂ z ∂q )
∂f ∂f ∂f
=x + y +( px+qy )
∂x ∂ y ∂z
∂f ∂f ∂f
=x + y +z
∂x ∂ y ∂z [Using (1)]
 =0 [Using (3)]
Hence (1) and (2) are compatible i.e., the differential equation z = px
+ qy is compatible with any differential equation f(x, y, z, p, q) = 0 which
is homogeneous in x, y, z.
Example 3. Show that the equation f(x, y, p, q) = 0, g(x, y, p, q) = 0
∂( f , g ) ∂( f , g )
+ =0
are compatible if ∂(x , p ) ∂( y , q ) .
Solution. The given equations are
f(x, y, p, q) =0 …(1)
and g(x, y, p, q) = 0 …(2)
The condition of compatibility of the equations
f(x, y, z, p, q) =0
and g(x, y, z, p, q) = 0 is
∂( f , g ) ∂( f , g ) ∂( f , g ) ∂( f , g )
[ f , g ]= +p + +q =0
∂( x , p ) ∂( z , p) ∂( y , q ) ∂( z , q )
…(3)
If we compare (1) and (2) with the equations f(x, y, z, p, q) =0 and
g(x, y, z, p, q) = 0, we observe that z is missing in (1) and (2)
 ∂f ∂g
=0 and =0
So, ∂z ∂z

∂( f , g) |∂ f ∂f ¿|¿
=¿ ∂ z ∂ p ¿¿
∂(z , p)
¿
∂(f ,g) |∂ f ∂ f ¿|¿
=¿ ∂ z ∂ q ¿¿
∂( z,q)
¿
Substituting these in (3), we get
∂(f , g ) ∂( f , g )
+ =0
∂(x , p ) ∂( y , q ) , which is the required condition.
Example 4. Examine whether the following system of partial
differential equations are compatible or not? If compatible, find their
common solution.
 ∂z ∂z
(i) =7 x+18 y−1, =9x+11 y−2
∂x ∂y
(i) p=6 x+3 y, q=3 x−4 y
(iii) p=ax+hy+g, q=hx +by +f .

Solution. (i) The given system of partial differential equations are

∂z
= p=7 x +18 y−1
∂x …(1)
∂z
=q=9 x +11 y−2
and ∂y …(2)
We know that the system of equations
p = P(x, y), q = Q(x, y)
∂ P ∂Q
=
are compatible if and only if ∂ y ∂ x
Here P = 7x + 18y – 1 and Q = 9x + 11y – 2
 ∂P ∂Q
=18 =9
\ ∂y and ∂ y
∂ P ∂Q
≠
Since ∂ y ∂ x , therefore equations (1) and (2) are not compatible.

(ii) The given system of partial differential equations are

p = 6x + 3y …(1)
and q = 3x – 4y …(2)
We know that the system of equations
p = P(x, y), q = Q (x, y)
∂ P ∂Q
=
are compatible if and only if ∂ y ∂ x
Here P = 6x + 3y, Q = 3x – 4y
∂P ∂Q
=3 , =3
\ ∂y ∂x
 ∂ P ∂Q
= =3
Since ∂ y ∂ x , therefore equations (1) and (2) are compatible.
Now, we shall find common solution of (1) and (2):
Substituting the values of p and q in dz = pdx + qdy, we get
dz = (6x + 3y) dx + (3x – 4y) dy …(3)
= Mdx + Ndy, where M = 6x + 3y, N = 3x – 4y

Here, R.H.S. of (3) is an exact differential, therefore integrating (3),

we have
z= ∫ M dx +∫ { terms in N not containing x } dy + c
y cons tan t

= ∫ (6 x+3 y)dx +∫ (−4 y)dy+ c

y cons tant
2 2
=3 x +3 xy −3 y +c , c being an arbitrary constant
which is the required common solution of (1) and (2).
(iii) The given system of partial differential equations are
p=ax+ hy+ g …(1)
and q=hx +by + f …(2)
Here P=ax+ hy+ g and Q=hx+ by+ f
∂P ∂Q
=h =h
\ ∂y and ∂ x
∂ P ∂Q
= =h
⇒ ∂y ∂x
Hence, (1) and (2) are compatible.
Now we shall find common solution of (1) and (2):
Substituting the value of p and q in dz = p dx + q dy, we get
dz = (ax + hy + g) dx + (hx + by+ f ) dy
= ax dx + by dy + h (ydx + xdy) + gdx + fdy

2 2
x y
z=a + b + hxy+ gx + fy +c
Integrating , 2 2
 1
= ( ax 2 +by 2 )+ hxy+ gx + fy+c , c
2 being an
arbitrary constant, which is the required common solution of (1) and (2).

CHARPIT’S METHOD
We shall develop this method for two independent variables.
Let the given equation be f(x, y, z, p, q) =0 …(1)
Since z is function of two independent variables, therefore
∂z ∂z
dz= dx+ dy
∂x ∂y
i.e., dz = pdx + qdy …(2)
Now our interest is to find another relation, say
F(x, y, z, p, q) = 0 …(3)
such that when they values of p and q obtained from it and equation (1),
are substituted in (2), it becomes integrable. The integral of (2) will give
the complete integral.
To obtain (3), differentiating (1) and (3) w.r.t. x, we get
 ∂f ∂f ∂f ∂ p ∂f ∂q
+ p+ + =0
∂x ∂z ∂ p ∂x ∂q ∂x …(4)
∂F ∂F ∂F ∂ p ∂F ∂q
+ p+ + =0
and ∂ x ∂ z ∂ p ∂ x ∂q ∂ x …(5)
Differentiating (1) and (3) w.r.t. y, we get
∂ f ∂ f ∂ f ∂ p ∂f ∂q
+ q+ + =0
∂ y ∂ z ∂ p ∂ y ∂q ∂ y …(6)
∂F ∂F ∂ F ∂ p ∂F ∂q
+ q+ + =0
and ∂ y ∂z ∂ p ∂ y ∂q ∂ y …(7)

