Ordinary Differential Equations (Math 2302)

 Õæ…Aë áÖ ß
@ .X . @
A®‚Ë@

2020-2019 úGAJË@ ɒ®Ë@

Chapter 4
Higher Order Linear Equations

In this chapter we generalize the structure and methods of solutions of second order linear equations to
third and higher order linear equations.

4.1 General Theory of nth Order Linear Equations

An nth order linear equation is an equation of the form
dn y dn−1 y dy
L[y] = n
+ p 1 (t) n−1
+ · · · + pn−1 (t) + pn (t)y = g(t),
dt dt dt
where p1 , p2 , . . . , pn , and g are continuous functions on some interval I = (α, β). The linear differential
operator L of order n defined above is similar to the second order operator introduced in Chapter 3.
Example 1. The equation y 000 + ty 00 − y = sin t is a third order linear equation.
Example 2. The equation 5y (4) + sin ty 0 + et y = t is a fourth order linear equation.
Theorem 1. Let
dn y dn−1 y dy
L[y] = n + p1 (t) n−1 + · · · + pn−1 (t) + pn (ty = g(t),
dt dt dt
(n−1)
y(t0 ) = y0 , y 0 (t0 ) = y00 , . . . , y (n−1) (t0 ) = y0
be an initial value problem. If the functions p1 , p2 , . . . , pn , and g are continuous on an interval I = (α, β)
containing the initial point t0 , then the initial value has a unique solution y = ϕ(t) valid for all t ∈ I.
Example. Find the largest interval in which the solution of the initial value problem
(t − 1)y 000 + tan ty 00 + 3y = cos t, y(0) = 1, y 0 (0) = 2, y 00 (0) = 3
is certain to exist.
Solution.

1
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
The Homogeneous Equation
Firstly, we discuss the homogeneous equation
L[y] = 0.
Theorem 2. (Principle of Superposition)
If y1 , y2 , . . . , yn are n solutions of the linear homogeneous equation
L[y] = 0,
then the linear combination ϕ(t) = c1 y1 + c2 y2 + · · · + cn yn is also a solution for any values of the
constants c1 , c2 , . . . , cn .
Theorem (2) helps us to construct the general solution of the linear homogeneous equation
L[y] = 0.
A linear combination ϕ(t) = c1 y1 + c2 y2 + · · · + cn yn is the general solution of L[y] = 0 if and only if it
is possible to choose c1 , c2 , . . . , cn such that ϕ(t) = c1 y1 + c2 y2 + · · · + cn yn satisfies any set of initial
conditions
(n−1)
y(t0 ) = y0 , y 0 (t0 ) = y00 , . . . , y (n−1) (t0 ) = y0 .
This required the condition
 
 y1 (t)
 0 ··· yn (t) 
yn0 (t)

 y1 (t) ··· 
W (y1 , y2 , . . . , yn ) =  .. ..  6= 0.
 
 . . 
 (n−1) 
y1 (t) · · · yn(n−1) (t)
Definition. (The Wronskian)
the Wronskian determinant of the functions f1 , f2 , . . . , fn is the determinant
 
 f1 (t)
 0 ··· fn (t) 
 f1 (t) ··· fn0 (t) 
W (f1 , f2 , . . . , fn ) =  .. .. .

 . . 
 (n−1) 
f1 (t) · · · fn(n−1) (t)
Example. Find the Wronskian of the functions f1 (t) = 1, f2 (t) = t, f3 (t) = cos t, f4 (t) = sin t.
Solution.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 2 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
Theorem 3. Assume that p1 , p2 , . . . , pn are continuous on the open interval I. Let y1 , y2 , . . . , yn be
solutions of the differential equation
dn y dn−1 y dy
L[y] = n + p1 (t) n−1 + · · · + pn−1 (t) + pn (t)y = 0.
dt dt dt
If W (y1 , y2 , . . . , yn )(t0 ) 6= 0 for some point t0 ∈ I, then y = c1 y1 + c2 y2 + · · · + cn yn , where c1 , c2 , . . . , cn ,
are arbitrary constants, is the general solution of L[y] = 0.
Definition. (Fundamental set of solutions)
If y1 , y2 , . . . , yn are solutions of L[y] = 0 with W (y1 , y2 , . . . , yn )(t0 ) 6= 0, then the set {y1 , y2 , . . . , yn } is
called a fundamental set of solutions of L[y] = 0.
Example. Show that y1 = 1, y2 = t and y3 = t3 form a fundamental set of solutions of the equation
ty 000 − y 00 = 0.
Solution.

