Nama : Geovani N Sihombing (190801010)

Delima br Waruwu (190801060)

Christine Feby Febyola S (190801066)
Nelly L Ompusunggu (190801076)
Desima Roida Glori Simorangkir (190801088)
Matkul : Fisika Matematika II

Find the eigenvalues and corresponding normalized eigenvectors of the matrices in Exercises
6.2.1 through 6.2.14. Orthogonalize any degenerate eigenvectors.
1 0 1
6.2.1
(
A= 0 1 0
1 0 1 )
Ans : λ = 0, 1, 2
λ 1=0 , r 1=( 1 ,0 ,−1 ) √2
λ 2=1 ,r 2 =( 0 ,1 , 0 )
λ 3=2 ,r 3=( 1 ,0 , 1 ) √ 2

1 √2 0
6.2.2
1 (
A= √ 2 0
0
0
0 )
Ans : λ=−1 , 0 ,2
λ 1=−1 ,r 1 =( 1 ,−√ 2 ,0 ) / √3
λ 2=0 , r 2=( 0 , 0 , 1 )
λ 3=2 ,r 3=( √ 2 ,1 , 0 ) / √ 3

1 1 0
6.2.3
(
A= 1 0 1
0 1 1 )
( 1 ,−2,1 )
Ans = R1=
√6
λ 1=¿-1
( 1 , ,0 , 1 )
R 2=
2
λ 2=¿ 1
( 1 ,1,1 )
R 3=
√3
 λ 3=¿ 2
Jadi ; ( -1 ,1 , 2 )
1 √8 0
(
6.2.4. A = √ 8
0
1 √8
√8 1 )
( 1 ,− √ 2 ,1 )
Ans = R1=
2
λ 1=¿-3
( 1 , ,0 ,−1 )
R 2=
√2
λ 2=¿ 1
( 1 , √2 , 1 )
R 3=
2
λ 3=¿ 5
Jadi ; ( -3 , 1 , 5)
Find the eigenvalues and corresponding normalized eigenvectors of the matrices in Exercices
6.2.1 through 6.2.14. Orthogonalize any degenerate eigenvectors.
1 0 0
6.2.5 A= 0 1 1
0 1 1( )
Ans : λ 1 = 0 , r1 = (0,1,-1) / √ 2
λ 2 = 1, r2 = (1,0,0)
λ 3 = 2, r3 = (0,1,1) / √ 2
1 0 0
6.2.6 A= 0
0( 1
√2
√2
0 )
Ans : λ 1= -1, r1 = (0,1, −√ 2)
λ 2 = +1, r2 = (1,0,0)
λ 3 = 2, r3 = (0,√ 2,0) / √ 3
0 1 0
6.2.7
(
A= 1 0 1
0 1 0 )
λ1 = − √ 2, r1 = (1, − √ 2, 1)/2
λ2 = 0, r2 = (1, 0, −1)/ √ 2
λ3 = √ 2, r3 = (1, √ 2, 1)/2.
λ=−√2,0,√2.
 2 0 0
6.2.8
(
A= 0 1 1
0 1 1 )
λ1 = 0, r1 = (0, 1, −1)/ √ 2
λ2 = 2, r2 = (0, 1, 1)/ √ 2
λ3 = 2, r3 = (1, 0, 0)
λ=0,2,2.
0 1 1
6.2.9
(
A= 1 0 1
1 1 0 )
λ1 = 2, r1 = (1, 1, 1)/ √ 3
λ2 = −1, r2 = (1, −1, 0)/ √ 2
λ3 = −1, r3 = (1, 1, −2)/ √ 6.
λ = 2, − 1 , − 1 ,
1 −1 −1
(
6.2.10 A = −1 1 −1
−1 −1 1 )
Ans :
1 −1 −1
(
A = −1 1 −1
−1 −1 1 )
( A−λI ) x=0
( A−λI )=0

1 −1 −1 1 0 0 x1 0
( −1 −1 1 0 0 1 x3
) ( )[ ] [ ]
−1 1 −1 −λ 0 1 0 x 2 = 0
0

1−λ −1 −1

[ −1 1−λ −1 =0
−1 −1 1− λ ]
Det A=
{( (1 )(−1 ) (−1 ) ) + ( (−1 ) (−1 )(−1 ) ) + ( (−1 )(−1 ) (−1 ) ) }−{( (−1 ) (−1 )( 1 ) ) + ( (−1 )(−1 ) ( 1 ) ) + ( (−1 ) (−1 )( 1 ) ) }
{ ( 1 ) + (−1 ) + (−1 ) } −{ ( 1 )+ ( 1 )+ (1 ) }
(−1 )−3=−4

[−11 −11 ]=¿ 0 [−11 −11 ]=¿ 0 [−11 −11 ]=¿ 0

λ 3−( 3 ) λ2 + ( 0 ) λ−| A|=0
λ 3−( 3 ) λ2 + ( 0 ) λ−|−4|=0
 λ 3−( 3 ) λ2 + ( 0 ) λ+ 4=0
λ 3−( 3 ) λ2 +4=0
( λ+1 )( λ−2 ) ( λ−2 )
λ=−1,2,2
λ1 = −1, r1 = (1, 1, 1)/ √ 3
λ2 = 2, r2 = (1, −1, 0)/ √ 2
λ3= 2, r3 = (1, 1, −2)/ √ 6.

1 1 1
( )
6.2.11 A = 1 1 1
1 1 1
Ans :
1 1 1
A= 1 1 1
1 1 1( )
( A−λI ) x=0
( A−λI )=0

1 1 1 1 0 0 x1 0
( ) (
1 1 1 −λ 0 1 0 x 2 = 0
1 1 1 0 0 1 x3 0
)[ ] [ ]
1−λ 1 1

[ 1
1
1−λ
1
1 =0
1− λ ]
Det A=
{( (1 )( 1 ) (1 )) + ( ( 1 )( 1 ) ( 1 ) )+ ( ( 1 )( 1 ) ( 1 ) ) }− {( ( 1 ) (1 )( 1 ) ) + ( ( 1 ) ( 1 )( 1 ) ) + ( ( 1 ) ( 1 )( 1 ) ) }
{ ( 1 ) + ( 1 ) + ( 1 ) }− { (1 )+ ( 1 ) + ( 1 ) }
( 3 ) −3=0

[ 11 11]=¿ 0 [ 11 11]=¿ 0 [ 11 11]=¿ 0

λ 3−( 3 ) λ2 + ( 0 ) λ−| A|=0
λ 3−( 3 ) λ2 + ( 0 ) λ−|0|=0
λ 3−( 3 ) λ2 + ( 0 ) λ+0=0
λ 3−( 3 ) λ2=0
( λ−3 ) ( λ−0 ) ( λ−0 )
λ=3,0,0
 λ1 = 3, r1 = (1, 1, 1)/ √ 3
λ2 = 0, r2 = (1, −1, 0)/ √ 2
λ3 = 0, r3 = (1, 1, −2)/ √ 6.