∂p ∂F
Eliminating ∂ x from (4) and (5) by multiplying (4) by ∂ p and (5) by
∂f
∂ p and substracting, we get
 ( ∂f ∂F ∂F ∂f
−
∂x ∂ p ∂x ∂ p) (
+p − ) (
∂ f ∂ F ∂ F ∂ f ∂q ∂ f ∂ F ∂ F ∂ f
+ −
∂z ∂ p ∂ z ∂ p ∂x ∂q ∂ p ∂q ∂ p )
=0
…(8)

∂q ∂F
Eliminating ∂ y from (6) and (7) by multiplying (6) by ∂ q and (7) by
∂f
∂q and substracting, we get

( ∂f ∂F ∂F ∂f
−
∂ y ∂q ∂ y ∂q) (
+q . − . +) (
∂ f ∂ F ∂ F ∂f ∂ p ∂ F ∂ f ∂ F ∂ f
−
∂ z ∂q ∂ z ∂ q ∂ y ∂ q ∂ p ∂ p ∂q
=0 )
…(9)

2
∂q ∂ z ∂ p
= =
Now adding (8) and (9) and using ∂ x ∂ x∂ y ∂ y and after
rearranging the terms, we obtain

( ∂f
+p ) (
∂f ∂F ∂f
+ +q
∂f ∂F
∂ x ∂ z ∂ p ∂ y ∂ z ∂q ) ( ∂f
+ − p −q
∂f ∂F
∂ p ∂q ∂ z )
 +−( ) ( )
∂f ∂F
∂ p ∂x
+−
∂f ∂F
∂q ∂ y
=0
…(10)
which is a linear equation of the first order with x, y, z, p, q as
independent variables and F as dependent variable. Using extension of
Lagrange’s auxiliary equations for eq. (10), we have
dp dq dz dx dy dF
= = = = =
∂f
+p
∂ f ∂f
∂x ∂z ∂y ∂z
+q
∂f ∂f
−p −q
∂f
∂ p ∂q
−
∂f
∂p
− ( )( )
∂f
∂q
0
…(11)
which are known as Charpits auxiliary equations.
Any integral of (11) will satisfy (10). Take the integral of simplest
relation involving p or q or both so that it is easy to find the solution of
(1). This integral can be taken for the assumed relation (3) i.e., for F = 0.
This integral with the help of (1) gives p and q when substituted in dz =
pdx + qdy and integrating, we obtain the desired solution.
Remark.
Charpits method can be used to obtain the solutions of some special
types of first order partial differential equations.
For example:
(i) Equations involving p and q only i.e., of the type f (p, q) = 0
 ∂f ∂f
=0 , =0 , ∂ f =0
Here ∂ x ∂y ∂z
Thus, Charpits equations are
dx dy dz dp dq
= = = =
∂f ∂f ∂f ∂f 0 0
− − − p −q−
∂p ∂q ∂p ∂q
\ dp = 0 and dq = 0 ⇒ p = a and q = b
Since f (p, q) = 0
\ f(a, b) = 0 ⇒ b = f (a) ⇒ p = a and q = f (a)
Substituting these values in dz = pdx + qdy, we have
dz = a dx + f (a) dy
Integrating, z = ax+ f (a) y + c, c being constant.
(ii) Equations involving p and q only i.e., of the type f (p, q) = 0
[Such equations are also called Clairaut equations]
In this case, Charpits equations implies p = a and q = b.
Substituting these values in dz = pdx + qdy and then integrating, we
get the required solution viz., z = ax + by + c, c = f (a, b).
 Remember the Charpits equation in this form:
dx dy dz dp dq dF
    
f f f f f f f f 0
   p  q p q
p q p q x z y z

SOLVED EXAMPLES:
Example1. Solve the equation: pxy + pq + qy = yz.
Solution. Here f (x, y, z, p, q) = pxy + pq + qy – yz = 0 …(1)
Charpits auxiliary equations are
dx dy dz dp dq dF
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f ∂f ∂f 0
− − − p −q− +p +q
∂p ∂q ∂p ∂q ∂ x ∂ z ∂ y ∂ z …(2)
∂f ∂f ∂f ∂f ∂f
= py , = px+q−z , =− y , =xy +q , = p+ y
Now, ∂ x ∂y ∂z ∂p ∂q
\ From (1), Charpit’s auxiliary equations are
 dx dy dz dp dq
= = = =
−( xy +q ) −( p+ y ) − p( xy +q )−q( p+ y ) py− py px +q−z−qy
…(3)
∂p
=each fraction
Eq. (3) gives 0
Eq. (2) gives dp=0 ⇒ p=c (constant )
Substituting p = c in (1), we get
cxy + cq + qy – yz = 0
or (c + y) q = yz – cxy
yz−cxy
q=
or c+ y
Now, dz = pdx + qdy
yz −cxy
dz=cdx+ dy
i.e., c+ y
y ( z−cx )
dz−cdx= dy
or c+ y
or
( dz−cdx
z −cx )
=
y
c+ y
dy

or
( dz−cdx
z −cx )(
= 1−
y
c+ y
dy)
Integrating, log (z – cx) = y – c log (c + y) + b
or log (z – cx) + c log (c + y)= y+b
which is the complete integral.

Example 2. Solve the partial differential equation px + qy = pq.

Obtain the general solution and show that there is no singular
solution.
Solution. Here f (x, y, z, p, q) = px + qy – pq = 0 …(1)
∂f ∂f ∂f ∂f ∂f
=x−q , = y− p , =0 , = p , =q
\ ∂p ∂q ∂z ∂x ∂y …(2)

