2. Course Outlines
Week Activity
1 Basic Concepts and Ideas, Differential Equations and their
Classifications
2 Formation of Differential Equations, Initial and Boundary
Conditions, Geometrical Meaning of yโ=f(x,y), Separable
Differential Equations
3 Homogeneous Equations, Differential Equations reducible
to Homogeneous Form, Exact Equations, Integrating
Factors
4 Linear Equations, Bernoulli Equations, Orthogonal
Trajectories of Curves, Homogeneous Linear Equations of
Second Order, Differential Operators
5 Euler-Cauchy Equation, Nonhomogeneous Linear
Equations, Reduction of Order, Existance and Uniqueness
Theory,Wronskian
6 Variation of Parameters
3. 7 Higher Order Differential Equations, Higher Order
Nonhomogeneous Equations, Introduction to Vectors,
Matrices, Eigenvalues
8 Homogeneous Systems with Constant Coefficients, Phase
Plane, Critical Points, Criteria for Critical Points
9 Stability, Nonhomogeneous Linear Systems, Power Series
Method,Theory of Power Series Method, Legendre
Equations
10 Legendre Polynomials, Frobenius Method,
11 Laplace Transforms, Properties of Laplace Transforms,
Table of Some Laplace Transforms, Inverse Transform,
Linearity, Shifting
12 Transforms of Derivative and Integral, Differential
Equations Solution of InitialValue Problems
4. Basic Concepts
Equation: Equations describe the relations between the
dependent and independent variables. An equal sign "=" is
required in every equation.
Differential Equations: Equations that involve dependent
variables and their derivatives with respect to the independent
variables are called differential equations.
Example: The following are differential equations involving the
unknown function y.
๐๐ฆ
๐๐ฅ
+ 2 = 0,
๐2
๐ฆ
๐๐ฅ2 +
๐๐ฆ
๐๐ฅ
= ๐๐ฅ,
๐2๐ฆ
๐๐ฅ2
=
๐๐ฆ
๐๐ก
.
5. Some Other Examples:
Newtonโs Second Law of Motion
Maxwell Equations
Kirchhoffโs Laws
Heat Equation
Wave Equation
Ordinary Differential Equation: Differential equations that
involve only one independent variable are called ordinary
differential equations.
Partial Differential Equation: Differential equations that
involve two or more independent variables are called partial
differential equations.
Examples: Equations 1 and 2 are examples of ordinary
differential equations, since the unknown function y depends
solely on the variable
x. Equation 3 is a partial differential equation, since y depends
on both the independent variables t and x.
6. Order: The order of a differential equation is the highest
derivative that appears in the differential equation.
Degree: The degree of a differential equation is the power of
the highest derivative term.
Linear Equations: A differential equation is called linear if
there are no multiplications among dependent variables and
their derivatives. In other words, all coefficients are functions
of independent variables.
Non-linear: Differential equations that do not satisfy the
definition of linear are non-linear.
8. Linear Homogeneous: A differential equation is homogeneous if
every single term contains the dependent variables or their
derivatives.
๐๐ ๐ก ๐ฆ(๐)
๐ก + ๐๐โ1 ๐ก ๐ฆ(๐โ1)
๐ก + โฏ + ๐1 ๐ก ๐ฆโฒ ๐ก + ๐0 ๐ก ๐ฆ(๐ก)
= 0.
Linear Non-homogeneous: Differential equations which do not
satisfy the definition of homogeneous are considered to be non-
homogeneous.
๐๐ ๐ก ๐ฆ(๐) ๐ก + ๐๐โ1 ๐ก ๐ฆ(๐โ1) ๐ก + โฏ + ๐1 ๐ก ๐ฆโฒ ๐ก + ๐0 ๐ก ๐ฆ(๐ก)
= ๐(๐ก).
9. Solutions
General Solution: Solutions obtained from integrating the
differential equations are called general solutions. The general
solution of a order ordinary differential equation contains
arbitrary constants resulting from integrating times.
Particular Solution: Particular solutions are the solutions
obtained by assigning specific values to the arbitrary constants
in the general solutions.
Singular Solutions: Solutions that can not be expressed by the
general solutions are called singular solutions.
10. Explicit Solution: Any solution that is given in the form y = y
(independent variable).
Examples: ๐ด ๐ = ๐๐2
+ ๐, ๐ฆ ๐ก = ๐โ๐ก
+ ๐ etc.
Implicit Solution: Any solution that isnโt in explicit form.
Examples:๐ฆ2
๐ฅ2
= ๐, sin ๐ฅ๐ฆ = ๐, ๐ฅ๐๐ฆ
= ๐ etc.
11. Conditions
Initial Condition: Constrains that are specified at the initial
point, generally time point, are called initial conditions.
Problems with specified initial conditions are called initial
value problems.
Examples: yโ+y=0 y(0)=1,
zโ=1, z(1)=-1.
Boundary Condition: Constrains that are specified at the
boundary points, generally space points, are called boundary
conditions. Problems with specified boundary conditions are
called boundary value problems.
Examples: zโ+zโ+z=1, z(0)=2, z(2)=-1,
fโ+f=0, f(0)=0, f(๏ฐ/2)=1.
14. First Order Differential Equations
Standard Form:
Standard form for a first-order differential equation in the
unknown function y(x) is
yโ=f(x,y).
Interpretation:
The form shows the slope of the tangent at each point of the
xy-plane. Thus the standard form of the first order differential
equation give the slope field. Solving the equation gives the
original curve family having the slope field expressed by the
differential equation.
15. Separable Equations
Form of Equation: ๐ ๐ฅ ๐๐ฅ = ๐ ๐ฆ ๐๐ฆ.
Method of Solution: Integrate both sides will give the solution
of such type of equations.