5 0 2
(
6.2.12 A = 0 1 0
2 0 2 )
Ans :
5 0 2
A= 0 1 0
2 0 2( )
( A−λI ) x=0
( A−λI )=0

5 0 2 1 0 0 x1 0
( 2 0 2
) ( )[ ] [ ]
0 1 0 −λ 0 1 0 x 2 = 0
0 0 1 x3 0

5−λ 0 2

[ 0
2
1− λ
0
0 =0
2−λ ]
Det A=
{( (5 )( 1 ) ( 2 ) ) + ( 0 ) + ( 0 ) }−{ ( ( 2 )( 1 ) ( 2 ) ) + ( 0 ) +( 0 ) }
( 10 )−4=6
1 0 5 2 5 0
[ ] 0 2
=¿ 2 [ ]
2 2
=¿6 [ ]
0 1
=¿ 5

2+6+5 = 13
λ 3−( 8 ) λ2 + ( 13 ) λ−|A|=0
λ 3−( 8 ) λ2 + ( 13 ) λ−|6|=0
λ 3−( 8 ) λ2 + ( 13 ) λ−6=0
λ 3−( 8 ) λ2 + ( 13 ) λ−6=0
( λ−1 )( λ−1 ) ( λ−6 )
λ=1,1,6
λ1 = 6, r1 = (2, 0, 1)/ √ 5
λ2 = 1, r2 = (1, 0, −2)/ √ 5
λ3 = 1,
 11 0
( )
6.2.13 A= 11 0
000
Ans :
λ1 = 2, r1 = (1, 1, 0)/√2
λ2 = 0, r2 = (1, −1, 0)/√2
λ3 = 0, r3 = (0, 0, 1).

5 0 √3
6.2.14 A=
( )
030
√3 0 3
Ans :
λ1 = 2, r1 = (1, 0, −√3)/2
λ2 = 3, r2 = (0, 1, 0)
λ3 = 6, r3 = (√3, 0, 1)/2
6.2.15 Describe the geometric properties of the surface
x 2+ 2 xy +2 y 2+ 2 yz + z 2=1
How is it oriented 3-D space? Is it a comic section? If so, which kind?
Ans :
Since the quadratic form

x 2+ 2 xy +2 y 2+ 2 yz + z 2=1
defining thesurface is obviously positive definite upon writing it as a sum of squares,(x
+ y)2 + (y + z)2 = 1, it is an ellipsoid or an ellipse. Finding the orientation in space
amounts to diagonalizing the symmetric 3 × 3 matrix of coefficients. The
characteristic polynomial is λ(1 − λ)(λ − 3) = 0, so that the eigenvalues are λ = 0
implying an ellipse, and λ = 1, and 3. For λ = 1 an eigenvector is v1 = (1, 0, −1) giving
one of its axes, for λ = 3 an eigenvector is v3 = (1, 2, 1) giving the other axis. v1 × v3 =
(2, −2, 2) is normal to the plane of the ellipse.
6.4.3 Show that a real matrix that is not symmetric cannot be diagonalized by an orthogonal
or unitary transformation.
Hint. Assume that the nonsymmetric real matrix can be diagonalized and develop a
contradiction.
Ans :
Assume that a unitary matrix U causes the real nonsymmetric matrix A to be diagonal,
i.e., that UA U T =D, a diagonal matrix. If we apply the inverse transformation to D, to
 recover A, we would have A=U T DU . But this form for A is symmetric :
(U T DU )T =U T DU .

6.4.4 The matrices representing the angular momentum components L x , L y ,∧Lz are all

Hermitian. Show that the eigenvalues of L2 , when L2=L2x + L2y + L2z , are real and
nonnegative.
Ans :
First, note that L2x has the same eigenvectors as L x, with eigenvalues that are the

squares of the (real) L x eigenvalues. Therefore, L2x (and for the same reason, L2y and

L2z ) have only nonnegative eigenvalues. Second, for vectors | x ⟩ of unit length, the

expectation value ⟨ x|L2x|x ⟩ will be real and have as its smallest possible value the

smallest eigenvalue of L2x. Proof of this statement is the topic of Exercise 6.5.5. Similar

statements are true for L2y and L2z, so ⟨ x|L2x + L2y + L2z| x ⟩ must always be nonnegative. We

theretore may conclude that all the eigenvalues of L2x + L2y + L2z are nonnegative.

6.4.5 A has eigenvalues λ i and corresponding eigenvectors ∣ x i ⟩ . Show that A−1 has the same

eigenvectors but with eigenvalues λ−1

i .

Ans :
If A ∣ x i ⟩ =λi ∣ xi ⟩ , then ∣ x i ⟩ =λi A ∣ x i ⟩ upon multiplying with the inverse matrix.
−1

Moving the (non-zero) eigenvalue to the left-hand side proves the claim.

6.4.6 A square matrix with zero determinant is labeled singular.

(a) If A is singular, show that there is at least one nonzero column vector v such that
A ∣ v ⟩=0.
(b) If there is a nonzero vector ∣ v ⟩ such that
A ∣ v ⟩=0 ,
show that A is a singular matrix. This means that if a matrix (or operator) has zero as
an eigenvalue, the matrix (or operator) has no inverse and its determinant is zero.
Ans :
(a) If A is singular, its determinant is zero. If A is transformed to diagonal form, its
determinant is seen to be the product of its eigenvalues, so a zero determinant
 indicates that at least one eigenvalue is zero. The eigenvector corresponding to a zero
eigenvalue will have the property that A ∣ v ⟩=0.
(b) If A ∣ v ⟩=0, then ∣ v ⟩ is an eigenvector with eigenvalue zero, the determinant of A
will be zero, and A will be singular.

6.4.7 Two Hermitian matrices A and B have the same eigenvalues. Show that A and B are
related by a unitary transformation.
Ans :
ϯ ϯ
If U 1 AU 1= [ λ1 ,… , λ n ]=U 2 BU 2 with unitary matrices U i , then

A=U ϯ1 U 2 BU ϯ2 U 1=U ϯ1 U 2 B (U ϯ1 U 2) ϯ
6.5.1 Find the eigenvalues and corresponding eigenvectors for

(21 42 )
Note that the eigenvectors are not orthogonal.
Ans : λ 1 = 0, c1 = (2, -1)
λ 2 = 4, c2 = (2, 1)
6.5.2 If A is a 2 ×2 matrix, show that its eigenvalues λ satisfy the secular equation
λ 2−λtrace(A) + det(A) = 0
Ans : The characteristic polynomial is
( λ−λ1) ( λ−λ2 ) = λ 2 – ( λ 1+ λ 2) λ+ λ1 λ2 = λ 2 – trace(A)λ+ det(A) = 0