Charpit’s auxiliary equations are

 dx dy dz dp dq dF
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f ∂f ∂f 0
− − − p −q− +p +q
∂p ∂q ∂p ∂q ∂ x ∂ z ∂ y ∂ z
dx dy dz dp dq
= = = =
\
−( x−q ) −( y + p ) − p (x−q )−q ( y+ p ) p q
…(3) [Using (2)]
dp dq
=
From last two fractions, p q
Integrating, log p = log q + log a
Or p = aq
Putting p = aq in (1), we have
aqx + qy – aq2 = 0 …(4)
y +ax
q= ⇒ aq=( y+ ax )
\ a
\ p = aq = (y + ax)
Now, dz = pdx + qdy
i.e.,
dz=( y +ax ) dx + ( y +ax
a
dy )
or a dz = a (y + ax) dx + (y + ax) dy
or a dz = (y + ax) (a dx + dy)
1
or a dz = 2 d (y + ax)2
1
Integrating, az = 2 (y + ax)2 + b …(5)
which is the complete integral.
To find the general solution:
Let b = f (a)
1
\ az = 2 (y + ax)2 + f(a) …(6)
Differentiating w.r.t. a, we have
z = (y + ax)x + f¢ (a) …(7)
General solution is obtained by eliminating a from (4) and (5).
 To find the singular solution:
Differentiating the complete integral (5) w.r.t. a and b.
z = x(y + ax)
i.e., 0 = 1(which is not possible)
It shows that there is no singular integral.
Example 3. Find the complete solution of partial differential equation
2xz – px2 – 2qxy + pq = 0
by using Charpit’s method.
Solution. Here f (x, y, z, p, q) = 2xz – px2 – 2qxy + pq = 0 …(1)
∂f
=2 z −2 px−2 qy
∂x
∂f
=−2 qx
∂y
∂f
=2 x
∂z
∂f
=−x 2 +q
∂p
 ∂f
=−2 xy + p
∂q
Charpit’s auxiliary equations are
dx dy dz dp dq dF
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f ∂f ∂f 0
− − − p −q− +q +q
∂p ∂q ∂p ∂q ∂ x ∂ z ∂ y ∂ z
dx dy dz dp dq dF
= = 2 = = =
\
2
x q 2 xy −p px − pq+2 xyq− pq 2 z−2qy 0 0
It gives dq = 0 i.e., q = a
Substituting this value of q in (1), we have
2xz – px2 – 2qxy + pq = 0
or 2x (z – ay) – p(x2 – a) = 0
2 x( z −ay )
p=
or x 2 −a
Now, dz = pdx + qdy
2 x ( z −ay )
dz= 2
dx+ a dy
i.e., x −a
 2 x ( z −ay )
dz−a dy =
or x 2 −a
dz−a dy ) 2x
= 2
or z −ay x −a
Integrating log (z – ay) = log (x2 – a) + log b
\ (z – ay) = b (x2 – a)
i.e., z = ay + b (x2 – a) .
which is the required complete solution.
Example 4. Apply Charpit’s method to find the complete integral of z =
px + qy + p2 + q2.
Solution. Here f (x, y, z, p, q) = z – px – qy – p2 – q2 = 0 …(1)
∂f ∂f ∂f
=− p, =−q , =1
∂x ∂y ∂z
∂f ∂f
=−x−2 p , =− y−2 q
∂p ∂q
Charpit’s auxiliary equations are
dx dy dz dp dq
= = = =
∂f ∂f ∂f ∂f ∂f ∂f ∂f ∂f
− − − p −q− +p +q
∂p ∂q ∂p ∂q ∂ x ∂ z ∂ y ∂ z
 dx dy dz dp dq dF
= = = = =
i.e., x+2 p y+ 2q p ( x +2 p)+q ( y +2 q ) − p+ p −q+ q 0
which gives dp = 0 and dq = 0
Integrating, p = a and q = b
Putting these values in (1), we have z = ax + by + a2 + b2.

Example 5. Solve p = (qy +z)2 by using Charpits method.

Solution. Here f (x, y, z, p, q) = – px + (qy +z)2 …(1)
∂f ∂f ∂f
=0 , =2(qy +z)q , =2(qy +z )
∂x ∂y ∂z
∂f ∂f
=−1 , and =2(qy +z) y
∂p ∂q

Charpit’s auxiliary equations are

dx dy dz dp dq dF
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f ∂f ∂f 0
− − − p −q− +p +q
∂p ∂q ∂p ∂q ∂ x ∂ z ∂ y ∂ z
i.e.,
dx dy dz dp dq
= = = =
−(−1) 2 y ( qy+ 2) p−2 q(qy + z ) y 0+2 p (qy + z ) 2q (qy + z )+2 q( qy+z )
dx dy dz dp dq
= = = =
or 1 −2 y (qy +2 ) p−2q (qy + z ) y 2 p(qy + z ) 4 q(qy +z )
dy dp
=
which gives −2 y ( qy+z ) 2 q (qy+ z )
dy dp
=
or −y p
dy dp
+ =0
or y p
Integrating, log y + log p = log a
a
a= yp ⇒ p=
\ y
Substituting this value of p in (1), we have
 a
=( qy + z )2
y

or √ a
y
=qy+ z

or √ a
y
−z =qy

q=
√a − z
\ y 3/2 y
Now dz = pdx + qdy

i.e.,
dz=
a
y y (
dx+ √3/ 2 − dy
a z
y )
or
ydz=adx+ √
√y
a
(
−z dy
)
ydz + zdy=adx+
√ a
dy
or √y
 d ( yz )=adx + √ dy
a
or √y
Integrating, yz=ax+2 √ a √ y +b , which is the required complete
integral.

JACOBI’S METHOD:
This method is used for solving partial differential equations
involving three or more independent variables. We shall develop this
method for three independent variables.
Consider the partial differential equation with three independent
variable x 1 , x2 , x 3 as
f ( x1 , x 2 , x 3 , p 1 , p2 , p3 )= 0
…(1)
∂z ∂z ∂z
p1 = , p2 = and p3 =
where ∂ x1 ∂ x2 ∂ x3
Here, z is the dependent variable not occurring in the equation except
in its partial derivatives p1 , p2 , p 3 with respect to x 1 , x2 , x 3 respectively.
 The main purpose of Jacobi’s method is to find two additional
equations
F 1 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 ) =a1
…(2)
F 2 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 ) =a2
and …(3)
where a 1 , a2 are arbitrary constants such that
(i) eqs. (1), (2) and (3) may be solved for p1 , p2 , p 3 as functions of
x 1 , x2 , x 3 and

