NATIONAL I NSTITUTE OF T ECHNOLOGY D URGAPUR

D EPARTMENT OF M ATHEMATICS

Higher Order Linear PDE with Constant Coefficients

Lakshmi Kanta Dey

July 26, 2023

1 L INEAR PARTIAL DIFFERENTIAL EQUATION WITH CONSTANT

COEFFICIENTS

In this section we consider a special type of linear partial differential equation, that with constant coeffi-
cients. We can write it as
∂nz ∂nz ∂nz ∂nz
a0 + a1 + a 2 + · · · + a n = f (x, y), (1.1)
∂xn ∂xn−1 ∂y ∂xn−2 ∂y 2 ∂y n
where a0 , a1 , . . . an are constants, n ∈ N. Equation (1.1) can be written in operator form as

a0 Dn + a1 Dn−1 D′ + a2 Dn−2 D′2 + · · · + an D′n z = f (x, y),

(1.2)

where D ≡ ∂x∂
, D′ ≡ ∂y
∂
. If we write F (D, D′ ) = a0 Dn + a1 Dn−1 D′ + a2 Dn−2 D′2 + · · · + an D′n ,
then the PDE can be written as
F (D, D′ )z = f (x, y). (1.3)
The solution containing the correct number of arbitrary constants of the corresponding nth order homo-
geneous linear partial differential equation F (D, D′ )z = 0, is called complementary function (C.F.) of
the equation (1.2) (i.e., (1.1)). Any solution (without containing any arbitrary constant) of (1.1) is called
particular integral (P.I.) of (1.1).
The general solution of (1.1) can be obtained by the following theorem:

Theorem 1.1. If u is a complementary function and z1 is a particular integral of (1.1), then a general
solution of (1.1) is given by z = u + z1 .

Remark 1. If the given partial differential equation is of the form F (D, D′ )z = 0, i.e., if f (x, y) = 0,
then z = u is the general solution of the equation, where u is the complementary function.

1
 1.1 M ETHOD OF FINDING COMPLEMENTARY FUNCTION
Let us consider a homogeneous linear partial differential equation with constant coefficients
a0 Dn + a1 Dn−1 D′ + a2 Dn−2 D′2 + · · · + an D′n z = f (x, y).


(1.4)
We know that the complementary function of (1.4) is the general solution of
a0 Dn + a1 Dn−1 D′ + a2 Dn−2 D′2 + · · · + an D′n z = 0.


(1.5)
Suppose that
a0 Dn + a1 Dn−1 D′ + a2 Dn−2 D′2 + · · · + an D′n
can be factorized as
(D − m1 D′ )(D − m2 D′ ) . . . (D − mn D′ ),
where m1 , m2 , . . . mn are all constants (real or complex), then the complementary function is the general
solution of
[(D − m1 D′ )(D − m2 D′ ) . . . (D − mn D′ )]z = 0,
i.e., of (D − m1 D′ )z = 0, (D − m2 D′ )z = 0, . . . (D − mn D′ )z = 0. (1.6)
We now choose a general form of the above factored equations
(D − mD′ )z = 0
=⇒ p − mq = 0,
which can be consider as a Lagrange’s equation. Lagrange’s auxiliary equations for the above equation
are
dx dy dz
= = .
1 −m 0
From this, we get the general solution as
z = ϕ(y + mx), ϕ being an arbitrary function of y + mx.
In this way, for each partial differential equations in (1.6), we have the general solutions as
z = ϕ1 (y + m1 x), z = ϕ2 (y + m2 x) . . . z = ϕn (y + mn x).
Then the complementary function will be
ϕ1 (y + m1 x) + ϕ2 (y + m2 x) + · · · + ϕn (y + mn x).
We can find the complementary function of (1.4) in an easy way by finding the roots and its nature of the
following algebraic equation:
a0 mn + a1 mn−1 + . . . an = 0. (1.7)
Equation (1.7) is obtained by replacing D by m and D′ by 1 is (1.4), and is known as auxiliary equation of
(1.4). First, suppose that the roots of (1.7) are all distinct and suppose that the roots are m1 , m2 , . . . , mn .
Then the complementary function will be
ϕ1 (y + m1 x) + ϕ2 (y + m2 x) + · · · + ϕn (y + mn x),
where ϕ1 , ϕ2 , . . . , ϕn are arbitrary functions of y + m1 x, y + m2 x, . . . , y + mn x respectively.
Next, suppose that there are r equal roots, say, m and (n − r) distinct roots, say, mr+1 , mr+2 , . . . , mn
of (1.7). Then the complementary function will be
ϕ1 (y + mx) + xϕ2 (y + mx) + · · · + xr−1 ϕr (y + mx) + ϕr+1 (y + mr+1 x) + · · · + ϕn (y + mn x),
where ϕ, ϕr+1 , . . . , ϕn are arbitrary functions of y + mx, y + mr+1 x, . . . , y + mn x respectively.

