Chapter 3

Higher Order Linear ODEs

3.1 Homogeneous Linear ODEs

Recall from §1.1 that an ODE is of nth order if the nth derivative y ( n ) = d n y dx n of
the unknown function y (x) is the highest occurring derivative. Thus the ODE is of the
form

F ( x, y, y ′, y ( n ) ) = 0 (y (n)
= d n y dx n )
where lower order derivatives and y itself may or may not occur. Such an ODE is
called linear if it is written

y ( n ) + pn−1 ( x) y ( n−1) +  + p1 y ′ + p0 y = r ( x)

(For n = 2 this is (1) in §2.1 with p1 = p and p0 = q .) The coefficients p0 , …, pn−1

and the function r on the right hand side are any given functions of x, and y is
unknown. y (n ) has the coefficient of 1. We call this standard form. An nth-order
ODE that cannot be written in this form is called nonlinear.

If r ≡ 0 , then y ( n ) + pn−1 ( x) y ( n−1) +  + p1 y ′ + p0 y = r ( x) becomes

y ( n ) + pn−1 ( x) y ( n−1) +  + p1 y ′ + p0 y = 0

and is called homogeneous. Otherwise, it is nonhomogeneous.

A solution of an nth-order (whether it is linear or nonlinear) ODE on some open

interval I is a function y = h (x) that is defined and n times differentiable on I and is
such that the ODE becomes an identity if we replace the unknown function y and its
derivatives by h and its corresponding derivatives.

※ THEOREM 1 Fundamental Theorem for the Homogeneous Linear ODE

For a homogeneous linear ODE, linear combinations of solutions on some open

interval I are again solutions on I. (This does not hold for a nonhomogeneous or
nonlinear ODE!)

※ Definition: General Solution, Basis, Particular Solution

1
A general solution of a HODE on an open interval I is a solution on I of the form

y ( x) = c1 y1 ( x) + c2 y 2 ( x) +  + cn y n ( x)

where y1 , y 2 ,…, y n is a basis (or fundamental system) of solutions on I; that is, these

solutions are linearly independent on I, as defined below.

A particular solution on I is obtained if we assign specific values to the n constants

c1 , c2 ,…, cn .

※ Definition: Linear Independence and Dependence

n functions y1 ( x) , y 2 ( x) ,…, y n (x) are called linearly independent on some interval I

where they are defined if the equation

k1 y1 ( x) + k 2 y 2 ( x) +  + k n y n ( x) = 0 on I

implies that all k1 , k 2 ,…, k n are zero. These functions are linearly dependent on I if

this equation also holds on I for some k1 , k 2 ,…, k n not all zero.

※ Initial Value Problems. Existence and Uniqueness

An initial value problem for the HODE consists of an ODE and n initial conditions

y ( x0 ) = K 0 , y ′( x0 ) = K1 ,…, y ( n−1) ( x0 ) = K n−1

with given x0 in the open interval I considered, and given K 0 , K1 ,…, K n−1 .

※ THEOREM 2 Existence and Uniqueness Theorem for Initial Value Problem

If the coefficients p0 ( x) , p1 ( x) ,…, pn−1 ( x) of a HODE are continuous on some open

interval I and x0 is in I, then the initial value problem of HODE has a unique solution
y (x) on I.

※ Linear Independence of Solution. Wronskian

The Wronskian W for n solutions y1 ( x) , y 2 ( x) ,…, y n (x) is defined as the nth order

2
determinant

y1 y2  yn
y1′ y 2′  y n′
W=
   
y1( n−1) y 2( n−1) ( n −1)
 yn

※ THEOREM 3 Linear Dependence and Independence of Solutions

Let the HODE have continuous coefficients p0 ( x) , p1 ( x) ,…, pn−1 ( x) on an open

interval I. Then n solutions y1 ( x) , y 2 ( x) , …, y n (x) on I are linearly dependent on

I if and only if their Wronskain is zero for some x = x0 in I. Furthermore, if W is zero

for x = x0 , then W is identically zero on I. Hence if there is an x1 in I at which W is

not zero, then y1 ( x) , y 2 ( x) , …, y n (x) are linearly independent on I, so that they

form a basis of solutions of HODE on I.

※ THEOREM 4 Existence of a General Solution

If the coefficients p0 ( x) , p1 ( x) ,…, pn−1 ( x) of a HODE are continuous on some open

interval I, then it has a general solution on I.

※ THEOREM 5 General Solution Include All Solutions

If a HODE has continuous coefficients p0 ( x) , p1 ( x) ,…, pn−1 ( x) on some open

interval I, then every solution y = Y (x) is of the form

Y ( x) = C1 y1 ( x) + C 2 y 2 ( x) +  + C n y n ( x)

where y1 ( x) , y 2 ( x) ,…, y n (x) is a basis of solutions of on I and C1 , C 2 ,…, C n are

suitable constants.