17. Homogeneous Equations
(of degree zero)
Form of Equation:
๐๐ฆ
๐๐ฅ
= ๐ ๐ฅ, ๐ฆ
๐ ๐ฅ, ๐ฆ is a function homogeneous of degree zero, i.e.,
๐ ๐ก๐ฅ, ๐ก๐ฆ = ๐ก0๐ ๐ฅ, ๐ฆ = ๐ ๐ฅ, ๐ฆ .
A function homogeneous of degree n can be defined as
๐ ๐ก๐ฅ, ๐ก๐ฆ = ๐ก๐๐ ๐ฅ, ๐ฆ .
Method of Solution: Substituting ๐ฆ = ๐ฃ๐ฅ,
๐๐ฆ
๐๐ฅ
= ๐ฃ + ๐ฅ
๐๐ฃ
๐๐ฅ
,
reduces the equation in form separable in variables v and x.
Integrate both sides and then substituting the value of v will
give the solution.
21. Bernoulli Equations
Form of Equation:
๐๐ฆ
๐๐ฅ
+ ๐ ๐ฅ ๐ฆ = ๐ ๐ฅ ๐ฆ๐
, ๐ โ โ.
Method of Solution: The substitution ๐ง = ๐ฆ1โ๐ transform the
Bernoulliโs equation into a linear equation in z. This linear
equation, after solving and back substituting the value of z
gives the solution of the Bernoulliโs equation.
23. A differential equation
๐ ๐ฅ, ๐ฆ ๐๐ฅ + ๐ ๐ฅ, ๐ฆ ๐๐ฆ = 0.
is exact if there exists a function ๐(๐ฅ, ๐ฆ) such that
๐๐ ๐ฅ, ๐ฆ = ๐ ๐ฅ, ๐ฆ ๐๐ฅ + ๐ ๐ฅ, ๐ฆ ๐๐ฆ.
Test for exactness: If ๐(๐ฅ, ๐ฆ) and ๐(๐ฅ, ๐ฆ) are continuous functions and
have continuous first partial derivatives on some rectangle of the ๐ฅ๐ฆ-
plane, then
๐ ๐ฅ, ๐ฆ ๐๐ฅ + ๐ ๐ฅ, ๐ฆ ๐๐ฆ = 0
is exact if and only if
๐๐(๐ฅ, ๐ฆ)
๐๐ฆ
=
๐๐(๐ฅ, ๐ฆ)
๐๐ฅ
.
How to obtain an exact equation from a function?
๐ ๐ฅ, ๐ฆ = ๐,
๐๐(๐ฅ, ๐ฆ)
๐๐ฅ
๐๐ฅ +
๐๐(๐ฅ, ๐ฆ)
๐๐ฆ
๐๐ฆ = 0,
๐ ๐ฅ, ๐ฆ ๐๐ฅ + ๐ ๐ฅ, ๐ฆ ๐๐ฆ = 0.
Exact Equations
24. Method of Solution
To solve an exact equation, first solve the equations
๐๐(๐ฅ, ๐ฆ)
๐๐ฅ
= ๐ ๐ฅ, ๐ฆ ,
๐๐(๐ฅ, ๐ฆ)
๐๐ฆ
= ๐(๐ฅ, ๐ฆ)
for ๐(๐ฅ, ๐ฆ). The solution to the exact equation is then given implicitly
by
๐(๐ฅ, ๐ฆ) = ๐,
where ๐ represents an arbitrary constant.
25. Examples
Example 1: Make an exact ODE using ๐ข ๐ฅ, ๐ฆ = ๐ฅ + ๐ฅ2๐ฆ3
Sol: If ๐ข(๐ฅ, ๐ฆ) = ๐ฅ + ๐ฅ2๐ฆ3 = ๐, then
๐๐ข = (1 + 2๐ฅ๐ฆ3)๐๐ฅ + 3๐ฅ2๐ฆ2๐๐ฆ = 0
๐ฆโฒ = ๐๐ฆ/๐๐ฅ = โ(1 + 2๐ฅ๐ฆ3)/3๐ฅ2๐ฆ2
26. Example 2:
Solve cos ๐ฅ + ๐ฆ ๐๐ฅ + 3๐ฆ2 + 2๐ฆ + cos ๐ฅ + ๐ฆ ๐๐ฆ = 0.
Sol: Here M = cos ๐ฅ + ๐ฆ , ๐ = 3๐ฆ2 + 2๐ฆ + cos ๐ฅ + ๐ฆ
Since
๐๐
๐๐ฆ
= โ sin ๐ฅ + ๐ฆ ,
๐๐
๐๐ฅ
= โ sin ๐ฅ + ๐ฆ , therefore
equation is exact.
๐ข ๐ฅ, ๐ฆ = cos ๐ฅ + ๐ฆ ๐๐ฅ + ๐ ๐ฆ = sin ๐ฅ + ๐ฆ + ๐ ๐ฆ
๐๐ข
๐๐ฆ
= cos(๐ฅ + ๐ฆ) +
๐๐(๐ฆ)
๐๐ฆ
= 3๐ฆ2 + 2๐ฆ + cos ๐ฅ + ๐ฆ
๐๐(๐ฆ)
๐๐ฆ
= 3๐ฆ2
+ 2๐ฆ โ ๐ ๐ฆ = ๐ฆ3
+ ๐ฆ2
+c
๐ข ๐ฅ, ๐ฆ = cos ๐ฅ + ๐ฆ ๐๐ฅ + ๐ ๐ฆ = sin ๐ฅ + ๐ฆ + ๐ฆ3 + ๐ฆ2+c
27. Integrating Factors
If an equation is inexact (not exact), it is possible to transform
Such equation into an exact differential equation by a judicious
multiplication.
A function ๐ผ(๐ฅ, ๐ฆ) is an integrating factor for an inexact equation if the
equation
๐ผ ๐ฅ, ๐ฆ ๐ ๐ฅ, ๐ฆ ๐๐ฅ + ๐ ๐ฅ, ๐ฆ ๐๐ฆ = 0
is an exact equation.