6.5.3 Assuming a unitary matrix U to satisfy an eigenvalue equation UC = λ C, show that the
eigenvalues of the unitary matrix have unit magnitude. This same result holds for real
orthogonal matrices.
Ans : If Ur = λr with |r 2|= 1, then 1 = r ˖r = r ˖ U ˖ Ur = |λ|2 r ˖r = |λ|2

6.5.4 Since an orthogonal matrix describing a rotation in real 3-D space is a special case of a
unitary matrix, such an orthogonal matrix can be diagonalized by a unitary
transformation.
a. Show that the sum of the three eigenvalues is 1+2 cosφ, where φ is the net angle of
rotation about a single fixed axis.
b. Given that one eigenvalue is 1, show that the other two eigenvalues must be e iϕ
and e−iφ .
Our orthogonal rotation matrix (real elements) has complex eigenvalues.
 Ans :
Choose a coordinate system in which the rotation is about the z-axis, and transform
our rotation matrix to these coordinates. This transformation will not change the trace
of the rotation matrix. Now the rotation matrix will have the form
cosφ sin φ 0
(
U = −sin φ sin φ 0
0 0 1 )
and the trace of U is obviously 1+2 cosφ
6.5.5 A is an nth-order Hermitian matrix with orthonormal eigenvectors |xi ⟩ and real

eigenvalues λ 1 ≤ λ2 ≤ λ 3 ≤ … ≤ λn. Show that for a unit magnitude vector | y i ⟩

λ 1 ≤ ⟨ y| A| y ⟩ ≤ λ n
Ans :

Expand | y i ⟩ in the eigenvectors: | y i ⟩ =∑ c i|x i ⟩

i

Then note that, because |yi is of unit magnitude and the | x i ⟩ are orthonormal

⟨ y| y ⟩−∑ c ¿i c j ⟨ x i|x j ⟩−∑ |c 2i|=¿ 1 ¿

ij i

Moreover, because the | x i ⟩ are orthonormal eigenvectors,

⟨ y| A| y ⟩ =∑ c ¿i c j ⟨ xi| A| x j ⟩ =∑|c2i | λ i
ij i

Lower and upper bounds for this expression can now be obtained by replacing λi by
the smallest or the largest eigenvalue, after which the |c 2 i | can be summed (yielding
unity).
6.5.6 A particular matrix is both Hermitian and unitary. Show that its eigenvalues are all ± 1.
Note. The Pauli and Dirac matrices are specific examples.
Ans :
if it is unitary, then
A+¿ A=I ¿
A=A +¿¿
AA= A 2=I
Since λ of I is 1, A2 x=λ x =x
6.5.7 For this relativistic electron theory Dirac required a set of four anticommuting
matrices. Assume that these matrices are to be Hermitian and unitary. If these are n × n
matrics,show that n must be even. With 2 ×2matrices inadequte (why?), this
demonstrates that the smallest possible matrices forming a set of four anticommuting,
Hermitian, unitary matrices 4 × 4.
 Answer: If γ μ and γ v anticommute, γ μ γ v =−γ v γ μ. Take thr determinants of the two
sides of this equation:
γ μ γ v =det ( γ μ ) det ( γ v ) =det (−γ v γ μ )=(−1 )n det ⁡( γ v ) det(γ ¿ ¿ μ)¿
Here the have used the fact that the determinant is not liniear operator, and that the
determinant of –A is (−1 )n det ( A ) , when n is the dimension og the determinant. Since
theγ are the unitary, they cannot be singular, and the anticommutation leads to
inconsistency unless n is even, making (−1 )n=+1

6.5.8 A is a normal matrix with eigenvalues λ n and orthonormal eigenvectors | X n ⟩ . Show

that A may be written as

A=∑ λn| X n ⟩ ⟨ X n|
n

Hint. Show that both this eigenvector form of A and the orginal A give the same result
acting on an arbitary vector | y ⟩

Answer: Expand| y ⟩ in the eigenfunctions: A ¿, with c n= ⟨ X n| y ⟩ . We get the same

result from the eigenvector form of A:

A| y ⟩ =∑ λn| X n ⟩ ⟨ X n| y ⟩ =∑ λ n c n| X n ⟩ ,
n n

6.5.9 A has eigenvalues 1 and -1 and corresponding eigenvectors (10 )and (01 ) . Construct A.
1 0
Answer : A=
0 −1 ( )
6.5.10 A non-Hermitian matrix A has eigenvalues λi and corresponding eigenvectors |ui>.
The adjoint matrix A† has the same set of eigenvalues but different corresponding
eigenvectors, |vi >. Show that the eigenvectors form a biorthogonal set in the sense
that
<vi |uj> = 0 for λi *≠ λj.
Ans :
Write
⟨ vi| A|uj ⟩ =λj ⟨ vi∨uj ⟩
¿ ⟨ A † vi∨uj ⟩= λi*⟨ vi∨uj ⟩ .
Subtracting the right-hand side of the second line from that of the first line,
(λj− λi*)*⟨ vi∨uj ⟩ =0,
 from which we conclude that⟨ vi∨uj ⟩ =0unless λi *¿ λj.

6.5.11 You are given a pair of equations:

A|fn> = λn|gn>
A˜|gn> = λn|fn> with A real.
(a) Prove that |fn> is an eigenvector of (A~A) with eigenvalue λn².
(b) Prove that |gn> is an eigenvector of (AA˜) with eigenvalue λn².
(c) State how you know that
(1) The |fn> form an orthogonal set.
(2) The |gn> form an orthogonal set.
(3) λn² is real.
Ans :
(a) and
~
(b) Apply A to the first equation and A to the second:
~ ~
A A∨f n ⟩ =λn A∨g n ⟩ =λn ²∨f n ⟩ ,
~
A A∨g n ⟩=λnA∨f n ⟩ =λn ²∨g n ⟩ .
~ ~
(c) Because A is real, A A and A A are both self-adjoint (Hermitian), and therefore
have eigenvectors that have real eigenvalues and form an orthogonal set.
6. 5. 12 Prove that A of the preceding exercise may be written as
∑ λn∨gn >¿ f n|,
n

With the | gn> and < f n| normalized to unity.

Hint. Expand an arbirary vector as a linear combination of | f n>.
Ans :
If the given formula for A gives the required result for every member of
an orthogonal set it is a valid expression for A. Apply the formula to an
| f j > of arbitrary j. Because the f n> are orthonormal, the result reduces to
λ j | g j>.

6. 5. 13 Given
1 2 2
A= (
√5 1 −4 )
(a) Construct the transpose Ᾱ and the symmetric forms ᾹA ∧ AᾹ
Ans :
 1 8 −6
AᾹ = (
5 −6 17 ) ᾹA = (10 04 )
(b) From AᾹ| gn> = λ 2n | gn>, find λ nand | gn>. Normalize the ¿ gn>.
Ans :
1 2
λ1 = 1 | g1 > = ()
√5 1
, |f 1 > = (10)
(c) From AᾹ| f n> = λ 2n | gn>, find λ n(same as (b)) and | f n>. Normalize the ¿ f n>.
Ans :
1 1
λ2 = 1 | g2 > = ( )
√ 5 −2
, |f 2 > = (01 )
(d) Verify that A| f n> = λ n∨gn> and Ᾱ| gn> = λ n∨f n>.