3
3.2 Homogeneous Linear ODEs with Constant Coefficients

In this section we consider nth-order homogeneous linear ODEs with constant

coefficients, which we write in the form

y ( n ) + an−1 y ( n−1) +  + a1 y ′ + a0 y = 0

where y ( n ) = d n y dx n . Substituting y = e λx , we obtain the characteristic equation

λn + an−1λ( n−1) +  + a1λ + a0 = 0

We again have three cases:

(I) Distinct Real Roots

If all the n roots λ1 , λ2 ,…, λn are real and distinct, then the n solutions

y1 ( x) = e λ1x , y 2 ( x) = e λ2 x , …, y n ( x) = e λn x

constitute a basis for all x. The corresponding general solution is

y ( x) = c1e λ1x + c2 e λ2 x +  + cn e λn x

※ THEOREM 1 Basis

Solutions y1 ( x) = e λ1x , y 2 ( x) = e λ2 x , …, y n ( x) = e λn x of a HODE (with any real or

complex λ j ' s ) form a basis of solutions on any open interval if and only if all n roots

are different.

※ THEOREM 2 Linear Independence

Any number of solutions of a HODE of the form e λx are linearly independent on an

open interval I if and only if the corresponding λ are all different.

(II) Simple Complex Roots

If complex roots occur, they must occur in conjugate pairs. If λ = a + bi is a simple

4
root of the characteristic equation, so is its conjugate λ = a − bi , and two
corresponding linearly independent solutions are

y1 = e ax cos bx , y 2 = e ax sin bx

(III) Multiple Real Roots

If a real double roots occur, say, λ1 = λ2 , then y1 = y 2 , and we take y1 and xy1 as
corresponding linearly independent solutions, as was done in §2.2. More generally, if
λ is a real root of order m, then m corresponding linearly independent solutions are

e λx , xe λx , x 2 e λx ,…, x m−1e λx

※ Multiple Complex Roots

In this case, real solutions are obtained as for complex simple roots above.
Consequently, if λ = a + bi is a complex double root, so is its conjugate λ = a − bi .
Corresponding linearly independent solutions are

e ax cos bx , e ax sin bx , xe ax cos bx , xe ax sin bx

The corresponding general solution is

y ( x) = e ax [ ( A1 + A2 x) cos bx + ( B1 + B2 x) sin bx ]

Examples:

Solve the following ODEs:

(1) y ′′′ − 6 y ′′ + 11y ′ − 6 y = 0 y ( x) = c1e x + c2 e 2 x + c3 e 3 x

(2) y ′′′ + 3 y ′′ + 3 y ′ + y = 0 y ( x) = e − x (c1 + c2 x + c3 x 2 )

(3) y ( 4 ) − 4 y ′′′ + 8 y ′′ − 16 y ′ + 16 y = 0 y ( x) = (c1 + c2 x) e 2 x + c3 cos 2 x + c4 sin 2 x

(4) y ( 4 ) + 8 y ′′ + 16 y = 0 y ( x) = (c1 + c2 x) cos 2 x + (c3 + c4 x) sin 2 x

※ Homogeneous Euler-Cauchy Equation

By substituting y = x m and its derivatives, we obtain the characteristic equation as a

5
function of m. The rest is exactly the same as §2.5.

(I) Distinct Real Roots

If all the n roots λ1 , λ2 ,…, λn are real and distinct, then the n solutions are

y1 ( x) = x m1 , y2 ( x) = x m2 , …, yn ( x) = x mn

constitute a basis for all x. The corresponding general solution is

y ( x) = c1 x m1 + c2 x m2 +  + cn x mn

(II) Simple Complex Roots

If complex roots occur, they must occur in conjugate pairs. If λ = a + bi is a simple

root of the characteristic equation, so is its conjugate λ = a − bi , and two
corresponding linearly independent solutions are

y1 = x a cos ( b ln x ) , y2 = x a sin ( b ln x )

(III) Multiple Real Roots

If a real double roots occur, say, λ1 = λ2 , then we take y1 and ( ln x ) y1 as

corresponding linearly independent solutions, as was done in §2.5. More generally, if

m is a real root of order n, then n corresponding linearly independent solutions are

x m , x m ( ln x ) , x m ( ln x ) ,…, x m ( ln x )
2 n −1

Examples:

Solve the following ODEs:

(1) x 3 y ′′′ − 3 x 2 y ′′ + 6 xy ′ − 6 y = 0 y h ( x) = c1 x + c2 x 2 + c3 x 3

(2) x 3 y ′′′ + xy ′ − y = 0 y h ( x) = c1 x + c2 x ln x + c3 x (ln x) 2

(3) x 3 y ′′′ − x 2 y ′′ + 2 xy ′ − 2 y = 0 y h ( x) = c1 x 2 + c2 x + c3 x ln x

6
3.3 Nonhomogeneous Linear ODEs

We now discuss nonhomogeneous linear ODE. The standard form is

y ( n ) + pn−1 ( x) y ( n−1) +  + p1 y ′ + p0 y = r ( x)

with y ( n ) = d n y dx n as the first term, and r ( x) ≠ 0 . A general solution on an open

interval I is of the form

y ( x) = y h ( x) + y p ( x)