The Nonhomogeneous Equation

Now we return to the nonhomogeneous equation

L[y] = y (n) + p1 (t)y (n−1) + · · · + pn−1 (t)y 0 + pm (t)y = g(t).

Theorem 4. The general solution of the nonhomogeneous equation L[y] = g(t) is given by

y = c1 y1 + c2 y2 + · · · + cn yn + yp ,

where y1 , y2 , . . . , yn are fundamental solutions of the corresponding homogeneous equation L[y] = 0

and yp is some particular solution of L[y] = g(t).

2020-2019 ú
GA JË@ ɒ®Ë@
Page 3 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
4.2 Homogeneous Equations with Constant Coefficients
In this section we consider the homogeneous constant coefficient equation

L[y] = an y (n) + an−1 y (n−1) + · · · + a0 y = 0, (1)

where an , . . . , a0 are constants and an 6= 0. As in the case of second order, we look for solutions in the
form y = ert . Substituting y into equation (1) yields

L[y] = ert Z(r) = 0.

Since ert 6= 0, it follows that y = ert is a solution of (1) if and only if r is a root of the polynomial Z(r).
Definition. (Characteristic polynomial)
The polynomial Z(r) = an rn +an−1 rn−1 +· · ·+a0 is called the characteristic polynomial of the differential
equation L[y] = an y (n) + an−1 y (n−1) + · · · + a0 y = 0.

An important question in mathematics for more than 200 years was whether every polynomial equation
has at least one root. The affirmative answer to this question, the fundamental theorem of algebra, was
given by Carl Friedrich Gauss in his doctoral dissertation in 1799, although his proof does not meet
modern standards of rigor. Several other proofs have been discovered since, including three by Gauss
himself. Today, students often meet the fundamental theorem of algebra in a first course on complex
variables, where it can be established as a consequence of some of the basic properties of complex
analytic functions.

Theorem 5. (Fundamental Theorem of Algebra)

A polynomial Z(r) = an rn + an−1 rn−1 + · · · + a0 of degree n has exactly n roots, say r1 , . . . , rn , and it
can be written as
Z(r) = an (r − r1 )(r − r2 ) . . . (r − rn ).

We have to distinguish between three case:

(1) All the roots are real and distinct.

(2) Roots are distinct but not all them are real.

(3) Roots are not distinct.

Real and Distinct Roots

If the roots r1 , . . . , rn of the characteristic polynomial Z(r) are real and no two of them are equal, then
the general solution of L[y] = 0 is given by

y = c1 er1 t + · · · + cn ern t .

Finding Roots of a Polynomial

Consider a polynomial Z(r) = an rn + an−1 rn−1 + · · · + a0 . For third and fourth degree polynomi-
als there are formulas, analogous to the formula for quadratic equations but more complicated, that
give exact expressions for the roots. Root-finding algorithms are readily available on calculators and

2020-2019 ú
GA JË@ ɒ®Ë@
Page 4 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
computers. Sometimes they are included in the differential equation solver, so that the process of fac-
toring the characteristic polynomial is hidden and the solution of the differential equation is produced
automatically.
If we need to factor the characteristic polynomial by hand, here is one result that is sometimes helpful.
Assume that an , . . . , a0 are integers.
(1) The polynomial equation Z(r) = 0 has a solution r ∈ Z only if r divides a0 .
p
(2) The polynomial equation Z(r) = 0 has a solution r = ∈ Q, where p and q have no common
q
factors, only if p divides a0 and q divides an .

Example 1. Find the roots of Z(r) = 12r4 + 31r3 + 75r2 + 37r + 5 = 0.

Solution.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 5 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
Example 2. Find the solution of the equation y 000 − 5y 00 + 6y 0 = 0.
Solution.

Example 3. Find the general solution of the equation y 000 − 2y 00 − y 0 + 2y = 0.

Solution.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 6 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
Example 4. Find the general solution of the equation y (3) − 5y 00 − 22y 0 + 56y = 0.