2
Example 1.1. Solve the PDE (D3 − 6D2 D′ + 11DD′2 − 6D′3 )z = 0.

Solution: The auxiliary equation of the given PDE is

m3 − 6m2 + 11m − 6 = 0
=⇒ m = 1, 2, 3

Therefore, C.F. is ϕ1 (y + x) + ϕ2 (y + 2x) + ϕ3 (y + 3x).

So the general solution of the given PDE is

z = ϕ1 (y + x) + ϕ2 (y + 2x) + ϕ3 (y + 3x).

Example 1.2. Find the complementary functions and the general solutions of the following partial dif-
ferential equations:
∂2z ∂2z
∂ z 2
1. ∂x2
+
+ 2 ∂x∂y
∂y 2
= 0.

2. D3 − 3D2 D′ + 2DD′ z = 0.

3. (D3 D′2 + D2 D′3 )z = 0.

4. r − 2s + 2t = 0.

Remark 2. Notice that where the solution of an ODE contains arbitrary constants, the solution to a PDE
contains arbitrary functions. In the same spirit, while an ODE of order n has n linearly independent
solutions, a PDE has infinitely many (there are arbitrary functions in the solution!). These are conse-
quences of the fact that a function of two variables contains immensely more (a whole dimension worth)
of information than a function of only one variable.

1.2 M ETHOD OF F INDING PARTICULAR I NTEGRALS

Let us consider a homogeneous linear pde of the form F (D, D′ )z = f (x, y). Then the particular integral
(P.I.) is
1
f (x, y),
F (D, D′ )
where 1
F (D,D′ ) is the inverse operator of F (D, D′ ), i.e., 1
F (D,D′ ) f (x, y) = g(x, y) if and only if

F (D, D′ )g(x, y) = f (x, y).

1
Before get into the general methods of obtaining F (D,D′ ) f (x, y), we recall the following necessary
results:

1. D, D2 , D3 , . . . will stand for differentiating partially with respect to x once, twice, thrice and so
on.

2. D′ , D′2 , D′3 , . . . will stand for differentiating partially with respect to y once, twice, thrice and so
on.
1 1 1
3. D , D2 , D3 , . . . will stand for integrating partially with respect to x once, twice, thrice and so on.
1 1 1
4. D′ , D′2 , D′3 , . . . will stand for integrating partially with respect to y once, twice, thrice and so on.

3
 1
We now discuss about different methods for obtaining F (D,D ′ ) f (x, y) for different forms of the function

f (x, y).
Case I: Suppose that f (x, y) is a function of ax + by for some constants a and b.

1. Replace D by a and D′ by b in F (D, D′ ) to get F (a, b) such that F (a, b) ̸= 0; and substitute
ax + by = u and integrate f (u) n times w.r.t. u. i.e.
ZZ Z
1
P.I. = · · · f (u)dudu . . . du.
F (a, b)

2. If F (a, b) = 0, differentiate F (D, D′ ) w.r.t. D partially and simultaneously multiply by x i.e.

1
P I = x. ∂
f (ax + by).
′
∂D [F (D, D )]

We proceed this method as long as the derivative of F (D, D′ ) vanishes when D is replaced by a
and D′ is replaced by b.

Theorem 1.2. If F (D, D′ ) be the homogeneous function of D and D′ of degree n, then P.I. of F (D, D′ )z =
f (n) (ax + by) is given by
1 1
′
f (n) (ax + by) = f (ax + by),
F (D, D ) F (a, b)

provided F (a, b) ̸= 0; where f (n) is the nth derivative of f w.r.t. (ax + by) as a whole.

Case II: Suppose that f (x, y) is a function of xm y n or polynomial function of x and y.

1. In this case, P.I. is obtained by expanding F (D, D′ ) in an infinite series in ascending powers of
D or D′ or D D
′ by using binomial theorem.

