Chapter 2

2.1 Ordinary Differential Equations

A differential equation (or DE) is an equation that involves one or more derivatives
of an unknown function. Solving the differential equation means finding a function
(or every such function) that satisfies the differential equation. Many physical laws
and relationships between quantities studied in various scientific disciplines are
expressed mathematically as differential equations.
For example, Newton’s second law of motion (F = ma) states that the position x(t)
at time t of an object of constant mass m subjected to a force F(t) must satisfy the
differential equation (equation of motion):
d 2x
m  F (t ) (2.1)
dt 2
Because differential equations arise so extensively in the abstract modeling of
concrete phenomena, such equations and techniques for solving them are at the heart
of applied mathematics.
Classifying Differential Equations
Differential equations are classified in several ways. The most significant
classification is based on the number of variables with respect to which derivatives
appear in the equation.
An ordinary differential equation (ODE) is one that involves derivatives with
respect to only one variable. Both of the examples given above are ordinary
differential equations.
A partial differential equation (PDE) is one that involves partial derivatives of the
unknown function with respect to more than one variable. For example, the one-
dimensional wave equation

1
  2u 2  u
2
 c (2.2)
t 2 x 2

models the lateral displacement u(x,t) at position x at time t of a stretched vibrating

string.
Differential equations are also classified with respect to order. The order of a
differential equation is the order of the highest order derivative present in the
equation. The one-dimensional wave equation is a second order PDE. The following
example records the order of two ODEs.

d2y
 x 3 y  sin x h a s o rd e r2,
dx 2
(2.3)
2
d3y  dy  d2y
 4 x    y  ey h a s o rd e r3
 dx 
3 2
dx dx

Like any equation, a differential equation can be written in the form F = 0, where F
is a function. For an ODE, the function F can depend on the independent variable
(usually called x or t), the unknown function (usually y), and any derivatives of the
unknown function up to the order of the equation. For instance, an nth order ODE
can be written in the form


F x, y, y, y,..., y ( n)   0 (2.4)

An important special class of differential equations is those that are linear. An nth
order linear ODE has the form

an ( x) y ( n ) ( x)  an 1 ( x) y ( n 1) ( x)  ...  a2 ( x) y( x) 

(2.5)
a1 ( x) y( x)  a0 ( x) y( x)  f ( x)

Each term in the expression on the left side is the product of a coefficient that is a
function of x, and a second factor that is either y or one of the derivatives of y. The
term on the right does not depend on y; it is called the nonhomogeneous term.

2
Observe that no term on the left side involves any power of y or its derivatives other
than the first power, and y and its derivatives are never multiplied together. A linear
ODE is said to be homogeneous if all of its terms involve the unknown function y,
that is, if f (x)=0. If f (x) is not identically zero, the equation is nonhomogeneous.
The equation

1  x  ddxy  sin x ddxy  4 dy

3 2
2
3
y 2
 0 (2.6)
dx

is a linear equation of order 3.

The coefficients are: a3 ( x)  1  x2 , a2 ( x)  sin x, a1 ( x)  4 , and a0 ( x)  1 . Since f ( x)  0 ,

this equation is homogeneous.

The following theorem states that any linear combination of solutions of a linear,
homogeneous DE is also a solution. This is an extremely important fact about linear,
homogeneous DEs.

an y ( n )  an 1 y ( n 1)  ...  a2 y 

(2.7)
a1 y  a0 y  0

then so is the linear combination

y  A y1 ( x)  B y2 ( x) (2.8)

for any values of the constants A and B.

Proof: from the given solutions y1 ( x) and y2 ( x) we have

( n 1)  
 an 1 y1  ...  a2 y1  a1 y1  a0 y1  0 and
( n)
an y1
(2.9)
( n 1)  
an y2  an 1 y2  ...  a2 y2  a1 y2  a0 y2  0
( n)

Multiplying the first equation by A and the second by B and adding the two gives

3
 ( n 1) ( n 1)  
 By2 )  an 1 ( Ay1  By2 )  ...  a2 ( Ay1  By2 ) 
(n) (n)
an ( Ay1
(2.10)
 
a1 ( Ay1  By2 )  a0 ( Ay1  By2 )  0

Thus y  A y1 ( x)  B y2 ( x) is also a solution of the equation

The same kind of proof can be used to verify the following theorem.

Theorem: If y  y1 ( x) is a solution of the linear, homogeneous equation given as

an y ( n)  an 1 y ( n 1)  ...  a2 y  a1 y  a0 y  0 (2.11)
and y  y2 ( x) is a solution of the linear, nonhomogeneous equation
an y ( n)  an 1 y ( n 1)  ...  a2 y  a1 y  a0 y  f ( x) (2.12)
then y  y1 ( x)  y2 ( x) is also a solution of the same linear, nonhomogeneous equation.
Exercises: Find two values of r such that y  erx is a solution of y  y  2 y  0 . Then
find a solution of the equation that satisfies y(0)  1 and y(0)  2 .