Solution.

Complex Roots
In the case where the coefficients a0 , . . . , an of Z(r) are real, then complex roots of Z(r) occur in
conjugate pairs. If r = λ + iµ is a root, so is r = λ − iµ, and the corresponding solutions are

u1 (t) = e(λ+iµ)t = eλt (cos(µt) + i sin(µt))

and
u2 (t) = e(λ−iµ)t = eλt (cos(µt) − i sin(µt)).
Since a linear combination of solutions of is also a solution, this implies that the real-valued functions
1 1
y1 = eλt cos(µt) = (u1 + u2 ), y2 = eλt sin(µt) = (u1 − u2 )
2 2i
are solutions. Therefore, if all the roots of Z(r) are distinct, then the general solution of L[y] = 0 is
still of the form
y = c1 er1 t + · · · + cn ern t .
However we can replace the complex functions e(λ±iµ)t by the real-valued functions eλt cos(µt) and
eλt sin(µt).

2020-2019 ú
GA JË@ ɒ®Ë@
Page 7 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
Example 1. Solve y (4) − 16y = 0.

Solution.

Example 2. Find the general solution of the equation y (4) + 5y 00 + 4y = 0.

Solution.

Example 3. Find the general solution of the equation y (4) + y (3) + 2y 00 + y 0 + y = 0.

Solution.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 8 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
The equation rn + 1 = 0
To find the n roots of the equation rn + 1 = 0 we write −1 = eiπ(2m+1) . As a result we obtain

rn = eiπ(2m+1)/n , m = 0, 1, . . . , n − 1.

Example 1. Solve r3 + 1 = 0.

Solution.

Example 2. Solve the differential equation y (5) + y = 0.

Solution.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 9 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@

Repeated Roots
If the roots of the characteristic equation are not distinct, that is, if r1 = r2 = · · · = rs , then the
fundamental solutions corresponding to r1 , r2 , . . . , rs are

er1 t , ter1 t , . . . , ts−1 er1 t .

Similarly, if a complex root r = λ + iµ is repeated s times, then the complex conjugate r = λ − iµ is

also repeated s times. The corresponding solutions are

eλt cos(µt), eλt sin(µt), teλt cos(µt), teλt sin(µt), . . . , ts−1 eλt cos(µt), ts−1 eλt sin(µt).

Example 1. Find the general solution of y (4) + 3y (3) + 3y 00 + y 0 = 0.

Solution.

Example 2. Find the solution of the initial value problem

y 000 − y 00 + y 0 − y = 0, y(π/2) = 2, y 0 (π/2) = 1 y 00 (π/2) = 0.

Solution.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 10 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
Example 3. Find the general solution of the equation 2y (4) + 11y (3) + 18y 00 + 4y 0 − 8y = 0.

Solution.

Example 4. Find the general solution of the equation y iv + 2y 00 + y = 0.

Solution.

4.3 The Method of Undetermined Coefficients

The method of undetermined coefficients can be used to obtain a particular solution of the nonhomo-
geneous nth order linear equation with constant coefficients

L[y] = an y (n) + an−1 y (n−1) + · · · + a0 y = g(t) (2)

provided that g(t) is a sum of polynomials, exponentials, sines, and cosines functions.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 11 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
Just as for the second order linear equation, we choose the suitable form for the particular solution
according to the following rules.
(1) If g(t) = eαt (an tn + · · · + a1 t + a0 ) with n = 0, 1, 2, . . . , then yp has the form

yp = ts eαt (An tn + · · · + A1 t + A0 ) ,

where s is the number of the roots of Z(r) that are equal to α.

(2) If g(t) = eαt [A cos(βt) + B sin(βt)] (an tn + · · · + a1 t + a0 ) with n = 0, 1, 2, . . . , then yp has the
form
yp = ts eαt [cos(βt)(An tn + · · · + A1 t + A0 ) + sin(βt)(Bn tn + · · · + B1 t + B0 )] ,
where s is the number of the roots of Z(r) that are equal to α + iβ.
Example 1. Find the general solution of y iv − 4y 00 = t2 + et .
Solution.