The famous formulas to obtain an integrating factor are as follows:
If
1
๐(๐ฅ,๐ฆ)
๐๐(๐ฅ,๐ฆ)
๐๐ฆ
โ
๐๐(๐ฅ,๐ฆ)
๐๐ฅ
โก ๐ ๐ฅ , then ๐ผ. ๐น. = ๐ผ ๐ฅ, ๐ฆ = ๐ ๐(๐ฅ)๐๐ฅ
.
If
1
๐(๐ฅ,๐ฆ)
๐๐(๐ฅ,๐ฆ)
๐๐ฆ
โ
๐๐(๐ฅ,๐ฆ)
๐๐ฅ
โก ๐ ๐ฆ , then ๐ผ. ๐น. = ๐ผ ๐ฅ, ๐ฆ = ๐โ ๐(๐ฆ)๐๐ฆ
.
If M = yf xy , N = xg(xy), then ๐ผ. ๐น. = ๐ผ ๐ฅ, ๐ฆ =
1
๐ฅ๐โ๐ฆ๐
.
30. OrthogonalTrajectories
Given a one-parameter family of curves ๐น(๐ฅ, ๐ฆ, ๐) = 0.
A curve that intersects each member of the family at right angles
(orthogonally) is called an orthogonal trajectory of the family.
The one-parameter families ๐น(๐ฅ, ๐ฆ, ๐) = 0 and ๐บ(๐ฅ, ๐ฆ, ๐) = 0 are
orthogonal trajectories if each member of one family is an orthogonal
trajectory of the other family.
A procedure for finding a family of orthogonal trajectories ๐บ(๐ฅ, ๐ฆ, ๐) = 0
for a given family of curves ๐น(๐ฅ, ๐ฆ, ๐) = 0 is as follows:
Step 1. Determine the differential equation for the given family ๐น(๐ฅ, ๐ฆ, ๐) =
0.
Step 2. Replace ๐ฆ0 in that equation by โ1/๐ฆ0; the resulting equation is the
differential equation for the family of orthogonal trajectories.
Step 3. Find the general solution of the new differential equation. This is
the family of orthogonal trajectories.
31. Example
Find family of curves orthogonal to one parameter family of
quadratic parabolas ๐ฆ = ๐๐ฅ^2.
Solution:
๐ฆ = ๐๐ฅ2
๐น ๐ฅ, ๐ฆ, ๐ = ๐ฆ โ ๐๐ฅ2 = 0
๐ฆ
๐ฅ2 = ๐ โ
๐ฅ2
๐ฆโฒ
โ 2๐ฅ๐ฆ
๐ฅ4 = 0 โ ๐ฆโฒ
๐๐๐ =
2๐ฆ
๐ฅ
.
Now the slope of the family of new curves is
๐ฆโฒ
๐๐๐ค
=
โ1
๐ฆโฒ
๐๐๐
=
โ1
2๐ฆ
๐ฅ
=
โ๐ฅ
2๐ฆ
.
2๐ฆ๐ฆโฒ = โ๐ฅ
2 ๐ฆ๐ฆโฒ๐๐ฆ = โ ๐ฅ๐๐ฅ
๐ฆ2
= โ
๐ฅ2
2
+ ๐
is required family of curves.
32. Exercises
1. Find the family of orthogonal trajectories of:
๐ฆ = ๐ถ๐ฅ2
+ 2
Answer: ๐ฅ2
+ 2๐ฆ2
โ 8๐ฆ = ๐ถ
2. Find the orthogonal trajectories of the family of parabolas with
vertical axis and vertex at the point (โ1, 3).
Answer: (๐ฅ + 1)2+2(๐ฆ โ 3)2 = ๐ถ
33. Linear Second Order DEs
The most general linear second order differential equation
is in the form.
๐ ๐ก ๐ฆโฒโฒ + ๐ ๐ก ๐ฆโฒ + ๐ ๐ก ๐ฆ = ๐ ๐ก .
The constant coefficient linear second order differential
equation is
๐๐ฆโฒโฒ
+ ๐๐ฆโฒ
+ ๐๐ฆ = ๐(๐ก)
where a, b, c are all constants.
Initially we will make our life easier by looking at differential
equations with ๐ ๐ก = 0. When ๐ ๐ก = 0 we call the
differential equation homogeneous and when ๐ ๐ก โ 0 we
call the differential equation nonhomogeneous.
34. Solving Second Order, Linear,
Homogeneous ODE with Constant
Coefficients
Consider the solution of the type ๐ฆ ๐ก = ๐๐๐ก
. Substituting in
๐๐ฆโฒโฒ
+ ๐๐ฆโฒ
+ ๐๐ฆ = 0
We have
๐ ๐2๐๐๐ก + ๐ ๐๐๐๐ก + ๐๐๐๐ก = 0
๐๐2
+ ๐๐ + ๐ = 0
As ๐๐๐ก โ 0.
This equation is typically called the characteristic equation
This will be a quadratic equation and so we should expect two
roots, r1 and r2. Once we have these two roots we have two
solutions to the differential equation.
๐ฆ1 ๐ก = ๐๐1๐ก and ๐ฆ2 ๐ก = ๐๐2๐ก.
35. The roots will have three possible forms. These are
Real, distinct roots, ๐1 โ ๐2 .
Complex root, ๐1 = ๐ + ๐๐, ๐2 = ๐ โ ๐๐ .
Double roots, ๐1 = ๐2 = ๐.
Real Roots
Example: Find two solutions to ๐ฆโฒโฒ โ 9๐ฆ = 0.
Solution:The characteristic equation is
๐2
โ 9 = 0 โ ๐ = ยฑ3.
The two roots are 3 and -3.Therefore, two solutions are
๐ฆ1 ๐ก = ๐3๐ก
and ๐ฆ2 ๐ก = ๐โ3๐ก
.