(ii) dz= p 1 dx 1 + p2 dx 2 + p3 dx 3 …(4)

becomes integrable
For which the conditions are
∂ p2 ∂2 z ∂ p1 ∂ p3 ∂ p 2 ∂ p 1 ∂ p3
= = , = , =
∂ x1 ∂ x1 ∂ x2 ∂ x2 ∂ x2 ∂ x3 ∂ x3 ∂ x1 …(5)
Now differentiating (1) and (2) partially w.r.t. x 1 , we have
∂ f ∂ f ∂ p1 ∂ f ∂ p 2 ∂ f ∂ p 3
+ . + . + . =0
∂ x 1 ∂ p 1 ∂ x1 ∂ p2 ∂ x 1 ∂ p3 ∂ x 1 …(6)
 ∂ F 1 ∂ F 1 ∂ p1 ∂ F 1 ∂ p2 ∂ F 1 ∂ p3
+ . + . + . =0
and ∂ x 1 ∂ p1 ∂ x 1 ∂ p2 ∂ x1 ∂ p 3 ∂ x1 …(7)
∂ p1 ∂ F1
Eliminating ∂ x1 from (6) and (7) by multiplying (6) by ∂ p1 and (7)
∂f
by ∂ p1 and then subtracting, we have

( ∂ f ∂ F1 ∂ f ∂ F1
. − .
∂ x 1 ∂ p1 ∂ p1 ∂ x1
+
)( . − . +
) (
∂ f ∂ F 1 ∂ f ∂ F 1 ∂ p2 ∂f ∂ F 1 ∂ f ∂ F 1 ∂ p3
. − .
)
∂ p2 ∂ p1 ∂ p1 ∂ p2 ∂ x 1 ∂ p3 ∂ p1 ∂ p1 ∂ p3 ∂ x 1
=

…(8)
∂f ∂ f ∂ p1 ∂ f ∂ p 2 ∂ f ∂ p3
= . + . + . =0
∂ x2 ∂ p1 ∂ x 2 ∂ p2 ∂ x 2 ∂ p 3 ∂ x2 …(9)
∂ F1 ∂ F 1 ∂ p1 ∂ F 1 ∂ p 2 ∂ F 1 ∂ p 3
+ . + . + . =0
and ∂ x2 ∂ p1 ∂ x2 ∂ p 2 ∂ x 2 ∂ p3 ∂ x 2 …(10)
 ∂ p2 ∂ F1
Eliminating ∂ x 2 from (9) and (10) by multiplying (9) by ∂ p2 and
∂f
(10) by ∂ p2 and then subtracting, we have

( ∂ f ∂ F1 ∂ f ∂ F1
. − .
∂ x 2 ∂ p2 ∂ p2 ∂ x 2)(
+ . − .
) (
∂ f ∂ F 1 ∂ f ∂ F 1 ∂ p1 ∂ f ∂ F 1 ∂ f ∂ F 1 ∂ p 3
+ . − .
∂ p1 ∂ p2 ∂ p2 ∂ p1 ∂ x 2 ∂ p3 ∂ p2 ∂ p2 ∂ x 3 ∂ x 2 )
=

…(11)

Similarly, differentiating (1) and (2) partially w.r.t. x 3 and then

∂ p3
eliminating ∂ x 3 from the resulting equation, we have

( ∂ f ∂ F 1 ∂ f ∂ F1
. − .
)(
∂ x 3 ∂ p3 ∂ p3 ∂ x 3
+ . − .
) ( + . − .
)
∂ f ∂ F1 ∂ f ∂ F 1 ∂ p 1 ∂ f ∂ F 1 ∂ f ∂ F 1 ∂ p 2
∂ p1 ∂ p3 ∂ p 3 ∂ p 1 ∂ x 3 ∂ p2 ∂ p3 ∂ p 3 ∂ x 2 ∂ x 3
=

…(12)
 Adding (8), (11) and (12) and using relation (5), we have

( )(
∂f ∂F 1 ∂f ∂F 1 ∂f ∂F 1 ∂f ∂F 1 ∂f ∂F 1 ∂f ∂F 1
. − . + . − . + . − . =0 ¿ ¿ ¿¿
)(
∂x1 ∂ p1 ∂ p1 ∂x1 ∂x2 ∂ p2 ∂ p2 ∂x2 ∂x3 ∂ p3 ∂ p3 ∂x 3 ) }
…(13)
Here, L.H.S. of (13) is denoted by ( f , F 1 ) . Eqn. (13) is a linear
equation of 1st order with x 1 , x2 , x 3 , p1 , p 2 , p 3 as independent variables and
F 1 as dependent variable.
Now, consider equation (1) and (3). Proceeding as above, we get

( )
3
∂ f ∂ F2 ∂ f ∂ F2
( f , F 1 ) ≡∑ . − .
∂ xr ∂ pr ∂ pr ∂ xr
=0
r=1
…(14)
which is a linear equation of 1st order with x 1 , x2 , x 3 , p1 , p 2 , p 3 as
independent variables and F 2 as dependent variable.
Again, repeating the above procedure for equations (2) and (3), we have

( )
3
∂ F 1 ∂ F 2 ∂ F 1 ∂ F2
( F 1 , F2 )≡ ∑ . − .
∂ x r ∂ pr ∂ pr ∂ x r
=0
r=1
…(15)
 Since eqns. (13) and (14) both are linear equation of first order,
therefore for both equations, Lagrange’s auxiliary equations are
dx 1 dp 1 dx 2 dp 2 dx 3 dp3
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f
− − −
∂ p1 ∂ x 1 ∂ p2 ∂ x 2 ∂ p3 ∂ x 3 …
(16)
which are known as Jacobi’s auxiliary equations.

From (16), we shall find two independent integrals

F 1 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 ) =a1
and F 2 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 ) =a2 .
If these
satisfy condition (15), then these relations are taken as two additional
equations.

Now, solving (1), (2), (3) we find p1 , p2 , p 3 in terms of x 1 , x2 , x 3 .

Then substituting these values in (4) and integrating, we get the complete
integral of the given equation containing three arbitrary constants of
integration.