(e) Verify that A = ∑ λn∨gn >¿ f n|,

n

6. 5. 14 Given the eigenvalues λ 1= 1, λ 2= -1 and the corresponding eigenvectors

| f 1> = (10 ) ,∨g > = √12 (11 ), ¿ f = (01 ) ,∧¿

1 2

1 1
√2 (−1 )
|g > =
2

(a) Construct A;
Ans :
1 1 −1
A= (
√2 1 1 )
(b) Verify that A| f n> = λ n∨gn> ;
Ans :
Disregard this exercise (it is ill-defined).

(c) Verify that Ᾱ| gn> = λ n∨f n>.

Ans :
Disregard this exercise (it is ill-defined).
6.5.15 Two matrices U and H are related by
U =eiaH
With a real
a. If H is Hermitian, show that U is unitary.
b. If U is unitary, show that H is Hermitian. (H is independent of a.)
c. If trace H = 0, show that det U = +1.
d. If det U = +1, show that trace H = 0.
Hint. H may be diagonalized by a similarity transformation. Then U is also diagonal.
The corresponding eigenvalues are given by uj = exp(iahj).
Ans :
a) Take the adjoint of U; because H is self-adjoint, the result is U † =exp(−iaH). Note
that an exponential can be interpreted as its powerseries expansion and the adjoint
taken termwise, thus validating the processing applied to the exponent. The result
shows that U† = U−1.
b) Form UUT = exp(iaH) exp(−iaH). Because H commutes with itself,this product
reduces to exp(iaH − iaH) = 1. Note that exponentials canbe combined in this way
only if the exponents commute.
c) If H is diagonalized by a similarity transformation, the zero trace implies that the
sum of its eigenvalues λn is zero. Then U, which is also diagonal, will have diagonal
elements exp(iaλn), and its determinant, which will then be the product of its
diagonal elements, will be exp ⁡¿.
d) Conversely, in a basis in which H and U are diagonal, a unit determinant for U
implies an exponential in which exp ⁡¿; this condition does not quite imply that trace
H = 0, but only that a(trace H) = 0 is an integer multiple of 2π.

6.5.16 An n × n matrix A has n eigenvalues Ai. If B = eA , show that B has the same
eigenvectors as A with the corresponding eigenvalues Bi given by Bi = exp(Ai).
Ans :
From A vi= Ai v i we obtain An =A ni v i for i = 0, 1, 2, . . .
∞
n
From B = exp(A) ¿ ∑ A /n ! we get
n=0

∞ ∞
B=∑ A v i /n!=∑ [ A n /n ! ] v i=( e A ) v i
n i

n=0 n=0
6.5.17 For any operator A, the egienvalues of A2 are the squares of the eigenvalues of A
6.5.18 Inserting the indicated expansion and using the orthogunality property of the
eigenvectors,
n
(x|A|x) = λ 1(x1|x1) + ∑ ¿ δ 1|2λn (x1|x1),
i=2

n
(x|x) = (x|x) + ∑ ¿ δ 1|2(x1|x1),
i=2

Because all λi for i > 1 are smaller than λi ,

n
(x|A|x) < λi ( (x1|x1) + ∑ ¿ δ 1|2(x1|x1) ),
i=2

So
(x∨ A∨x)
< λi
( x∨x)
The error when this ration in used to approximate λi is approximately
n
1
¿¿ 1 – λi) |δ i|2(x1|x1),|
(x∨x ) ∑
i=2

Which is of order |δ i|2

6.5.19 Two equal masses are conneted to each other and to wals by springs as shown in Fig.
6.5. The masses are constrained to stay on a horizontal line.
a) Set up the Newtonian acceleration equation for each mass.
b) Solve the secular equation for the eigenvectors
c) Determine the eigenvectors and thus the normal modes of motion.
Ans :
a) Letting x 1 and x 2 be the displacements of the two movable masses, each of the
same mass m, measured from their equilibrium positions (with the positive
direction for both x i the same), the equations of motion are
mẍ1=−kx 1+ k ¿)
mẍ2 =−kx 2 + k ( x 2−x 1 )
In a normal mode of oscillation x i = X e e iωt, with the same angular frequency ω for
both masses. Inserting these expressions,
−mω 2
X1 = -2X1 + X2
k
−mω 2
X2 = X1 −¿2X2
k
These equations are equivalent to the matrix equation
 (−12 −12 )( xx ) = 𝜆( xx )
1

2
1

With eigenvalues 𝜆 = 1 and 𝜆 = 3

b) This is an eigenvalue equation which has soluction only if

|2−λ
−1
−1
2−λ
=0 |
With 𝜆 = mω2/k
c) For 𝜆 = 1, the equation solition is X1 = X2 , corresponding to the two masses
moving, in phase, back and forth. For 𝜆 = 3, the equation soluction is X1 = - X2
,corresponding to a periodic motion in which the masses oscillate relative to each
other.
6.5.20 Given a normal matrix A with eigenvalues λ j, show that A+ has eigenvalues λ *j, its real
part (A + A+) / 2 has eigenvalues Ҡe (λ j), and its imaginary part (A - A+) /2i has
eigenvalues ℑm(λ j).

Ans :
Relying on the proof that a normal matrix A and its adjoint have the save eigenvectors
xj ,
⟨ x j| A|x j ⟩ = 𝜆j⟨ x j|x j ⟩
+¿ x j ¿
=⟨A x j ⟩ = μ*j⟨ x j|x j ⟩ ,
Where μ j is the eigenvalue of A+ corresponding to x j . We see that μ*j = 𝜆j. Since A
and A+ have common eigenvectors,
(A + A+)| x j ⟩ = ( λ j + λ j)| x j ⟩ = 2ℜe𝜆j.
¿

Likewise, A - A+ has eigenvalues 𝜆j – 𝜆*j,or 2i I m𝜆j.