The general solution is ๐ฆ๐ ๐ก = ๐1๐3๐ก + ๐2๐โ3๐ก
36. Example: Solve the following IVP ๐ฆโฒโฒ + 11๐ฆโฒ + 24๐ฆ =
0, ๐ฆ 0 = 0, ๐ฆโฒ 0 = โ7.
Solution:The characteristic equation is ๐2 + 11๐ + 24 =
0 โ ๐ + 8 ๐ + 3 = 0 โ ๐ = โ8, โ3.
The general solution and its derivative is
๐ฆ๐ ๐ก = ๐1๐โ3๐ก + ๐2๐โ8๐ก
๐ฆโฒ๐ ๐ก = โ3๐1๐โ3๐ก โ 8๐2๐โ8๐ก
Putting the initial conditions, we have the following system
of equations
0 = ๐ฆ๐ 0 = ๐1 + ๐2
โ7 = ๐ฆโฒ๐ 0 = โ3๐1 โ 8๐2
Solving we get ๐1 =
โ7
5
and ๐2 =
7
5
.Thus the solution is
๐ฆ๐ ๐ก =
โ7
5
๐โ3๐ก +
7
5
๐โ8๐ก
37. Complex Roots
Example: Solve the following IVP ๐ฆโฒโฒ
โ 4๐ฆโฒ
+ 9๐ฆ = 0, ๐ฆ 0 =
0, ๐ฆโฒ
0 = โ8.
Solution:The characteristic equation is ๐2
โ 4๐ + 9 = 0 โ ๐1,2 =
2 ยฑ 5๐.
The general solution and its derivative is
๐ฆ๐ ๐ก = ๐1๐2๐ก
cos 5๐ก + ๐2๐2๐ก
sin 5๐ก
Putting the initial conditions, we have the following system of
equations
0 = ๐ฆ๐ 0 = ๐1
so
๐ฆ๐ ๐ก = ๐2๐2๐ก
sin 5๐ก
๐ฆโฒ๐ ๐ก = 2๐2๐2๐ก
sin 5๐ก + 5๐2๐2๐ก
cos 5๐ก
๐ฆโฒ๐ 0 = 5๐2 โ โ8 = 5๐2 โ ๐2 =
โ8
5
Thus the solution is ๐ฆ ๐ก =
โ8
5
๐2๐ก sin 5๐ก
38. Repeated Roots
Example: Solve the following IVP ๐ฆโฒโฒ
โ 4๐ฆโฒ
+ 4๐ฆ = 0, ๐ฆ 0 =
12, ๐ฆโฒ
0 = โ3.
Solution:The characteristic equation is ๐2
โ 4๐ + 4 = 0 โ
๐ โ 2 2 = 0 โ ๐ = 2,2.
The general solution and its derivative is
๐ฆ๐ ๐ก = ๐1๐2๐ก + ๐2๐ก๐2๐ก
๐ฆโฒ๐ ๐ก = 2๐1๐2๐ก + ๐2๐2๐ก + 2๐2๐ก๐2๐ก
Putting the initial conditions, we have the following system of
equations
12 = ๐ฆ๐ 0 = ๐1
โ3 = ๐ฆโฒ
๐
0 = 2๐1 + ๐2
Solving we get ๐1 = 12 and ๐2 = โ27.Thus the solution is ๐ฆ ๐ก =
12๐2๐ก
โ 27๐ก๐2๐ก
42. Solutions of Linear Homogeneous
Equations; theWronskian
Theorem . (Existence and UniquenessTheorem)
Consider the initial value problem
๐ฆโฒโฒ + ๐(๐ก)๐ฆโฒ + ๐(๐ก)๐ฆ = ๐(๐ก), ๐ฆ(๐ก0) = ๐ฆ0, ๐ฆโฒ(๐ก0) = ๐ฆโฒ0
If ๐(๐ก), ๐(๐ก) and ๐(๐ก) are continuous and bounded on an open
interval ๐ผ containing ๐ก0, then there exists exactly one solution
๐ฆ(๐ก) of this equation, valid on ๐ผ.
43. Wronskian: Given two functions ๐(๐ก), ๐(๐ก), the Wronskian is
defined as
๐(๐, ๐)(๐ก) = ๐๐โฒ โ ๐โฒ๐
Remark: One way to remember this definition could be using the
determinant,
๐(๐, ๐)(๐ก) =
๐ ๐
๐โฒ ๐โฒ
Main property of the Wronskian:
โข If ๐(๐, ๐) โก 0, then ๐ anf ๐ are linearly dependent.
โข Otherwise, they are linearly independent.
44. Example: Check if the given pair of functions are linearly
dependent or not.
(a) ๐(๐ก) = ๐๐ก, ๐(๐ก) = ๐โ๐ก
(b) ๐(๐ก) = sin๐๐ก, ๐(๐ก) = cos๐๐ก
(c) ๐(๐ก) = ๐ก + 1, ๐(๐ก) = 4๐ก + 4
(d) ๐(๐ก) = 2๐ก, ๐(๐ก) = |๐ก|
Suppose๐ฆ_1 (t),๐ฆ_2 (t) are two solutions of
๐ฆโฒโฒ + ๐(๐ก)๐ฆโฒ + ๐(๐ก)๐ฆ = 0
Then
(I)We have either W(๐ฆ_1,๐ฆ_2)โก0 orW(๐ฆ_1,๐ฆ_2) never zero;
(II) IfW(๐ฆ_1,๐ฆ_2)โ 0, then y =๐_1 ๐ฆ_1+๐_2 ๐ฆ_2is the general
solution.They are also called to form a fundamental set of
solutions. As a consequence, for any Ics ๐ฆ(๐ก_0) = ๐ฆ_0, ๐ฆโฒ(๐ก_0)=ใ
๐ฆโฒใ_0, there is a unique set of (๐_1,๐_2) that give a unique
solution.