2.2 First-Order Separable and Homogeneous Equations

Solving differential equations typically involves integration; indeed, the process of

solving a DE is called integrating the DE. Nevertheless, solving DEs is usually more
complicated than just writing down an integral and evaluating it. The only kind of
DE that can be solved that way is the simplest kind of first order, linear DE that can
be written in the form
dy
 f (x) (2.13)
dx

The solution is then just the antiderivative of f :

y   f ( x)dx (2.14)
Separable Equations

A separable equation is of the form described below is one of the simplest equation
to be solved.

4
 dy
 f ( x) g ( y ) (2.15)
dx
This equation can be rewrite in an equivalent differential form as
dy
 f ( x)dx (2.16)
g ( y)

where the variables have been separated on opposite sides of the equation. This was
possible because the derivative dy / dx was a product of a function of x alone times a
function of y alone, rather than a more general function of the two variables x and y.
We solve the separated equation by integrating both sides,
dy
 g ( y) 
 f ( x)dx (2.17)
Assuming both of the integrals can be evaluated, we can obtain a general solution
to the given DE in the form
G ( y )  F ( x)  C (2.18)
Exercise: solve the following equation
dy  y
 ky 1   (2.19)
dt  L

Solution The equation can be separated to give

Ldy
 kdt (2.20)
y L  y 

and solved by integrating both sides.

1 1 
    dy  kt  C
 y L y
(2.21)

By assuming that 0  y  L , we obtain

ln y  ln ( L  y )  kt  C
 y  (2.22)
ln    kt  C
L y

This can be further solved for y,

5
 C1Lekt
y (2.23)
1  C1e kt
A first-order differential equation of the form
dy  y
 f  (2.24)
dx x
can be transformed into a separable equation (and therefore solved) by means of a
change of dependent variable. If we set
y
v , or equivalently y  xv(x)
x
then we have
dy dv
vx (2.25)
dx dx
and the original differential equation transforms into
dv f (v)  v
 (2.26)
dx x
which is separable.
dy x 2  xy
Exercise: solve for 
dx xy  y 2

Exact Equations and Integrating Factors

A first-order differential equation expressed in differential form as

M ( x, y) dx  N ( x, y) dy  0 (2.27)
dy M ( x, y )
which is equivalent to  , is said to be exact if the left hand side is the
dx N ( x, y )

differential of a function  ( x, y) :
d ( x, y)  M ( x, y) dx  N ( x, y) dy (2.28)
The function  is called an integral function of the differential equation. The level
curves  ( x, y)  C of  are the solution curves of the differential equation
A necessary condition for the exactness of the DE M ( x, y) dx  N ( x, y) dy  0 is that

6
 M N
 ; (2.29)
y x

 2  2
this just means that the mixed partial derivatives and of the integral
xy yx

function  must be equal.

Integrating Factors

Any ordinary differential equation of order 1 can be expressed in differential form:

M ( x, y) dx  N ( x, y) dy  0 . However, this latter equation will usually not be exact. It

may be possible to multiply the equation by an integrating factor u( x, y) so that the

resulting equation
u( x, y)M ( x, y) dx  u( x, y) N ( x, y) dy  0 (2.30)
is exact. In general such integrating factors are difficult to find; they must satisfy the
partial differential equation
u u  N M 
M ( x, y)  N ( x, y)  u ( x, y )    (2.31)
y x  x y 
which follows from the necessary condition for exactness stated above.
Exercise: find an integrating factor of the form u(x,y) for the equation
 y2   x sin x 
 x cos x   dx    y  dy  0
 x   y 

First-Order Linear Equations

A first-order linear differential equation is one of the type
dy
 p ( x) y  q ( x) (2.32)
dx
where p(x) and q(x) are given functions, which we assume to be continuous. The
equation is homogeneous provided that q(x)=0. In that case, the given linear equation
is separable:

7
 dy
  p( x)dx (2.33)
y

which can be solved by integrating both sides. Nonhomogeneous first-order linear

equations can be solved by the following procedure involving the calculation of an
integrating factor.
Let u(x) be any antiderivative of p(x):
du
u ( x)   p( x)dx, and  p ( x) (2.34)
dx
If y  y(x) satisfies the given equation, then we calculate, using the Product Rule,

dx

d u ( x)

e y ( x )  eu ( x )
dy
dx
 eu ( x )
dy
dx
y ( x)
(2.35)
 dy 
 eu ( x )   p ( x ) y   eu ( x ) q ( x)
 dx 
Therefore
eu ( x ) y ( x)   eu ( x ) q( x)dx
or (2.36)
y ( x)  e  u ( x )  eu ( x ) q( x)dx

dy y
Exercise: solve  1 for x0
dx x
x C
Solution: y( x)   for any constant C
2 x

2.3 Differential Equations of Second Order

The general second-order ordinary differential equation is of the form

 d 2 y dy 
F  2 , , y, x   0 (2.37)
 dx dx 

for some function F of four variables. When such an equation can be solved
explicitly for y as a function of x, the solution typically involves two integrations and
therefore two arbitrary constants. A unique solution usually results from prescribing
the values of the solution y and its first derivative y  dy / dx at a particular point.