Working Rule for Jacobi’s Method:

Let the given equation with three independent variables be s
f ( x1 , x 2 , x 3 , p 1 , p2 , p3 )= 0
…(1)
Step 1. Consider Jacobi’s auxiliary equations
dx 1 dp 1 dx 2 dp 2 dx 3 dp3
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f
− − −
∂ p1 ∂ x 1 ∂ p2 ∂ x 2 ∂ p3 ∂ x 3

Step 2. Solve these equations to obtain two additional equations

F 1 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 ) =a1
…(2)
F 2 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 ) =a2
and …(3)
where a 1 and a 2 are arbitrary constants.
Step 3. Verify that relations (2) and (3) satisfy the condition
 ( )
3
∂ F 1 ∂ F 2 ∂ F 1 ∂ F2
( F 1 , F2 )≡ ∑ . − .
∂ x r ∂ pr ∂ pr ∂ x r
=0
r=1
…(4)
Step 4. If eqn. (4) is satisfied then solve eqns. (1), (2) and (3) for
p1 , p2 , p 3 in terms of x 1 , x2 , x 3 .

Substitute these values p1 , p2 , p 3 in

dz= p 1 dx 1 + p2 dx 2 + p3 dx 3
and integrate to get the complete integral.
SOLVED EXAMPLES:
p1 x 1 + p2 x 2 = p
Example 1. Find the complete integral of 32 by using
Jacobi’s method.
f ( x1 , x 2 , x 3 , p 1 , p2 , p3 )= p 1 x 1 + p 2 x2 − p 2= 0
Solution. Here 3 …
(1)
∂f ∂f ∂f ∂f ∂f ∂f
= p1 , =p 2 , =0 , =x1 , =x 2 , =−2 p3
\
∂ x1 ∂ x2 ∂ x3 ∂ p1 ∂ p2 ∂ p3

Jacobi’s auxiliary equations are

 dx 1 dp 1 dx 2 dp 2 dx 3 dp. 3
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f
− − −
∂ p1 ∂ x1 ∂ p2 ∂ x2 ∂ p3 ∂ x3
dx 1 dp1 dx 2 dp 2 dx 3 dp 3
= = = = =
\
−x1 p1 −x 2 p2 2 p3 0 …(2)
From first two fractions of (2), we have
dx 1 dp1
=
−x1 p 1
i. e . ,
1
x1
dx 1 +
1
p1 ( ) ( )
dp1 =0

Integrating, log x1 + log p1 = log a1

⇒ x1 p1 = a1
F 1 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 ) =x1 p1 =a 1
Let …(3)
From third and fourth fractions of (2), we have
dx 2 dp 2
=
−x 2 p2
i. e . ,
1
x2
dx 2 +
1
p2
dp 2 =0
( ) ( )
Integrating, log x2 + log p2 = log a2 ⇒ x2 p2 = a2
Let F 2 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 ) =x 2 p 2=a2 …(4)
 ( )
3
∂ F 1 ∂ F 2 ∂ F 1 ∂ F2
( F 1 , F2 )= ∑ . − .
∂ x r ∂ pr ∂ pr ∂ x r
Now, r=1

=
( .
∂ x1 ∂ p1 ∂ p1 ∂ x1
+ .
)(
∂ F 1 ∂ F2 ∂ F1 ∂ F 2 ∂ F 1 ∂ F 2 ∂ F 1 ∂ F2
. − − .
∂ x 2 ∂ p2 ∂ p2 ∂ x 2 )
+
(
∂ F1 ∂ F 2 ∂ F 1 ∂ F 2
. − .
∂ x 3 ∂ p3 ∂ p3 ∂ x3 )
=( p1 ).(0 )−( x 1 ).(0)+( 0).( x 2 )−( 0 ).( p2 )+( 0).( 0 )−( 0 ) .( 0)
=0 [By (3) and (4)]
Thus, ( F 1 , F2 )=0 is verified for the relations (3) and (4).
Hence (3) and (4) may be taken as additional equations.
a1
p1 =
From (3), x1
a2
p2 =
From (4), x2
p 2= p1 x 1 + p 2 x 2 =a 1 +a2 ⇒ p3 = √ a 1 +a 2
From (1), 3
 Substituting these values in dz= p 1 dx 1 + p2 dx 2 + p3 dx 3 , we have
dz=
a1
x1( ) ( )
dx 1 +
a2
x2
dx 2 + √ a1 +a 2 . dx 3

Integrating, z=a1 log x1 +a2 log x 2 + x 3 √ a 1 +a2 . a 3

which is the complete integral of the given equation containing three
arbitrary constants a 1 , a2 and a 3 .
Example 2. Find complete integral of p3 x 3 ( p1 + p 2 )+ x 1 + x2 =0
using Jacobi’s method.
Solution. Here f ( x1 , x 2 , x 3 , p 1 , p2 , p3 )= p 3 x3 ( p1 + p2 )+ x 1 + x 2 =0 …
(1)
∂f ∂f ∂f
=1 , =1 , = p3 ( p 1 + p2 )
\ ∂ x1 ∂ x2 ∂ x3
∂f ∂f ∂f
= p3 x3 , =p 3 x 3 , =x 3 ( p 1 + p2 )
∂ p1 ∂ p2 ∂ p3
Jacobi’s auxiliary equations are
 dx 1 dp 1 dx 2 dp 2 dx 3 dp. 3
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f
− − −
∂ p1 ∂ x 1 ∂ p2 ∂ x 2 ∂ p3 ∂ x 3
dx 1 dp 1 dx 2 dp 2 dx 3 dp 3
= = = = =
\
−p 3 x 3 p1 − p 3 x3 1 −x 3 ( p1 + p 2 ) p3 ( p1 + p2 )

…(2)
Taking the second and fourth fractions of (2), we have
dp1 dp2
=
1 1

⇒ dp 1−dp 2 =0

Integrating, p1 − p2 =a1
F 1 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 ) =p 1− p 2=a1
\ Let …(3)
Taking the firth and sixth fractions of (2), we have
dx 3 dp 3
=
−x 3 ( p1 + p2 ) p 3 ( p1 + p 2 )
or
( ) ( )
1
x 3
dx 3 +
1
p 3
dp 3 =0