45. Theorem (AbelโsTheorem):Let y1, y2 be two (linearly independent)
solutions to ๐ฆโฒโฒ + ๐(๐ก)๐ฆโฒ + ๐(๐ก)๐ฆ = 0 on an open interval I.Then, the
Wronskian ๐(๐ฆ1, ๐ฆ2) on I is given by
๐(๐ฆ1, ๐ฆ2)(๐ก) = ๐ถ ยท exp( โ๐(๐ก) ๐๐ก),
for some constant ๐ถ depending on ๐ฆ1, ๐ฆ2, but independent on ๐ก in ๐ผ.
46. Example: Given
๐ก2๐ฆโฒโฒ โ ๐ก(๐ก + 2)๐ฆโฒ + (๐ก + 2)๐ฆ = 0.
Find ๐(๐ฆ1, ๐ฆ2) without solving the equation.
Answer.We first find the ๐(๐ก)
p(t) = โ(t + 2)t
which is
which is valid for ๐ก โ 0. By AbelโsTheorem, we have
๐(๐ฆ1, ๐ฆ2)(๐ก) = ๐ถ ยท exp( โ๐(๐ก) ๐๐ก)=๐ถ ยท exp ๐ก ๐ก + 2 ๐๐ก =
C. ๐๐ก+2 ln ๐ก = ๐ก2๐ถ๐๐ก
Example: If y1, y2 are two solutions of
tyโฒโฒ + 2yโฒ + tety = 0,
andW(y1, y2)(1) = 2, findW(y1, y2)(5).
Example 5. IfW(f, g) = 3e4t, and f = e2t, find g.
47. Example: For the equation
2๐ก2๐ฆโฒโฒ + 3๐ก๐ฆโฒ โ ๐ฆ = 0, ๐ก > 0
given one solution ๐ฆ1 =
1
๐ก
, find a second linearly independent solution.
Solution: Use AbelโsTheorem and Wronskian. By Abelโs
Theorem, and choose C = 1, we have
๐(๐ฆ1, ๐ฆ2)(๐ก) = ๐ถ ยท exp( โ๐(๐ก) ๐๐ก)=๐ถ ยท exp โ
3๐ก
2๐ก2 ๐๐ก = ๐กโ3/2.
By definition of theWronskian,
๐ ๐ฆ1, ๐ฆ2 = ๐ฆ1๐ฆ2
โฒ โ ๐ฆ1
โฒ๐ฆ2 =
๐ฆ2
โฒ
๐ก
+
๐ฆ2
๐ก2
= ๐กโ3/2
Solve this for ๐ฆ2 (taking c=0): ๐ฆ2 =
2
3
๐ก
Example: Consider the equation
๐ก2
๐ฆโฒโฒ โ ๐ก(๐ก + 2)๐ฆโฒ + (๐ก + 2)๐ฆ = 0, ๐ก > 0
Given ๐ฆ1 = ๐ก, find the general solution.
Example: Given the equation ๐ก2
๐ฆโฒโฒ
โ (๐ก โ
3
16
) ๐ฆ = 0, ๐ก > 0
and ๐ฆ1 = ๐ก1/4
๐2 ๐ก
, find ๐ฆ2.
48. Linear Differential Equations
An nth-order linear differential equation has the form
(4.1)
where g(x) and the coefficients bj(x) ( j = 0,1,2,..., n) depend
solely on the variable x. In other words, they do
not depend on y or any derivative of y.
If g(x) = 0, then above Equation is homogeneous; if not, 4.1 is
nonhomogeneous.
A linear differential equation has constant coefficients if all
the coefficients bj(x) in above equation are constants; if one or more
of these coefficients
is not constant, the above equation has variable coefficients.
49. Theorem 4.1. Consider the initial-value problem given by the
linear differential
equation 4.1 and the n initial conditions
y(x0)=c0, yโ(x0)=c1, yโโ(x0)=c2,โฆ, y(n-1)(x0)=cn-1
50. The general solution to the linear differential equation ๐ฟ(๐ฆ) =
๐(๐ฅ), ๐ฆ = ๐ฆโ + ๐ฆ๐ where ๐ฆ๐ denotes one solution to the
differential equation and ๐ฆโ is the general solution to the
associated homogeneous equation, ๐ฟ(๐ฆ) = 0 . Methods for
obtaining ๐ฆโ when the differential equation has constant
coefficients are given in previous lectures. In this lecture, we give
methods for obtaining a particular solution ๐ฆ๐ once ๐ฆโ is known.
Second Order Non-homgeneous
Linear Differential Equations
51. Method of UndeterminedCoefficients
This method is applicable only if ๐(๐ฅ) and all of its derivatives can be
written in terms of the same finite set of linearly independent functions,
which we denote by {๐ฆ1(๐ฅ), ๐ฆ2(๐ฅ), โฆ , ๐ฆ๐(๐ฅ)}. The method is initiated by
assuming a particular solution of the form
๐ฆ๐(๐ฅ) = ๐ด1๐ฆ1(๐ฅ) + ๐ด2๐ฆ2(๐ฅ) + โฏ + ๐ด๐๐ฆ๐(๐ฅ),
where ๐ด1, ๐ด2, โฆ , ๐ด๐ denote arbitrary multiplicative constants. These
arbitrary constants are then evaluated by substituting the proposed
solution into the given differential equation and equating the
coefficients of like
terms.
52. ๐(๐) = ๐๐(๐), an nth-degree polynomial in ๐.
Assume a solution of the form
๐ฆ๐ = ๐๐๐ฅ(๐ด๐๐ฅ๐ + ๐ด๐ โ 1๐ฅ๐ โ 1 + โฏ +
๐ด2๐ฅ2 + ๐ด1๐ฅ + ๐ด0)
where ๐ด๐ (๐ = 0,1,2, โฆ , ๐) is a constant to be determined.