2. The operator D gives differentiation partially w.r.t x, D′ gives differentiation partially w.r.t y, 1
D
stands for integration partially w.r.t x and D1′ stands for integration partially w.r.t y.

Case III: Suppose that f (x, y) is any function of x and y.

Assume that F (D, D′ ) can be expressed as (D − m1 D′ ).(D − m2 D′ ) . . . (D − mn D′ ). Hence
1 1 1
P.I. = ′
. ′
... f (x, y).
(D − m1 D ) (D − m2 D ) (D − mn D′ )

We have the following rules:

1
R
1. (D−mD ′ ) f (x, y) = f (x, a − mx)dx where a = y + mx;
1
R
2. (D+mD′ ) f (x, y) = f (x, a + mx)dx where a = y − mx.

By repeated applications of the above method, we can find the desired particular integral.

Example 1.3. Find the particular integral of (D2 + DD′ − 6D′2 )z = y sin x.

4
Solution:
1
P.I. = y sin x
D2 + DD′ − 6D′2
1
= y sin x
(D − 2D′ )(D + 3D′ )
Z
1
= (a + 3x) sin xdx, where a = y − 3x
(D − 2D′ )
 Z 
1
= −(a + 3x) cos x + 3 cos xdx
(D − 2D′ )
1
= [−y cos x + 3 sin x]
(D − 2D′ )
Z
−(a′ − 2x) cos x + 3 sin x dx, where a′ = y + 2x

=

= −y sin x − cos x.

Example 1.4. Solve the PDE D3 − 4D2 D′ + 4DD′2 z = cos(2x + 3y).

Solution: The auxiliary equation of the given PDE is

m3 − 4m2 + 4m = 0
=⇒ m = 0, 2, 2

Therefore, C.F. is ϕ1 (y) + ϕ2 (y + 2x) + xϕ3 (y + 2x).

Now,
1
P.I. = cos(2x + 3y)
D(D2 − 4DD′ + 4D′2 )
Z Z Z
1
= cos v dv dv dv, where v = 2x + 3y
2(4 − 24 + 86)
Z Z
1
= sin v dv dv
32
1
= − sin v
32
1
= − sin(2x + 3y).
32

Therefore, the required solution is

1
z = ϕ1 (y) + ϕ2 (y + 2x) + xϕ3 (y + 2x) − sin(2x + 3y).
32
Example 1.5. Solve the PDE D2 − a2 D′2 z = x2 y + ex+2y , a ̸= 21 .

Solution: The auxiliary equation of the given PDE is

m2 − a2 = 0
=⇒ m = a, −a

Therefore, C.F. is ϕ1 (y + ax) + ϕ2 (y − ax).

5
Now,
1  2
P.I = x y + ex+2y
(D2 − a2 D′2 )
1 1
= x2 y + ex+2y
(D2 − a2 D′2 ) (D2 − a2 D′2 )
Z Z
1 2 1 1
=  x y + 2 2 2
ev dv dv where v = x + 2y and a ̸=
D2 1 − a2 D
′2 (1 − a 2 ) 2
D2

a2 D′2
 
1 1
= 1+ + . . . x2 y + ev
D2 D2 1 − 4a2
1 2 1
= 2
x y+ ex+2y
D 1 − 4a2
1 4 1
= x y+ ex+2y .
12 1 − 4a2

Therefore, the required solution is

1 4 1
z = ϕ1 (y + ax) + ϕ2 (y − ax) + x y+ ex+2y .
12 1 − 4a2
Exercise 1.1. Solve the following:

1. D2 − 3DD′ + 2D′2 z = e2x−y + ex+y + cos(x + 2y).

1 2x−y
Ans: P.I= 12 e − xex+y − 31 cos(x + 2y).

2. D2 − 6DD′ + 9D′2 z = tan(3x + y) + 12x2 .

3. D2 − DD′ − 2D′2 z = (y − 1)ex . Ans: P.I = yex .

4. D2 + D′2 z = cos mx cos ny. 1

Ans: P.I = − m2 +n2 cos mx cos ny.

2
5. D2 − DD′ − 6D′2 z = x2 sin(x + y). Ans: P.I = ( x4 − 13 3x


32 ) sin(x + y) − 8 cos(x + y).

6. 2D2 − 5DD′ + 2D′2 z = 24(y − x). Ans: P.I = 6x2 y + 3x3 .

7. D2 + DD′ − 6D′2 z = y sin x.

Ans: P.I = −y sin x − cos x.