Integrating, log x 3 + log p3 =log a2

or x 3 p3 =a2
F 2 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 ) =x 3 p 3 =a2
Let …(4)

( )
3
∂ F 1 ∂ F 2 ∂ F 1 ∂ F2
( F 1 , F2 )= ∑ . − .
∂ x r ∂ p r ∂ pr ∂ x r
Now, r=1

=
( .
∂ x1 ∂ p1 ∂ p1 ∂ x1
+ . −
)(
∂ F 1 ∂ F 2 ∂ F 1 ∂ F 2 ∂ F 1 ∂ F 2 ∂ F 1 ∂ F2
. − .
∂ x 2 ∂ p2 ∂ p2 ∂ x 2 )
+
(
∂ F1 ∂ F 2 ∂ F 1 ∂ F 2
. − .
∂ x 3 ∂ p3 ∂ p3 ∂ x3 )
=( 0 ).(0 )−( 1)( 0 )+(0 )( 0)−(−1).( 0)+( 0)( x 3 )−( 0 )( p3 )
=0 [By (3) and (4)]
Thus, it is verified that for the relations (3) and (4), (F1, F2) = 0
Hence, (3) and (4) may be taken as additional equations.
From (1) and (4), we have
 a 2 ( p1 + p2 )+ x 1 + x 2 =0
x 1+ x2
p1 + p2 =−
or a2 …(5)
a1 x 1+ x 2
p1 = −
Adding (3) and (5), we have 2 2 a2
Subtracting (3) from (5), we have
a1 x 1+ x 2
p2 =− −
2 2 a2
a2
p3 =
From (4), x3

Substituting p1 , p2 , p 3 in dz= p1 dx 1 + p 2 dx 2 + p3 dx 3 , we have

dz=
2(
a1 x 1 + x 2
−
2 a2
dx 1 + − −
2 ) (
a 1 x 1 + x2
2 a2
a2
dx 2 + dx 3
x3 )
a1
= ( dx 1−dx 2 )−
2
x 1 + x2
2 a2 ( ) a2
( dx 1 + dx 2 )+ dx 3
x3
 a1 1
z= ( x 1 −x 2 )− ( x 1 + x 2 )2 + a2 log x 3 +a 3 ,
Integrating, 2 4 a2
which is the required complete integral.
Example 3. Find the complete integral of
2
( x 2 + x 3 )( p 2 + p3 ) + zp 1 =0 .
Solution. The given equation is
( x 2 + x 3 )( p 2 + p3 )2 + zp 1 =0 …
(1)
Here eqn. (1) involves the dependent variable z. So, we shall first
reduce (1) to the form that does not involve dependent variable so that
Jacobi’s method may be applied.
Now, eqn. (1) can be rewritten as

( )
2
1 ∂z 1 ∂z 1 ∂z
( x 2 +x 3 ) + + =0
z ∂ x2 z ∂ x3 z ∂ x1
…(2)

Let
()
1
z
dz=dZ
so that Z = log z. Then from (2), we have
 ( )
2
∂Z ∂Z ∂Z
( x 2+ x 3 ) + + =0
∂ x 2 ∂ x 3 ∂ x1
…(3)

Let
( )
∂Z
∂ x1
=P1 ,
( )
∂Z
∂ x2
=P2 ,
( )
∂Z
∂ x3
=P3 .
Then from (3), we have
2
( x 2 + x 3 )( P 2 + P3 ) + P1 =0
This equation does not involve dependent variable except in its
partial derivatives P1 ,P2 ,P3 w . r .t . x 1 , x 2 , x3 respectively. So, we can now
apply Jacobi’s Method.
2
Here f ( x1 , x 2 , x 3 , P 1 ,P2 ,P3 ) =( x 2 + x3 )( P2 + P 3 ) + P1 =0 …(4)
∂f ∂f ∂f
=0 , =( P2 +P 3 )2 , =( P2 +P 3 )2
\ ∂ x1 ∂ x2 ∂ x3
∂f ∂f ∂f
=1 , =2( P2 +P 3 )( x 2 +x 3 ), =2 (P2 +P3 )( x2 +x 3 )
∂ P1 ∂ P2 ∂ P3
Jacobi’s auxiliary equations are
 dx 1 dp 1 dx 2 dP2 dx 3 dP. 3
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f
− − −
∂ P1 ∂ x 1 ∂ P2 ∂ x 2 ∂ P3 ∂ x3
\
dx 1 dP1 dx 2 dP2 dx 3 dP 3
= = = = =
−1 0 −2( P2 + P 3 )( x 2 + x 3 ) ( P2 + P 3 )2 −2( x 2 + x3 )( P2 + P 3 ) ( P2 + P3 )2
…(5)
Taking the second fraction of (5), dP1 = 0
Integrating, P1 = a1
Let F 1 ( x 1 , x 2 , x 3 , P1 ,P2 ,P3 )=P1 =a 1 …(6)
Taking the fourth and sixth fractions of (5), we have
dP2 dP3
2
=
( P 2 + P3 ) ( P 2 + P3 )2 ⇒ dP 2−dP3 =0

Integrating, P2 −P3 =a2 …(7)

Let F 2 ( x 1 , x 2 , x 3 , P1 ,P2 ,P3 )= P2 −P3 =a 2 …
(8)
 As in the previous example, verify that the condition (F1, F2) is satisfied
for relations (6) and (8). Hence (6) and (8) may be taken as two additional
equations.
From (4) and (6), we have

{ }
1/2
−a1
P2 + P 2= ±
x 2+ x 3
…(9)
Adding (8) and (9), we have

[ ( )]
1/ 2
1 −a 1
P2 = a 2±
2 x 2 + x3

Subtracting (8) and (9), we have

[( ) ]
1/ 2
1 −a 1
P2 = ± −a2
2 x2+ x3

Substituting the values of P1, P2, P3 in dZ = P 1dx1+ P2dx2 + P3dx3, we get

[ ( ) ] [ ( ) ]
1/ 2 1/2
1 −a1 1 −a1
dZ=a1 dx 1 + a2 ± dx 2 + ± −a2
2 x 2+ x 3 2 x 2 + x3
 a2
=a1 dx 1 + ( dx 2 −dx 3 )±
√−a1
( x 2 + x 3 )−1/ 2 ( dx 2 +dx 3 )
2 2
a2
Z=a1 x 1 + ( x 2 −x 3 )±√ −a1 ( x2 + x 3 )1 /2 +a3
Integrating, 2
Since Z = log z, therefore we have
a2
log z =a1 x 1 + ( x 2 −x 3 )±√ −a1 ( x 2 + x3 )1/ 2 + a3 ,
2
which is the required complete integral containing a1, a2, a3 as arbitrary
constants.
a2
c 1 =−a1 , c2 = , c 3 =a3 , then
If 2
log z =−c 1 x 1 +c 2 ( x 2 −x 3 )±√ c 1 ( x 2 + x 3 )+ c 3 , where c1 , c2 , c3 are
arbitrary constants.