๐(๐) = ๐๐๐๐ where ๐ and ๐ are known constants.
Assume a solution of the form
๐ฆ๐ = ๐ด๐๐๐ฅ
where ๐ด is a constant to be determined.
๐(๐) = ๐๐๐๐๐๐๐ + ๐๐๐๐๐๐๐ where ๐๐, ๐๐, and ๐ are known
constants.
Assume a solution of the form
๐ฆ๐ = ๐ด sin ๐๐ฅ + ๐ต cos ๐๐ฅ,
where ๐ด and ๐ต are constants to be determined.
53. Generalizations
If ๐(๐ฅ) is the product of terms considered in all the cases given above,
take ๐ฆ๐ to be the product of the corresponding assumed solutions and
algebraically
combine arbitrary constants where possible. In particular, if ๐(๐ฅ) =
๐๐๐ฅ๐๐(๐ฅ) is the product of a polynomial with an exponential, assume
๐ฆ๐ = ๐๐๐ฅ(๐ด๐๐ฅ๐ + ๐ด๐ โ 1๐ฅ๐ โ 1 + โฏ +
๐ด2๐ฅ2 + ๐ด1๐ฅ + ๐ด0)
where Aj is as in 1st case. If, instead, ๐(๐ฅ) = ๐๐๐ฅ๐๐(๐ฅ)๐ ๐๐๐๐ฅ is the
product of a polynomial, exponential, and sine term, or if ๐(๐ฅ) =
๐๐๐ฅ๐๐(๐ฅ)๐๐๐ ๐๐ฅ is the product of a polynomial, exponential, and cosine
term, then assume
๐ฆ๐
= ๐๐๐ฅ(๐ด๐๐ฅ๐ + ๐ด๐ โ 1๐ฅ๐ โ 1 + โฏ +
๐ด2๐ฅ2 + ๐ด1๐ฅ + ๐ด0)๐ ๐๐๐๐ฅ + ๐๐๐ฅ(๐ต๐๐ฅ๐
+ ๐ต๐ โ 1๐ฅ๐ โ 1 + โฏ +
๐ต2๐ฅ2 + ๐ต1๐ฅ + ๐ต0)๐๐๐ ๐๐ฅ
where ๐ด๐ and๐ต๐ (๐ = 0,1,2, โฆ , ๐) are constants which still must be
determined.
If ๐(๐ฅ) is the sum (or difference) of terms already considered, then we
take ๐ฆ๐ to be the sum (or difference) of the corresponding assumed
solutions and algebraically combine arbitrary constants where possible.
54. Modifications
If any term of the assumed solution, disregarding multiplicative
constants, is also a term of
(the homogeneous solution), then the assumed solution must be
modified by multiplying it by ๐ฅ๐, where ๐ is the smallest positive
integer such that the product of ๐ฅ๐ with the assumed solution has no
terms in common with ๐ฆโ.
Limitations of the Method
In general, if ๐(๐ฅ) is not one of the types of functions considered
above, or if the differential equation does not have constant
coefficients, then the method of variations of parameters is preferred.
55. Variation of parameters is another method for finding a particular solution of
the nth-order linear differential equation ๐ณ(๐ฆ) = ๐(๐ฅ), once the solution of the
associated homogeneous equation ๐ณ(๐ฆ) = 0 is known. If ๐ฆ1 ๐ฅ , ๐ฆ2 ๐ฅ , โฆ , ๐ฆ๐ ๐ฅ
are ๐ linearly independent solutions of ๐ณ(๐ฆ) = 0, then the general solution of
๐ณ(๐ฆ) = 0 is
๐ฆโ = ๐1๐ฆ1(๐ฅ) + ๐2๐ฆ2(๐ฅ) + โฏ + ๐๐๐ฆ๐(๐ฅ)
Methodology:
A particular solution of ๐ณ(๐ฆ) = ๐(๐ฅ) has the form
๐ฆ๐ = ๐ฃ1๐ฆ1(๐ฅ) + ๐ฃ2๐ฆ2(๐ฅ) + โฏ + ๐ฃ๐๐ฆ๐(๐ฅ)
where ๐ฆ๐ = ๐ฆ๐ ๐ฅ (๐ = 1,2, โฆ , ๐) is given in above Equation and ๐ฃ๐ (๐ =
1,2, โฆ ๐) is an unknown function of ๐ฅ which still must be determined.
Variation of Parameters
56. To find ๐ฃ๐, first solve the following linear equations simultaneously for ๐ฃโฒ๐:
๐ฃโฒ1๐ฆ1 + ๐ฃโฒ2๐ฆ2 + โฏ + ๐ฃโฒ
๐๐ฆ๐ = 0
๐ฃโฒ1๐ฆโฒ1 + ๐ฃโฒ2๐ฆโฒ2 + โฏ + ๐ฃโฒ
๐๐ฆโฒ๐ = 0
โฎ
๐ฃโฒ1๐ฆ1
(๐โ2)
+ ๐ฃโฒ2๐ฆ2
(๐โ2)
+ โฏ + ๐ฃโฒ
๐๐ฆ๐
(๐โ2)
= 0
๐ฃโฒ1๐ฆ1
(๐โ1)
+ ๐ฃโฒ2๐ฆ2
(๐โ1)
+ โฏ + ๐ฃโฒ
๐๐ฆ๐
๐โ1
= ๐ ๐ฅ .
Then integrate each to obtain ๐ฃ๐, disregarding all constants of integration.
This is permissible because we are seeking only one particular solution.