8
Such a prescription constitutes an initial-value problem for the second-order
equation.
The most frequently encountered ordinary differential equations arising in
applications are second-order linear equations. The general second-order linear
equation is of the form
d2y dy
a2 ( x) 2
 a1 ( x)  a0 ( x) y  f ( x) (2.38)
dx dx
If f(x)= 0 identically, then we say that the equation is homogeneous. If the
coefficients a2 ( x), a1 ( x) , and a0 ( x) are continuous on an interval and a2 ( x)  0 there,
then the homogeneous equation
d2y dy
a2 ( x) 2
 a1 ( x)  a0 ( x) y  0 (2.39)
dx dx
has a general solution of the form
yh  C1 y1 ( x)  C2 y2 ( x) (2.40)
where y1 ( x) and y2 ( x) are two independent solutions. Whenever one solution, y1 ( x) ,
of a homogeneous linear second-order equation is known, another independent
solution (and therefore the general solution) can be found by substituting
y  v( x) y1 ( x) y into the differential equation. This leads to a first-order, linear,

separable equation for v .

Non-Linear Ordinary Differential Equation

On the other hand, the differential equation of the form

2
d3y  dy  d2y
 4 x    y  ey  0 (2.41)
dx3  
2
dx dx

is not linear (we say it is nonlinear) because the second term involves the square of
a derivative of y, the third term involves the product of y and one of its derivatives,
and the fourth term is not y times a function of x.

9
Assignment: read more information on non-linear differential equations

2.4 Partial Differential Equation

Basic Concepts

An equation involving one or more partial derivatives of an (unknown)

function of two or more independent variables is called a partial differential
equation. The order of the highest derivative is called the order of the equation.
Just as in the case of an ordinary differential equation, we say that a partial
differential equation is linear if it is of the first degree in the dependent variable and
its partial derivatives. If each term of such an equation contains either the dependent
variable or one of its derivatives, the equation is said to be homogeneous; otherwise
it is said to be nonhomogeneous.
Some important linear partial differential equations of the second order are:
𝜕2 𝑢 𝜕2 𝑢
= 𝑐2 (one-dimensional wave equation) (2.42)
𝜕𝑡 2 𝜕𝑥 2
𝜕𝑢 𝜕2 𝑢
= 𝑐2 (one-dimensional heat equation ) (2.43)
𝜕𝑡 𝜕𝑥 2
𝜕2 𝑢 𝜕2 𝑢
2
+ =0 (two-dimensional Laplace equation) (2.44)
𝜕𝑥 𝜕𝑦 2

𝜕2 𝑢 𝜕2 𝑢
+ = 𝑓 (𝑥, 𝑦) (two-dimensional Poisson equation) (2.45)
𝜕𝑥 2 𝜕𝑦 2

𝜕2 𝑢 𝜕2 𝑢 𝜕2 𝑢
2
+ 2
+ =0 (three-dimensional Laplace equation) (2.46)
𝜕𝑥 𝜕𝑦 𝜕𝑧 2

Here c is constant, t is the time, and x, y, z are Cartesian coordinates. Equation (2.45)
(with f ≠ 0) is nonhomogeneous, while the other equations are homogeneous.
A solution of a partial differential equation in some region R of the space of
the independent variables is a function which has all the partial derivatives appearing
in the equation in some domain containing R, and satisfies the equation everywhere
in R. (Often one merely requires that the function is continuous at the boundary of

10
R, has those derivatives in the interior of R, and satisfies the equation in the interior
of R.)
In general, the totality of solutions of a partial differential equation is very
large. For example, the functions
u = x 2 – y2 , u = ex cos y, u = In (x2 +y2), (2.46)
Which are entirely different from each other, are solutions of (2.44). We shall see
later that the unique solution of a partial differential equation corresponding to a
given physical problem will be obtained by the use of additional information arising
from the physical situation. For example, in some cases the values of the required
solution of the problem on the boundary of some domain will be given (“boundary
conditions”); in other cases when the time t is one of the variables, the values of
the solution at t = 0 will be prescribed (“initial conditions”).
We know that if an ordinary differential equation is linear and homogeneous,
then from known solutions further solutions can be obtained by superposition. For a
homogeneous linear partial differential equation the situation is quite similar. In fact,
the following theorem holds.
Fundamental Theorem 1. If u1 and u2 are any solutions of a linear homogeneous
partial differential equation in some region, then
U = c1u1 + c2u2, (2.47)
Where c1 and c2 are any constants, is also a solution of that equation in that region.

11