6.5.21 Consider a rotation given by Eular angles α =π/4, β=π /¿2, γ =5 π /¿4
a) Using the formula of Eq. (3.37), contructyhe matrix U representing this rotation.
b) Find the eigenvalues and eigenvectors of U, and from them describe this rotation
by specifying a single rotation axis and an angle of rotation about that axis.
Ans:
a) Using Eq. (3.37), the matrix U of the rotation is
 1/2 −1 /2 1/ √ 2

(
U = 1/2 −1 /2 −1/ √ 2
1/ √ 2 1 / √ 2 0 )
U has the following eigenvalues and eigenvectors:
𝜆1 = 1 r1 = √ 2/3 e^ x+ √ 1/3 e^ z
1
𝜆2 = (-1 +i √3) r2 = -√ 1/6 e^ x +i √1/2 e^ y+ √ 1/3 e^ z
2
1
𝜆3 = (-1 - i √ 3) r3 = -√ 1/6 e^ x - i √1/2 e^ y+ √ 1/3 e^ z
2
From these data we see that the rotation of the coordinate axes corresponding
to U is equivalent to a single rotation about φ 1 by an angle given as the phase
of 𝜆2 (the angle it makes with the real axis), which is 1200.
7.2.1 From Kirchoff’s law the current Ι in an RC (resistance-capasitance) circuit (Fig 7.1)
obeys the equation.
dl 1
R + I =0
dt C
a) Find I(t)
b) For a capacitance of 10,000 μF charged to 100 V and discharging though a
resistance of 1 MΩ, find the current I for t = 0 and for t = 100 seconds.
0
Q
Note. The initial voltage is IοR or where Q=∫ I ( t ) dt
C ∞

Ans :
dl 1
a) R + I =0
dt C
t
α ( t )=exp∫ RC dt=0,
❑
−t
I ( t )=I o e .
RC

Dimana Io adalah integrasi konstanta.

b) Disini Ω = 106 Ohm, kemudian Io = 10-4 Amp, RC = 104 sekon dan t = 100,
−t
-4
e RC =e−0,01 ≈ 0,99, lalu I = 0,99 × 10 Amp. Waktu 100 sekon hanya 1 % dari
konstanta waktu RC

7.2.2 The laplce transform of Bessel’s equation (n = 0) leads to

( s2 +1 ) f ' ( s ) +sf ( s )=0
Solve for f(s)
 Ans :
f ' (s ) sds −1 ( 2 )
ln f ( s )=∫ ds=−∫ 2 = ln s + 1 + ln C,
f (s) s +1 2
C
f ( s )=
√ s2 +1
7.2.3 The decay of a population by catastropic two-body collisions is described by
dN
=−k N 2
dt
This is a first-order, nonlinear differential equation. Derive the solution
−1
1
( )
N ( t ) =N ο 1+
τ0
Where τ0 = (kN0)-1. This impliesan infinite population at t = −τ0
Ans :
N t

∫ dN
N 2
=−∫ kdt=−kt=
−1 1
+
N N
No 0 o

No
N=
t , τ0 =
1
τ
7.2.4 The rate of a particular chemical reaction A + B → C is proportional to the
concentrations of the reactants A and B:
dC ( t )
=α [ A ( 0 )−C ( t ) ][ B ( 0 )−C ( t ) ]
dt
(a) Find C(t) for A(0) 6 ≠ B(0).
(b) Find C(t) for A(0) = B(0).
The initial condition is that C(0) = 0.
Ans :
(a) Set A0 = A(0), B0 = B(0). Separating variables and using a partial fraction
expansion obtain
dC 1 1 1
α ∫ dt=α t=∫ = ∫ − (
( A 0−C ) ( B 0−C ) B0 −A 0 A 0−C B0−C
dC )
A 0−C A0
Thus ln = ( A0 −B 0)αt + ln .
B0−C B0

A 0 B0 [ e( A − B )α t−1 ]
0 0

Rewrite this as C(t) = . Then C(0) = 0.

A 0 e( A −B )α t −1−B 0
0 0

dC
(b) From ∫ 2
=α t get 1 = αt + 1 ,
( A0 −C ) A 0−C A0
 α A 02 t
which yields C(t) = . Again C(0) = 0.
1+ α A0 t
7.2.5 A boat, coasting through the water, experiences a resisting force proportional to v n , v
being the boat’s instantaneous velocity. Newton’s second law leads to
dv
m =−k v n
dt
With v(t = 0) = v 0, x(t = 0) = 0, integrate to find v as a function of time and v as a
function of distance.
Ans :
The values n < 0 are unphysical as the acceleration diverges.
The case n = 0 gives
m[v − v(0)] = −kt, v(t) = v(0) − kt/m, x(t) = x(0) + v(0)t − kt 2/2m .
The case n = 1 gives
mv( 0)
v(t) = v(0)e−kt /m, x(t) = x(0) + ( 1−e−kt /m ).
k
For n ≠ 0, 1, 2 and n > 0 we integrate to get
v 1−n−v (0)1−n dv −k
=∫ n = t
1−n v m
Integrating again gives
mv( 0)2−n kt
x(t) = x(0) +
(2−n)k [ (
1− 1+ ( n−1 ) v (0)n−1
m )]
The case n = 2 leads to
v(0) m kv ( 0 ) t
ẋ=
1+α t
, x(t) = x(0) +
k (
ln 1+
m ).

7.2.6 In the first-order differential equation dy/dx = f (x,y), the function f (x,y) is a function
of the ratio y/x:
dy
= g(y/x).
dx
Show that the substitution of u = y/x leads to a separable equation in u and x.
Ans :
The substitution u = y/x, or y = xu, corresponds to dy = xdu + udx, and our ODE
assumes the form
x du + u dx = g(u) dx , or x du = [g(u) − u] dx ,
which is separable.
7.2.7 The differential equation
P ( x , y ) dx+Q ( x , y ) dy=0
 Is exact. Show that its solution is of the form

x y
φ ( x , y ) =∫ P ( x , y ) dx+ ∫ Q ( x 0 , y ) dy =constant.
x0 y0

Ans :
x
∂φ
If =P ( x , y ) then φ ( x , y ) =∫ P ( X , y ) dX + α ( y ) follows.
∂x x0

∂φ
Differentiating this and using =Q ( x , y ) we obtain
∂y
x
dα ∂ P( X , y )
Q ( x , y )= +∫ dX ,
dy x0 ∂y
So
x
dα ∂Q ( X , y )
=Q ( x , y ) −∫ dX .
dy x0 ∂y
y
dα
So =Q ( x , y ) =Q ( x 0 , y )∧α ( y )=∫ Q ( x 0 ,Y ) dY . Thus
dy y0

x y
φ ( x , y ) =∫ P ( X , y ) dX + ∫ Q ( x 0 ,Y ) dY
x0 y0

∂φ
From this we get =P ( x , y ) and
∂x
x y x
∂φ ∂Q ( X , y )
=∫ P ( X , y ) dX + ∫ Q ( x 0 ,Y ) dY =¿∫ dX +Q ( x , y ) =Q ( x , y ) . ¿
∂ y x0 y0 x0 ∂y
7.2.8 The differential equation
P ( x , y ) dx+Q ( x , y ) dy=0
Is exact. If
x y
φ ( x , y ) =∫ P ( x , y ) dx+ ∫ Q ( x 0 , y ) dy =constant.
x0 y0