Here y h ( x) = c1 y1 ( x) + c2 y 2 ( x) +  + cn y n ( x) is a general solution of the corresponding

homogeneous ODE

y ( n ) + pn−1 ( x) y ( n−1) +  + p1 y ′ + p0 y = 0

on I. y p (x) is any solution of the NHODE on I containing no arbitrary constants. If

the NHODE has continuous coefficients and a continuous r (x) on I, then a general

solution exists and includes all solutions (no singular solutions). To obtain y p (x) ,

there are at least two methods as before for 2nd-order ODEs. (Methods of
undetermined coefficients & variation of parameters)

An initial value problem for NHODE consists of the NHODE and n initial
conditions

y ( x0 ) = K 0 , y ′( x0 ) = K1 ,…, y ( n−1) ( x0 ) = K n−1

with x0 in I. Under those continuous conditions it has a unique solution.

※ Method of Undetermined Coefficients

For a constant-coefficient equation

y ( n ) + an−1 y ( n−1) +  + a1 y ′ + a0 y = r ( x)

( a0 , a1 ,…, an constant) and special r (x) as in §2.7, such a y p (x) can be determined by

7
the method of undetermined coefficients, as in §2.7, using the following rules.

(A) Basic Rule as in §2.7.

(B) Modification Rule. If a term in your choice for y p (x) is a solution of the

homogeneous equation, then multiply y p (x) by x k , where k is the smallest positive

integer such that x k y p (x) is not a solution of the HODE.

(C) Sum Rule as in §2.7.

Example:

Find y p (x) for the ODE:

y ( 4 ) − 4 y ′′′ + 3 y ′′ + 4 y ′ − 4 y = e 3 x , e 2 x

Solution:

The homogeneous solution: y h ( x) = c1e x + c2 e − x + (c3 + c4 x) e 2 x

1
Try y p ( x) = C1e 3 x , C1 =
8

1
Try y p ( x) = C2 x 2 e 2 x , C2 = .
6

※ Method of Variation of Parameters

Similar to the 2nd-order ODE case

n
Wk ( x)
y p ( x) = ∑ y k ( x) ∫ dx
k =1 W ( x)
W ( x) W ( x) W ( x)
= y1 ( x) ∫ 1 dx + y 2 ( x) ∫ 2 dx +  + y n ( x) ∫ n dx
W ( x) W ( x) W ( x)

on an open interval I on which the coefficients and r (x) are continuous. The

functions y1 , y 2 , …, y n form a basis of the homogeneous ODE, with Wronskain

8
W, and Wk ( k = 1, 2, …, n) is obtained from W by replacing the k-th column by
 0 
 0 
  . Thus, when n = 2, this becomes identical with the case in §2.10.
  
 
 r ( x) 

Example:

Find y p (x) for the ODE:

ln x
x 3 y ′′′ − 3 x 2 y ′′ + 6 xy ′ − 6 y = −2 x 2 ln x , r ( x) = −2 .
x

Solution:

Homogeneous solution: y h ( x) = c1 x + c2 x 2 + c3 x 3

Wronskian:

x x2 x3 0 x 2 x3
W = 1 2 x 3x 2 = 2 x 3 , W1 = 0 2 x 3 x 2 = −2 x 3 ln x
0 2 6x − 2 lnxx 2 6x

0x x3 xx2 0
W2 = 1 0 3 x 2 = 4 x 2 ln x , W3 = 1 2 x 0 = −2 x ln x
0 − 2 lnxx 6x 0 2 − 2 lnxx

W1
u1′ = = − ln x , u1 = x − x ln x
W

W2 ln x
u 2′ = =2 , u 2 = (ln x) 2
W x

W3 ln x 1 ln x
u3′ = =− 2 , u3 = +
W x x x

Particular solution: y p ( x) = u1 y1 + u 2 y 2 + u3 y3 = x 2 [ 2 + (ln x) 2 ] .

Solution:
9
 y ( x) =c1 x + c2 x 2 + c3 x3 + x 2  2 + ( ln x )  =c1 x + c2 x 2 + c3 x3 + x 2 ( ln x ) .
2 2
 

Example:

Find y p (x) for the ODE:

6
x 3 y ′′′ + xy ′ − y = 6 x , r ( x) = .
x2

Solution:

Homogeneous solution: y h ( x) = c1 x + c2 x ln x + c3 x(ln x) 2

Wronskian:

x x ln x x(ln x) 2 0 x ln x x(ln x) 2
6 (ln x) 2
W = 1 ln x + 1 (ln x) + 2 ln x = 2 , W1 = 0
2
ln x + 1 (ln x) 2 + 2 ln x =
6 x
0 1x 2 x (1 + ln x) 1x 2 x (1 + ln x)
x2

x 0 x(ln x) 2 x x ln x 0
ln x 6
W2 = 1 0 (ln x) + 2 ln x = −12
2
, W3 = 1 ln x + 1 0 = .
6 x 6 x
0 2 x (1 + ln x) 0 1x 2
x2 x

Particular solution: y p ( x) = x (ln x) 3

Solution:

c1 x c2 x ln x + c3 x ( ln x ) + x ( ln x ) .
y ( x) =+
2 3

10