Example 6.1: For the special case ๐ = 3, Equations reduce to
๐ฃโฒ1๐ฆ1 + ๐ฃโฒ2๐ฆ2 + ๐ฃโฒ
3๐ฆ3 = 0
๐ฃโฒ1๐ฆโฒ1 + ๐ฃโฒ2๐ฆโฒ2 + ๐ฃโฒ
3๐ฆโฒ3 = 0
๐ฃโฒ1๐ฆ1โฒโฒ + ๐ฃโฒ2๐ฆ2โฒโฒ + ๐ฃโฒ
3๐ฆ3โฒโฒ = ๐ ๐ฅ .
For the case ๐ = 2, Equations become
๐ฃโฒ1๐ฆ1 + ๐ฃโฒ2๐ฆ2 = 0
๐ฃโฒ1๐ฆโฒ1 + ๐ฃโฒ2๐ฆโฒ2 = ๐ ๐ฅ .
and for the case ๐ = 1, we obtain the single equation
๐ฃโฒ1๐ฆ1 = ๐ ๐ฅ .
57. Since ๐ฆ1 ๐ฅ , ๐ฆ2 ๐ฅ , โฆ , ๐ฆ๐ ๐ฅ are ๐ linearly independent solutions of the same
equation L(y)=0, theirWronskian is not zero.This means that the system 6.9
has a nonzero determinant and can be solved uniquely for ๐ฃโฒ๐.
Scope of the Method
The method of variation of parameters can be applied to all linear differential
equations. It is therefore more powerful than the method of undetermined
coefficients, which is restricted to linear differential equations with constant
coefficients and particular forms of ๐ ๐ฅ .
Nonetheless, in those cases where both methods are applicable, the method
of undetermined coefficients is usually the more efficient and, hence,
preferable.
As a practical matter, the integration of may be impossible to perform. In
such an event other methods (in particular, numerical techniques) must be
employed.
58. Initial-Value Problems
Initial-value problems are solved by applying the initial conditions
to the general solution of the differential equation. It must be
emphasized that the initial conditions are applied only to the
general solution and not to the homogeneous solution ๐ฆโ that
possesses all the arbitrary constants that must be evaluated.The
one exception is when the general solution is the homogeneous
solution; that is, when the differential equation under consideration
is itself homogeneous.
59. Example: Solve ๐ฆโฒโฒ โ ๐ฆโฒ โ 2๐ฆ = 4๐ฅ2.
Solution: ๐ฆโ = ๐1๐2๐ฅ + ๐2๐โ๐ฅ. Here ๐ ๐ฅ = 4๐ฅ2, a second degree polynomial.We
assume that
๐ฆ๐ = ๐ด2๐ฅ2 + ๐ด1๐ฅ + ๐ด0
Thus ๐ฆโฒ๐ = 2๐ด2๐ฅ + ๐ด1 and ๐ฆโฒโฒ๐ = 2๐ด2. Substituting these results into the
differential equation, we have
2๐ด2 โ (2๐ด2๐ฅ + ๐ด1) โ 2(๐ด2๐ฅ2
+ ๐ด1๐ฅ + ๐ด0) = 4๐ฅ2
.
Equating the coefficients of like powers of ๐ฅ, we obtain
โ2๐ด2 = 4, โ2๐ด2 โ 2๐ด1 = 0,2๐ด2 โ ๐ด1 โ 2๐ด0 = 0
Solving this system, we find that ๐ด2 = โ2, ๐ด1 = 2, ๐ด0 = โ3.
Hence ๐ฆ๐ = โ2๐ฅ2
+ 2๐ฅ โ 3 and the general solution is
๐ฆ๐ = ๐ฆโ + ๐ฆ๐ = ๐1๐2๐ฅ + ๐2๐โ๐ฅ โ 2๐ฅ2 + 2๐ฅ โ 3
60. Example: Solve ๐ฆโฒโฒ
โ ๐ฆโฒ
โ 2๐ฆ = sin(2๐ฅ) .
Solution:Again ๐ฆโ = ๐1๐2๐ฅ
+ ๐2๐โ๐ฅ
.
Assume that ๐ฆ๐ = ๐ด sin 2๐ฅ + ๐ต cos 2๐ฅ .
Thus, ๐ฆโฒ
๐ = 2๐ด cos 2๐ฅ โ 2๐ต sin 2๐ฅ and ๐ฆโฒโฒ
๐ = โ4๐ด sin 2๐ฅ โ
4๐ต cos 2๐ฅ Substituting these results into the differential equation, we
have (โ6๐ด + 2๐ต) sin 2๐ฅ + (โ2๐ด โ 6๐ต) cos 2๐ฅ = sin(2๐ฅ)
Equating coefficients of like terms, we obtain
โ6๐ด + 2๐ต = 1, โ2๐ด โ 6๐ต = 0
Solving this system, we find that ๐ด = โ3/20 and ๐ต = 1/20.Then
๐ฆ๐ = โ
3
20
sin 2๐ฅ +
1
20
cos 2๐ฅ .
and the general solution is
๐ฆ๐ = ๐ฆโ + ๐ฆ๐ = ๐1๐2๐ฅ + ๐2๐โ๐ฅ โ
3
20
sin 2๐ฅ +
1
20
cos 2๐ฅ .
61. Example: Solve ๐ฆโฒโฒโฒ
+ ๐ฆโฒ
= sec ๐ฅ .
This is a third-order equation with ๐ฆโ = ๐1 + ๐2 cos ๐ฅ + ๐3 sin ๐ฅ .
It follows that
๐ฆ๐ = ๐ฃ1 + ๐ฃ2 cos ๐ฅ + ๐ฃ3 sin ๐ฅ .
So the set of Equations becomes
๐ฃโฒ1 + ๐ฃโฒ2 cos ๐ฅ + ๐ฃโฒ
3 sin ๐ฅ = 0
๐ฃโฒ2 sin ๐ฅ + ๐ฃโฒ
3 cos ๐ฅ = 0
โ๐ฃโฒ
2 cos ๐ฅ โ ๐ฃโฒ
3 sin ๐ฅ = sec ๐ฅ .