Show that
∂φ ∂φ
=P ( x , y ) , =Q ( x , y ) .
∂x ∂y
Hence, φ ( x , y ) =¿ constant is a solution of the original differential equation.
Ans :
See proof of Exercise 7.2.7
7.2.9 Prove that Eq. (7.12) is exact in the sense of Eq. (7.9), provided that α (x) satisfies Eq.
(7.14).
Ans :
 For α dy+ α ( py−q ) dx =0 to be exact requires
∂α ∂
= α ( x ) ( py −q ) =αp .
∂x ∂y
Which is Eq. (7.14)
7.2.10 A certain differential equation has the form
f (x)dx + g (x) h (y) dy = 0
with none of the functions f (x), g (x), h (y) identically zero. Show that a necessary
and sufficient condition for this equation to be exact is that g (x) =constant
Answer : For f (x)dx + g (x) h (y) dy = 0 to be exat requires
∂ f (x ) ∂ g ( x ) h( y ) ∂ g (x)
=0= = h (y) , i.e., g = const.
∂y ∂x ∂y

7.2.11 Show that

x x

y (x) = exp ¿ {∫ exp

1
[∫ p ( t ) dt ¿¿q ( s ) ds +c
1
}
is a solution of
dy
+ p (x) y (x) = q (x)
dx
by differentiating the expression for y (x) and substituting into the differential equation
x

Answer : y I= − pe−Fx pdt f

[∫ 1
}
p ( t ) dt ¿¿q ( s ) ds +c + e−fx pdtI : e fx dt I q(x)

Implies : y I+ p(x) y (x) = q (x)

7.2.12 The motion of a body falling in a resisting medium may be described by
dv
m =mg−bv
dt
when the retarding force is proportional to the velocity, v. Find the velocity.
Evaluatethe constant of integration by demanding that v ( 0 )=0
Ans : Separating variables we get
−bt b b
=ln (g− v)−ln ( A )
m m m
with A an integration constant. Exponentiating this we obtain
−bt
mg m mg
v ( t )= −A e ,thus v 0=v ( 0 )= −A.
b b
−bt
mg mg
(
Hence v ( t )= v 0−
b
e ) m
+
b
. Set v 0=0 here .
 The velocity dependent resistance force opposes the gravitational acceler-ation
implying the relative minus sign.

7.2.13 Radioactive nuclei decay according to the law

dN
=−λN ,
dt
N being the concentration of a given nuclide and λ , the particular decay constant. Ina
radioactive series of two differentnuclides, with concentrations N 1(t)∧N 2(t ),
wehave
dN 1
=−λ 1 N 1 ,
dt
dN 2
=λ 1 N 1− λ 2 N 2.
dt
FindN 2 ( t )for the conditions N 1 ( 0 )=N 0∧N 2 ( 0 )=0
Ans : Solve first for N 1 ,which is separable and has the general solution
N 1=C e−λ1 t . Since N 1 ( 0 )=N 0 , we have N 1(t)=N 0 e− λ1 t
Substitute this result into the equation for N 2, which is now an inhomoge-neous
equation in which N 2 is the only unknown. We look for a particularintegral of the
inhomogeneous equation, guessing the form of the solutionto be N 2= A exp(−λ 1t ).
Thus,
dN 2
+ λ 2 N 2=λ 1 N 0 e−λ 1t becomes− Aλ1 e− λ1 t + Aλ 2e− λ1 t =λ 1 N 0 e− λ1 t ,
dt
confirming that with a proper choice of A our guess will work. We findthat
A=λ 1 N 0 /(λ 2−λ 1).
To this particular integral we must add the multiple of the solution to thehomogeneous
equation that is needed to satisfy the condition N 2(0)=0.The homogeneous equation
has solutione− λ2 t ,so our complete solution is
λ 1 N 0 − λ1 t −λ 2t
N 2 ( t )= ( e −e )
λ 2− λ 1
7.2.14 The rate of evaporation from a particular spherical drop of liquid (constant density) is
proportional to its surface area. Assuming this to be the sole mechanism of mass
loss,find the radius of the drop as a function of time.
Ans : We have dV /dt=−C 4 π r 2 with V =4 π r 3 /3the volume and C a
positiveconstant. So dr /dt=−C∧r (t)=r 0−Ct

.2.15 In the linear homogeneous differential equation

 dv
=−av
dt
the variables are separable. When the variables are separated, the equation is exact.
Solve this differential equation subject to v(0)=v 0 by the following three methods:
(a) Separating variables and integrating.
(b) Treating the separated variable equation as exact.
(c) Using the result for a linear homogeneous differential equation.
Ans :
(a) Separating variables, dv /v =−a dt yields
v
ln =−at , v=v 0 e−at .
v0
(b) dv / v+ a dt =0 yields
ϕ (t , v )=ln v+ at.
ϕ (t , v )=ln v 0=const . is equivalent to (a).
(c) Substituting into the form of solution written in Exercise 7.2.11 with q=0 , p=a we
get v(t)=Ce−at . Setting t=0 we identify C as c 0.

.2.16 (a) Solve Example 7.2.1, assuming that the parachute opens when the parachutist’s
velocity has reached vi =60 mi/h (regard this time as t=0). Find v(t).
(b) For a skydiver in free fall use the friction coefficient b=0.25 kg /m and mass
m=70 kg. What is the limiting velocity in this case?
Ans :
Separating variables as in Example 7.2.1 we get the velocity
t vi
v ( t )=v 0 tanh
[ T
+t an h−1
( )]
v0

For vi ≥ 0.
7.2.17 Solve the ODE
( x y 2 – y )dx + x dy = 0
7.2.18 Solve the ODE
( x 2- y 2 e y/ x ¿ dx +( x¿ ¿2+ xy ) e y/ x dy=0 ¿
Answer
7.2.17 This ODE is isobaric, and becomes separable under the substitution v= xy. Removing
x via this substitution,
the ODE becomes
 7.2.18 This ODE is homogeneous, so we substitute y = vx, obtaining initially

7.3.1 y ' ' ' −2 y ' ' − y ' + 2 y=0

Ans :
Try solution e mx. The condition on m3−2 m2−m+2=0 with roots m = 2, m = 1, m = -1
The general solution to the ODE is therefore c 1 e2 x + c2 e x +c 3 e−x .
7.3.2 y ' ' ' −2 y ' ' + y ' −2 y=0
Ans :
Try solution e mx. The condition on m is m 3−2 m 2+ m−2=0 , with roots m = 2, m = i, m
= -i. The solution e ix and e−ix can be expresed in terms of the real quantities sin x and

cos x, so the general solution to the ODE is c 1 e2 x + c2 sin x+ c3 cos x.