Example 2. Solve the initial value problem y iv + 2y 000 + y 00 + 8y 0 − 12y = 12 sin t − 20e−t ,
y(0) = 3, y 0 (0) = 0, y 00 (0) = −1, y 000 (0) = 2.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 12 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
Solution.

Example 3. Find a suitable form for a particular solution of the differential equation
y (4) − y 000 − y 00 + y 0 = t2 + 4et + t sin t.
Do not evaluate the constants.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 13 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
Solution.

Example 4. Find a suitable form for a particular solution of the differential equation
y 000 + y 00 − 4y 0 − 4y = 8t + 8 + 6e−t .
Do not evaluate the constants.
Solution.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 14 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
Example 5. Find a suitable form for a particular solution of the differential equation

y iv + 2y 000 + 2y 00 = 3et + 2te−t + e−t sin t.

Do not evaluate the constants.

Solution.

4.4 The Method of Variation of Parameters

The method of variation of parameters is a general method to find a particular solution of the nonho-
mogeneous nth order linear differential equation

dn y dn−1 y dy
L[y] = n
+ p 1 (t) n−1
+ · · · + pn−1 (t) + pn (t)y = g(t).
dt dt dt
In this method we start by finding the general solution of the corresponding homogeneous equation
L[y] = 0, say
yc = c1 y1 + · · · + cn yn .
Then we look for a particular solution of the nonhomogeneous equation L[y] = g(t) in the form

yp = u1 y1 + · · · + un yn ,

where u1 , . . . , un are functions to be determine. Since we have n functions to determine, we will have to
specify n conditions. One of these is clearly that yp satisfy L[y] = g(t). The other n − 1 conditions are

2020-2019 ú
GA JË@ ɒ®Ë@
Page 15 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
chosen so as to make the calculations as simple as possible. Since we can hardly expect a simplification
in determining yp if we must solve high order differential equations for u1 , . . . , un , it is natural to impose
conditions to suppress the terms that lead to higher derivatives of u1 , . . . , un . These n equations are
u01 y1 + u02 y2 + · · · + u0n yn = 0,
u01 y10 + u02 y20 + · · · + u0n yn0 = 0,
.. (3)
.
(n−1) (n−1)
u01 y1 + u02 y2 + · · · + u0n yn(n−1) = g(t).
The system (3) is a linear algebraic system for the unknown quantities u01 , . . . , u0n . By solving this
system and then integrating the resulting expressions, you can obtain the coefficients u1 , . . . , un . The
system of equations (3) has a solution if the determinant of coefficients is nonzero for each value of
t. However, the determinant of coefficients is precisely W (y1 , y2 , . . . , yn ), and it is nowhere zero since
y1 , . . . , yn are linearly independent solutions of the homogeneous equation L[y] = 0. Hence it is possible
to determine u01 , . . . , u0n . Using Cramer’s rule, we find that the solution of the system of equations (3)
is
g(t)Wm (t)
u0m (t) = , m = 1, 2, . . . , n, (4)
W (t)
where W (t) = W (y1 , y2 , . . . , yn )(t) and Wm is the determinant obtained from W by replacing the mth
column by the column (0, 0.....0, 1). Integrating we obtain
Z
g(t)Wm (t)
um (t) = dt, m = 1, 2, . . . , n. (5)
W (t)
π
Example 1. Find the general solution of y 000 + y 0 = tan t, 0<t< .
2
Solution.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 16 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
Example 2. Solve y 000 − 2y 00 − y 0 + 2y = e4t .
Solution.

Example 3. Consider the equation L[y] = ty 000 + 3y 00 = t2 + 1, t > 0.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 17 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
(a) Show that y1 = 1, y2 = t, and y3 = t−1 are fundamental solutions of L[y] = 0.
(b) Find the general solution of the equation L[y] = t2 + 1.
Solution.

Example 4. Find a particular solution

t3 y 000 − t2 (t + 3)y 00 + 2t(t + 3)y 0 − 2(t + 3)y = −4t4

2020-2019 ú
GA JË@ ɒ®Ë@
Page 18 of 19
Ordinary Differential Equations (Math 2302)
 áÖ ß
@ .X . @
A®‚Ë@
given the fundamental set of solutions t, t2 , tet .


Solution.

2020-2019 ú
GA JË@ ɒ®Ë@
Page 19 of 19