Solving this set of equations simultaneously, we obtain
๐ฃโฒ1 = sec ๐ฅ , ๐ฃโฒ2 = โ1, ๐ฃโฒ
3 = โ tan ๐ฅ .
Thus ๐ฃ1 = ln sec ๐ฅ + tan ๐ฅ , ๐ฃ2 = โ๐ฅ, ๐ฃ3 = ln cos ๐ฅ ,
Gives ๐ฆ๐ = ln sec ๐ฅ + tan ๐ฅ โ ๐ฅ cos ๐ฅ + ln cos ๐ฅ sin ๐ฅ .
The general solution is therefore
๐ฆ๐ = ๐ฆโ + ๐ฆ๐
= ๐1 + ๐2 cos ๐ฅ + ๐3 sin ๐ฅ + ln sec ๐ฅ + tan ๐ฅ โ ๐ฅ cos ๐ฅ + ln cos ๐ฅ sin ๐ฅ .
62. LaplaceTransform
Let ๐(๐ฅ) be defined for 0 โค ๐ฅ < โ and let ๐ denote an arbitrary real variable.
The Laplace transform of ๐(๐ฅ), designated by either โ{๐(๐ฅ)} or ๐น(๐ ),is
โ ๐ ๐ฅ = ๐น ๐ =
0
โ
๐โ๐ ๐ฅ
๐(๐ฅ)๐๐ฅ
for all values of ๐ for which the improper integral converges. Convergence
occurs when the limit
lim
๐ โโ 0
๐
๐โ๐ ๐ฅ
๐(๐ฅ)๐๐ฅ
exists. If this limit does not exist, the improper integral diverges and ๐(๐ฅ) has no
Laplace transform. When evaluating the integral, the variable ๐ is treated as a
constant because the integration is with respect to ๐ฅ.
The Laplace transforms for a number of elementary functions can be found in
Annexure of the Book.
63. โบ (Linearity). If โ ๐ ๐ฅ = ๐น ๐ and โ ๐ ๐ฅ = ๐บ ๐ , then
for any two constants ๐1 and ๐2
โ ๐1๐ ๐ฅ + ๐2๐ ๐ฅ = ๐1โ ๐ ๐ฅ + ๐2โ ๐ ๐ฅ = ๐1๐น(๐ ) + ๐2๐บ(๐ )
โบ If โ ๐ ๐ฅ = ๐น(๐ ), then for any constant ๐, โ ๐๐๐ฅ๐ ๐ฅ = ๐น(๐ โ ๐)
โบ If โ ๐ ๐ฅ = ๐น(๐ ), then for any positive integer ๐, โ ๐ฅ๐
๐ ๐ฅ = โ1 ๐ ๐๐
๐๐ ๐ [๐น ๐ ]
โบ If โ ๐ ๐ฅ = ๐น(๐ ) and if lim
๐ฅโ0
๐ฅ>0
๐(๐ฅ)
๐ฅ
exists, then โ
1
๐ฅ
๐(๐ฅ) = ๐
โ
๐น(๐ก) ๐๐ก
โบ If โ ๐ ๐ฅ = ๐น(๐ ), then โ 0
๐ฅ
๐(๐ก) ๐๐ก =
1
๐
๐น(๐ )
โบ If ๐(๐ฅ) is periodic with period ๐ค, that is, ๐ ๐ฅ + ๐ค = ๐(๐ฅ), then
โ ๐ ๐ฅ = 0
๐ค
๐โ๐ ๐ฅ
๐(๐ฅ) ๐๐ฅ
1 โ ๐โ๐ค๐
Properties of LaplaceTransforms
64. Inverse LaplaceTransform
An inverse Laplace transform of ๐น(๐ ) designated by โโ1{๐น(๐ )}, is another
function ๐(๐ฅ) having the property that โ ๐ ๐ฅ = ๐น(๐ ).
Methodology
The simplest technique for identifying inverse Laplace transforms is to
recognize them, either from memory or from a table.
If ๐น(๐ ) is not in a recognizable form, then occasionally it can be transformed
into such a form by algebraic manipulation.
65. Manipulating Denominators
-The method of completing the square deals with quadratic polynomials
-The method of partial fractions
Manipulating Numerators
A factor ๐ โ ๐ in the numerator may be written in terms of the factor ๐ โ ๐,
where both ๐ and ๐ are constants, through the identity ๐ โ ๐ = ๐ โ ๐ +
(๐ โ ๐).
โฅ (Linearity). If the inverse Laplace transforms of two functions ๐น(๐ ) and
๐บ(๐ ) exist, then for any constants ๐1 and๐2,
โโ1 ๐1๐น ๐ + ๐2๐บ ๐ = ๐1โโ1{๐น ๐ } + ๐2โโ1{๐บ ๐ }
66. Convolution
The convolution of two functions ๐(๐ฅ) and ๐(๐ฅ) is
๐ ๐ฅ โ ๐ ๐ฅ =
0
๐ฅ
๐(๐ก)๐(๐ฅ โ ๐ก)๐๐ก
Theorem ๐ ๐ฅ โ ๐ ๐ฅ = ๐(๐ฅ) โ ๐(๐ฅ).
Theorem. (Convolution Theorem). If โ ๐ ๐ฅ = ๐น(๐ ) and โ ๐ ๐ฅ =
๐บ(๐ ), then
โ ๐ ๐ฅ โ ๐ ๐ฅ = โ ๐ ๐ฅ โ ๐ ๐ฅ = ๐น(๐ )๐บ(๐ )
The inverse Laplace transform of a product is computed using a
convolution.
โโ1 ๐น(๐ )๐บ(๐ ) = ๐ ๐ฅ โ ๐ ๐ฅ = ๐ ๐ฅ โ ๐(๐ฅ)
If one of the two convolutions in above Equation is simpler to calculate,
then that convolution is chosen when determining the inverse Laplace
transform of a product.