7.3.3 y ' ' ' −3 y ' +2 y=0

Ans :
Try solution e mx .the condition on m is m 3−3 m+2=0, with roots m= 1, m = 1, m = -2.
Two independent solution for m = 1 are e x and xe x , so the general solution to the ODE

is c 1 e x + c 2 xe x +c 3 e−2 x.
7.3.4 Ans : The condition on m is m2 +2 m+2=0,with roots m=−1+i and m=−1−i. We
can cpmbine e ( −1 +i) x and e (x−1−i) to form e− x sin x and e− x cos x ,so the general solution to

the ODE is c 1 e−x sin x +c 2 e−x cos x

−2 x l(l+1)
7.4.1 Ans : For P ( x ) = 2
,Q ( x )=
1−x 1−x 2
 (1∓ x ¿ P and ( 1 ∓ x ) Q are regular at x=± 1,respectively. So these are regular
singularities.
2
z z Q(z−1) l(l+ 1)
As z → 0 , 2 z − =2(z+ 2 is regular, and
) = 2 2 z−2 diverges.
1 1−z z 4
z ( z −1)
1−
22
So ∞ is a regular singularity.
7.4.2   Show that Laguerre’s equation, like the Bessel equation, has a regular singularity at x
= 0 and an irregular singularity at x = ∞.
Ans.
1−x n
For P= ,Q= , x=0 is a regular singularity.
x x
2 z−P(z−1 ) z +1
For z → 0 , 2
= 2 1/ z 2 diverges more rapidly than 1/ z, so ∞ is an
z z
irregular singularity.
7.4.3   Show that Chebyshev’s equation, like the Legendre equation, has regular singularities
at x = −1, 1, and ∞.
Ans. Writing the Chebyshev equation in the form
x n2
y' '+
( 1−x 2) (
y '
+
1−x2 )
y=0 ,

We see that the coefficients of y’ and y become singular (for finite x) only at x = ±1
and that each singularity is first order, so the ODE has regular singularities at these
points.
1
2 x−P ( x−1 ) 2 x
= − 2 ,
X 2
x x −1

Q(x−1) n2
= .
x4 x 4−x 2
These have, at x=0, singularities that are respectively of first and second order,
indicating that ODE has a regular singularity at infinity.
7.4.4   Show that Hermite’s equation has no singularity other than an irregular singularity at x
= ∞.
Ans. Hermite’s ODE has no coefficient thath are singular at finite x, and therefore is
regular for all finite x. At infinity,
2 x−P ( x−1 ) 2 2
= + 3
X2 x x
 Has a singularity of order 3 at x = 0, so the ODE will have an irregular singularity at
infinity.
7.5.6 Obtain two series solutions of the confluent hypergeometric equation
x y” + (c − x)y’ − ay = 0.
Test your solutions for convergence.
Ans:
Compare with Eq. (18.136). Convergent for all finite x provided the series exists [ c 6≠ n ]
, a negative integer, in Eq. (18.137), 2−c 6 ≠ n in Eq. (18.136)].

7.5.7 A quantum mechanical analysis of the Stark effect (parabolic coordinates) leads to the
differential equation
d du 1 m2 1
ξ( ) (
dξ dξ
++ E ξ+ α
2
− F ξ 2 u=0.
4ξ 4 )
Here α is a constant, E is the total energy, and F is a constant such that Fz is the
potential energy added to the system by the introduction of an electric field.
Using the larger root of the indicial equation, develop a power-series solution about ξ =
0. Evaluate the first three coefficients in terms of a 0.
Ans :
The point P ξ = 0 is a regular singularity of the ODE. The trial solution ∑ j a j ξ k+ j yields
the given indicial equation. For k = m/2, a 0 ≠ 0 0 and non-negative m we set the
coefficient of the term ξ k+ 1 to zero. This gives a 1. = −αa 1/(m+1). Setting the coefficient
of ξ k+ 2 to zero gives the second given term, etc.
m2
Indicial equation. s2 − = 0,
4
m
α2 α2
{
u ( ξ )=a 0 ξ 2 1− ξ
[ −
E
m+1 2 ( m+1 ) ( m+ 2 ) 4 ( m+ 2 )] }
ξ 2 +… .

7.5.8 For the special case of no azimuthal dependence, the quantum mechanical analysis of
the hydrogen molecular ion leads to the equation
d du
dη [
( 1−η2 )
dη ]
+ α u+ βη2 u=0

Develop a power-series solution for u(η). Evaluate the first three nonvanishing
coefficients in terms of a 0.
Ans:
 ∞
j+k
Substituting ∑ a j η ∧its derivatives into
j=0

d du
1−1−η2 ¿ + α u+ β u 2=0
dη dη
we obtain the recursion relation
a j +2( j+ k +2)( j+k +1)−a j [( j+k )( j+ k +1)−α ]+ β a j−2=0.
For j = −2, a−2 = 0 = a−4 by definition and the indicial equation k(k − 1)a 0 = 0 comes
out, i.e. k = 0 or k = 1 for a 0 6≠0.
For j = −1 with a−3 = 0 = a−1 we have a 1k(k + 1) = 0. If k = 1, then a 1 = 0 implying a 3 =
0 = a 5.
For j = 0, k = 0 we get 2 a2 = −a 0 α and 6a 2 = a 0 (2 − α) for k = 1.
For j = 1, k = 0 we find 6 a 3 = 12a 3 = a 1(6 − α) for k = 1.
Finally, for j = 2, k = 1 we have 20 a 4 − (12 − α)a 2 + βa 0 = 0, giving the expansion listed
in the problem set.
Indicial equation s(s − 1) = 0,

2−α 2 (2−α )(12−α ) β 4

uk=1=a0 η 1+ { 6
η+
120 [
−
20
η +… . ] }
7.5.9 To a good approximation, the interaction of two nucleons may be described by a
mesonic potential
A e−ax
V= ,
x
attractive for A negative. Show that the resultant Schrödinger wave equation
h2 d 2 ψ (
+ E−V ) ψ=0
m d x2
has the following series solution through the first three nonvanishing coefficients:
1 1 1 '
{
ψ=a0 x + A ' x 2 +
2 6 2 [
A −E' −a E' x 3 +…
2

] }
where the prime indicates multiplication by 2m/h2
Ans :
Substituting
ψ=a0 +a1 x+ a2 x 2 +a3 x 3+ … ,
and setting
2 mA ' 2 mE A
A' = , E = 2 , V = e−ax , A< 0 , a>0 , we obtain
h 2
h x
1
[ ]
2 a2 +6 a3 x +…+ − A ' + ( E' + a A ' ) x− A ' a2 x 2+ … ( a1 +a 2 x+ … ) =0
2
 where the coefficients of all powers of x vanish. This implies
a 0=0 , 2 a2= A' a1 ,6 a3 +a1 ( E' +a A' )− A ' a2 =0 , etc
Thus, we get the given series.