Example 4. Find the complete integral of

2 2 2
( p1 + x 1 ) +( p 2 + x 2 ) +( p3 + x3 ) =3( x 1 + x 2 + x 3 ) .
 2 2
Solution. Here f ( x1 , x 2 , x 3 , p 1 , p2 , p3 )=( p1 + x 1 ) +( p 2 + x 2 )
+( p 3 + x 3 )2−3( x1 + x 2 + x 3 )=0 …
(1)
∂f ∂f ∂f
=2( p 1 +x 1 )−3 , =2( p 2 +x 2 )−3 , =2( p3 + p3 )-3
\ ∂ x1 ∂ x2 ∂ x3
∂f ∂f ∂f
=2( p 1 +x 1 ), =2( p2 +x 2 ), =2( p3 +p 3 )
∂ p1 ∂ x2 ∂ x3
Jacobi’s auxiliary equations are
dx 1 dp 1 dx 2 dp 2 dx 3 dp. 3
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f
− − −
∂ p1 ∂ x 1 ∂ p2 ∂ x 2 ∂ p3 ∂ x 3
dx1 dp 1 dx 2 dp 2
= = =
\
−2( p 1 + x 1 ) 2( p1 + x 1 )−3 −2( p2 + x2 ) 2( p 2 + x 2 )−3
dx 3 dp 3
= =
−2( p3 + x3 ) 2( p3 + x 3 )−3 …(2)
 dp 1 +dx 1 dp 2 +dx 2 dp 3 + dx 3
= = =
Each fraction of (2) −3 −3 −3 …(3)
From (3), we have dp 1 +dx 1 =dp 2 +dx 2

Integrating, p1 + x 1 = p2 + x 2 + a1 ⇒ x 1 + p1 −x 2− p2 =a 1 ,
where a1 is an arbitrary constant.
\ Let F 1 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 )=x 1 + p1 −x2 − p2 =a 1 …(4)
also from (3) dp 3 +dx 3 =dp2 +dx 2

Integrating, p3 + x3 = p2 + x2 +a2
⇒ x3 + p3 −x 2− p2 = a2 ,
where a2 is an arbitrary constant.
\ Let F 2 ( x 1 , x 2 , x 3 , p1 , p 2 , p 3 )=x 3 + p3 −x 2 −p 2=a2 …(5)

( )
3
∂ F 1 ∂ F 2 ∂ F 1 ∂ F2
( F 1 , F2 )= ∑ . − .
∂ x r ∂ pr ∂ pr ∂ x r
Now, r=1

=
( .
∂ x1 ∂ p1 ∂ p1 ∂ x1
+ . −
)(
∂ F 1 ∂ F 2 ∂ F 1 ∂ F 2 ∂ F 1 ∂ F 2 ∂ F 1 ∂ F2
. − .
∂ x 2 ∂ p2 ∂ p2 ∂ x 2 )
 +
(∂ F1 ∂ F 2 ∂ F 1 ∂ F 2
. − .
∂ x 3 ∂ p3 ∂ p3 ∂ x3 )
=(1).(0)−(1 ).(0 )+(−1).(−1 )−(−1 ).(−1)+(0).(1)−(0 ).(1 )
=0 [By (4) and (5)]
Thus, (F1, F2) = 0 is verified for the relations (4) and (5).
Hence, (4) and (5) may be taken as additional equations.
From (4), x 1 + p 1 = x 2 + p 2 + a1

From (5), x 3 + p3 = x 2 + p2 +a2

Using these in (1), we have
2 2 2
( x 2 + p2 + a1 ) +( x 2 + p 2 ) ( x2 + p2 + a2 ) −3 ( x 1 + x2 + x3 )=0
2
3( x 2 + p2 ) + 2( x 2 + p 2 )( a1 + a2 )+ a 2 +a 2−3 ( x1 + x2 + x 3 )=0 ,
⇒ 1 2

which is a quadratic in x2 + p2.

\

x 2 + p2 =
√ 2
−2( a1 + a2 )± 4 ( a1 +a2 ) −12{a 2 + a 2 −3( x 1 + x 2 + x 3 )}
1 2
6
 ⇒

1
3 √
p2 =−x 2 + −( a1 +a 2 )± 9( x 1 + x 2 + x 3 )−2 a 2−a 2 +2 a1 a2
1 2

1
√[ 2
=−x2 + ( a 1 +a2 )± ( x 1 + x 2 + x 3 )− ( a 2 +a 2 −a1 a 2 )
3
1
9 1 2 ]
=−x2 − ( a1 + a2 )±t
3 …
(6)

[ ]
1/2
2
t = ( x1 + x 2 + x 3 )− ( a 2 + a 2−a1 a2 )
where 9 1 2

From (4) and (6), we have

2 a1 a 2
p1 =−x 1 + − ±t
3 3
From (5) and (6), we have
2 a2 a1
p3 =−x 3 + − ±t
3 3
Substituting the values of p1 p2, p3 in dz = p1 dx1 + p2 dx2 + p3, we get
 (
dz= −x 1 +
2 a1 a 2
3 ) (
− ±t dx 1 + −x 2 −
3
( a1 + a2 )
3 ) (
±t dx 2 + −x 3 +
2 a 2 a1
3
− ±t dx 3
3 )
=−( x dx + x dx + x dx )+ ( − ) dx −
2a a ( a +a )
1 2 2a a
1 2 2 1
1 1 2 2 3 3 dx +1 − dx 2 3
3 3 3 3 3