Some special cases:

Case 1: If f (x, y) = eax+by , then
1 n
ax+by
o
P.I = e
F (D, D′ )
1
= eax+by , provided F (a, b) ̸= 0.
F (a, b)

6
Case 2: If f (x, y) = sin(ax + by) or cos(ax + by), then
1
P.I = {sin(ax + by)}
F (D, D′ )
1
= {sin(ax + by)}
G(D , DD′ , D′2 )
2

1
= sin(ax + by), provided G(−a2 , −ab, −b2 ) ̸= 0.
G(−a , −ab, −b2 )
2

Example 1.6. We can solve the PDE (D2 − 2DD′ + D′2 )z = sin x using the above method for finding
PI easily. Now for C.F, we first find the auxiliary equation as m2 − 2m + 1 = 0. Roots are −1, −1.
Therefore, C.F. is ϕ1 (y − x) + xϕ2 (y − x), where ϕ1 , ϕ2 are arbitrary functions.
Now,
1
P.I = {sin x}
− 2DD′ + D′2
D2
1
= sin x
(−1 + 0 + 0)
= − sin x.

Hence, the complete integral is z = ϕ1 (y −x)+xϕ2 (y −x)−sin x, where ϕ1 , ϕ2 are arbitrary functions.

Case 3: If f (x, y) = eax+by V (x, y), then

1 n
ax+by
o
P.I = e V (x, y)
F (D, D′ )
1
= eax+by V (x, y).
F (D + a, D′ + b)

Case 4: If F (a, b) = 0, then

1
P.I = f (ax + by)
F (D, D′ )
1
= x ∂F (D,D′ ) f (ax + by).
∂D

1.3 P ROBLEMS
1. Solve the following homogeneous PDE (with constant coefficients):
a) (D3 − 6D2 D′ + 5D′3 )z = 0.
b) (D4 − D′4 )z = 0. Ans: z = ϕ1 (y + x) + ϕ2 (y − x) + ϕ3 (y + ix) + ϕ4 (y − ix).
c) 25r − 40s + 16t = 0.
d) (D2 + 3DD́ + 2D́2 )z = x + y. Ans: P.I = − 13 x3 + 12 x2 y or 1 3
36 (x + y) .
1 2x+3y
e) (D + D́)2 z = e2x+3y . Ans: P.I = 25 e .

7
 x2
f) (D − D́)2 z = sin(x + y). Ans: P.I = 2 sin(x + y).
1 1
g) zxx − zxy = cos x cos 2y. Ans: P.I = 2 sin(x + 2y) + 6 sin(x − 2y).
h) (D2 − 3DD′ + 2D′ 2 )z = 2xy. Ans: P.I = x3 y/3 + x4 /4.
i) (D3 − 3D2 D′ + 4D′3 )z = ex+2y . Ans: P.I = 1 x+2y
27 e .
j) (D2 + 5DD′ + 6D′2 )z = 1
y−2x . Ans: P.I = x − (y − 3x) ln(y − 2x).
k) (D2 + 2DD′ + 2D′2 )z = 1
y−x . Ans: P.I = x ln(y − x).
∂3z 3 3
l) ∂x3
− 4 ∂x∂2 ∂y
z ∂ z
+ 4 ∂x∂y 2 = cos(2x + y).

m) (D − 3D′ )2 z = 12x2 + 36xy. Ans: P.I = 10x4 + 6x3 y.

n) (D2 + 3D′2 )z = −4π(x2 + y 2 ).
o) (D3 + 2D2 D′ − DD′2 − 2D′3 )z = (y + 2)ex . Ans: P.I = (y − 1)ex .
p) (D3 D′ + D2 D′3 )z = 0.
q) (D2 − D′ )z = 2y − x2 . Ans: Expand as in Case 1.2. PI = x2 y.

ACKNOWLEDGEMENT

This part of the course material is jointly prepared with my research student Dr Ashis Bera.

R EFERENCES
[1] Lecture note: https://ocw.mit.edu/courses/mathematics/18-152-introduction-to-partial-differential-
equations-fall-2011/lecture-notes/MIT18-152F11-lec-01.pdf

[2] Lecture note of Prof. Zachary S Tseng, PSU, https://www.math.psu.edu/tseng/class/Math251/Notes-

PDE

[3] I. N. Sneddon. Elements of Partial Differential Equations., McGraw-Hill,1957