±t ( dx 1 +dx 2 + dx 3 )
1 1
=−( x 1 dx 1 + x 2 dx 2 + x 3 dx 3 )+ (2 a1 −a2 )dx 1− (a1 + a2 )dx 2
3 3

[ ]
1/ 2
1 2
+ ( 2a 2−a1 ) dx 3 ± ( x 1 + x 2 + x3 )− ( a 2 + a 2 −a 1 a2 ) ( dx 1 +dx 2 +dx 3 )
3 9 1 2

[Putting the value of t]

Integrating
1 1 1 1
z=− (x 2 + x 2 + x 2 )+ (2 a1 −a2 ) x1 − (a1 + a2 )x 2 + (2 a2 −a1 )x 3
2 1 2 3 3 3 3

[ ]
3 /2
2 2
± ( x1 + x 2 + x 3 )− (a 2 + a 2−a1 a2 ) +a3
3 9 1 2
which is the required complete integral containing a1, a2, a3 as arbitrary
constants.
If we put a1 = 3c1, a2 = 3c2, a3 = c3, then
1
z=− (x 2 + x 2 + x 2 )+(2 c 1−c2 ) x1 −(c 1 +c 2 )x 2 +(2 c2 −c 1 )x 3
2 1 2 3

2
± [ x 1 + x 2 + x 3 −2(c 2 + c 2 −c 1 c 2 )]3/2 +c 3
3 1 2

where c1, c2, c3 are arbitrary constants.

Example 5. Find the complete integral by Jacobi’s method for the
equation

( )( ) ( ) ( )
2 2
∂u ∂u ∂ u ∂ u
2 x2 y −x 2 −2 y =0
∂x ∂z ∂y ∂x
Solution. In the given equation, putting
∂u ∂u
x=x 1 , y=x 2 , z=x3 , = p2 and = p 3 , we have
∂y ∂z
2 x 2 x2 p p3 −x p2 −2 x 2 p 2=0
1 12 12 1
 p 2 2 p1 2
2 p 2 p3 − − =0
1 x2 x 2
or 1

p2 2 p12
f ( x1 , x 2 , x 3 , p 1, p 2 , p3 )=2 p 2 p 3 − − =0
1 x2 x 2
Let 1 …(1)
4p 2 p2
∂f 1 ∂f ∂f
= , = , =0 ,
∂ x1 x 3 ∂ x2 x 2 ∂ x3
\ 1 2

( )
∂f 1 ∂f 1 ∂f
=4 p1 p 3 − , =− , =2 p 2
∂ p1 x2 ∂ p2 x2 ∂ p3 1
1

Jacobi’s auxiliary equations are

dx 1 dp 1 dx 2 dp 2 dx 3 dp. 3
= = = = =
∂f ∂f ∂f ∂f ∂f ∂f
− − −
∂ p1 ∂ x 1 ∂ p2 ∂ x 2 ∂ p3 ∂ x 3
dx 1 dp 1 dx 2 dp 2 dx 3 dp 3
= = = = =

( )
1 4p 1 p2 2p 0
−4 p1 p3 − 12 12
x 2 x x2 x 2
1
\ 13 2
…(2)
From the sixth fraction of (2), we have
 dp3 = 0 ⇒ p3 = a1
\ Let F 1 ( x 1 , x 2 , x 3 , p 1 , p 2 , p 3 )=p 3 =a1 …(3)
From the third and fourth fractions of (2), we have
dx 2 dp 2
=
1 p2
x2 x 2
2

or
( ) ( )
1
p2
dp 2 =
1
x2
dx 2

Integrating, log p2 = log x2 + log a2

p2
=a 2
or p2 =x 2 a2 ⇒ x 2
p2
F 2 ( x 1 , x 2 , x 3 , p1 , p 2 , p 3 )= =a2
\ Let x2 …(4)
 ( )
3
∂ F 1 ∂ F 2 ∂ F 1 ∂ F2
( F 1 , F2 )= ∑ . − .
∂ x r ∂ pr ∂ pr ∂ x r
Now, r=1

=
( .
∂ x1 ∂ p1 ∂ p1 ∂ x1
+ .
)(
∂ F 1 ∂ F 2 ∂ F 1 ∂ F 2 ∂ F 1 ∂ F 2 ∂ F 1 ∂ F2
. − − .
∂ x 2 ∂ p2 ∂ p2 ∂ x 2 )
(
+
∂ F1 ∂ F 2 ∂ F 1 ∂ F 2
. − .
∂ x 3 ∂ p3 ∂ p3 ∂ x3 )
( ) ( )
1 − p2
=( 0 ).( 0 )−( 0) .( 0 )+( 0 ). 2 −( 0 ). +( 0) .( 0)−( 1 ).( 0 )
x x 2
2

=0
Since, the condition (F1, F2) = 0 is verified by (3) and (4), therefore
(3) and (4) may be taken as additional equations.
From (1), (3) and (4), we have

( )
1
2 p 2 a1 − −a2 =0
1 x 2
1

2 p 2 ( a1 x 2 −1)−a2 x 2 =0
or 1 1 1
or
p1 =
2√
a2
.
x1
√
a 1 x 2−1
1

Substituting the values p1, p2, p3 in dz = p1dx1 + p2dx2 + p3dx3, we

have

dz=
√ a2
2
x1
√ a1 x 12 −1
dx 1 + a2 x 2 dx 2 +a1 dx 3

Integrating,
z=
√
a2 1 1/ 2
. ( a x −1) +
2 a1 1 1 2
a2
2 ( )
. x 2 + a1 x 3 +a 3
2

which is the required complete integral with a1, a2, a3 as arbitrary

constants.

If we put √ a2
2
=c2 , a1 =c 1 , a 3=c3 ,
then

z=
( )
c2
c1
.( c 1 x 2−1 )1/ 2 +c 2 x 2 +c 1 x 3 + c3 ,
1 2 2

where c1, c2, c3 are arbitrary constants.