7.5.10 If the parameter b2 in Eq. (7.53) is equal to 2, Eq. (7.53) becomes

y } + {1} over {{x} ^ {2}} {y} ^ {'} - {2} over {{x} ^ {2}} y = ¿
From the indicial equation and the recurrence relation, derive a solution y = 1 + 2x +
2x2. Verify that this is indeed a solution by substituting back into the differential
equation.
Ans :
Even though the point x = 0 is an essential singularity we try substituting
∞ ∞
j+ k
, y =∑ a j ( j+k ) x j +k−1 , y } = sum from {j =0} to {∞} {{a} rsub {j} left (j + k right ) left (j + k - 1 right ) {x} ^ {j + k} ¿
'
∑ aj x
j=0 j=0

into our ODE we obtain the recursion relation

a j [ ( j+k ) ( j+k −1 )−2 ] + a j +1 ( j+ k−1 )=0
For j = −1, a-1 = 0 by definition, so k = 0 for a 0≠0 is the indicial equation. For j = 0,
−2a0 + a1 = 0, and for j = 1, −2a 1 + 2a2 = 0, while j = 2 yields a 3 = 0, etc. Hence our
solution is y = a0 (1 + 2x + 2x2), and this is readily verified to be a solution

7.5.11 The modified Bessel function I0(x) satisfies the differential equation
2 d2 ( ) d 2
x I x + x I 0 ( x )−x I 0 ( x ) =0
2 0
dx dx
Given that the leading term in an asymptotic expansion is known to be
ex
I 0 ( x)
√ 2 πx
assume a series of the form
ex
I 0 ( x) { 1+ b1 x −1 +b 2 x−2 +… }
√ 2 πx
Determine the coefficients b1 and b2.
Ans :
ex
Writing the solution to the ODE as f (x)
√ 2 πx
we find that f(x) satisfies the ODE x 2 f” + 2x2 f’ + f /4 = 0. Substituting into this ODE
the series expansion f(x) = b0 +b1/x+b2/x2 +· · · , we find that the bn satisfy the
recurrence formula
 1
n ( n+1 ) +
4
b n+1= bn
2n+ 2
which, with the initial value b0 = 1, we can use to obtain the coefficients in the
asymptotic expansion. The first two coefficients are
b1 = (1/4)/2 = 1/8 and
b2 = (2 + 1/4)b1/4 = 9/128.
7.6.2 The criterion for the linear independence of three vectors A, B, and C is that the
equation
aA+bB+cC=0,
analogous to Eq. (7.54), has no solution other than the trivial a = b = c = 0. Using
components A = (A1, A2, A3), and so on, set up the determinant criterion for the exis-
tence or nonexistence of a nontrivial solution for the coefficients a, b, and c. Show that
your criterion is equivalent to the scalar triple product A · B × C ̸= 0.
Ans : If A, B, C are linearly independent, geometry tells us that their volume (A X
B). C ≠ 0, and vice versa.

7.6.3 Using the Wronskian determinant, show that the set of functions

xn
{
1,
n!
(n=1,2, … , N ) }
Is linearly independent.
Xn ' X n−1
Ans : Using y n= , y n= ,etc. for n = 0, 1, … N we get
n! ( n−1 ) !

x2
W 1= 1
0| | X =1 , W =
1 2
1 x
0 1
0 0
| ||
and, continuing, W2 = … = WN = 1
2 = 1 x =W =1
x
1
0 1 1|

7.6.4 If the Wronskian of two functions y1 and y2 is identically zero, show by direct integration that
y 1=c y 2
that is, that y1 and y2 are linearly dependent. Assume the functions have continuous
derivatives and that at least one of the functions does not vanish in the interval under
consideration.

' y '1 y '2

'
Ans : If W = y 1 y − y y 2=0 ,then = .
2 1
y1 y2
Integrating gives ln y 1=ln y2 + lnC . Hence y 1=C y 2, and vice versa.
7.6.5 The Wronskian of two functions is found to be zero at x 0 −ε ≤ x ≤ x 0 +ε for arbitrarily
small ε > 0. Show that this Wronskian vanishes for all x and that the functions are
linearly dependent.
Ans :
' '
∂ x1 x 1 x 2 −x 1 x 2 −W
( )
∂ t x2
=
x2
2
= 2
x2
∂
( constant )=0
∂t
The Wronskian W(x) is written as a Taylor series at x 0, all of its coefficients must be
zero.

7.6.6 The three functions sin x, e x , and e− x are linearly independent. No one function can be
written as a linear combination of the other two. Show that the Wronskian of sinx, e x ,
and e− x vanishes but only at isolated points.
Ans :
General Equation : x '' + Px' +Qx=0
W =x1 x 2' −x1' x 2
W ' =x 1' x 2 + x 1 x 2' ' −(x 1' ' x 2 + x 1' x2 )
W ' =x 1 x2 ' ' −x 1' ' x 2
w h ere: x 1' ' =−P x 1' −Q x 1 ; x 2'' =−P x 2' −Q x 2
so
❑
W ' =x 1 (−P x 2' −Q x 2 )−(−P x1' −Q x 1) x 2

W ' =−P x 1 x 2' −Q x1 x 2❑+ P x 1' x2 + Q x1 x 2

❑

W ' =−P ( x1 x 2' −x 1' x 2 )

W ' =−PW
∂w
+ PW =0 → W (t )− A e−∫ P ∂ t ≠ 0
∂t
W = 4 sin x,
W = 0 for x = ±nπ, n = 0, 1, 2, . . . .
7.6.7 Consider two functions φ 1 = x dan φ 2 = |x|. Since φ ʹ1= 1 and φ ʹ2= x/|x|, W ¿ ¿ for any
interval, including [ −1 ,+1 ]. Does the vanishing of the wronskian over [ −1 ,+1 ] prove
that φ 1∧φ2 are linierly dependent? Clearly , they are not. what is wrong?
Ans :
Diketahui: φ 1 = x
 φ 2 = |x|

φ ʹ1= 1
φ ʹ2= x/|x|
x
Maka nilai φ ʹ2=
|x|
1
=
|1|
=1
φ ʹ2 ≠ 0
7.6.8 Explain that linier independence does not mean the absense of any dependence,
illustrate your argument with cosh x and e x .
Ans :
1
Persamaan 2 y 1(x)- y 2(x)- = 0,
y 2( x)
Persamaan itu adalah non linier.
7.6.9 Legendre’s differensial equation
(1-x2) y ʺ -2 xy ʹ +n(n+1)y = 0
Has a reguler solution Pn(x) and in irreguler solution Q n(x). Show that the wornskian of
Pn∧Q n is given by
An
Pn(x)Q 'n(x)- P'n(x)Q n(x) =
1−x 2
Ans :
Dik: (1-x2) y ʺ -2 xy ʹ +n(n+1)y = 0
An
Pn(x)Q 'n(x)- P'n(x)Q n(x) =
1−x 2
Maka, Pn Q 'n- P'n Q n = W (x)
An
Pn Q 'n- P'n Q n = An c− ∫ x Pdt =
1−x 2
2t
Karena - ∫x Pdt = ∫x dt
1−t 2
= -ln (1- x2)
7.6.10 Show, by means of the Wronskian, that a linear, second-order, homogeneous ODE of
the form
Y’’(x) + P(x)y’(x) + Q(x)y(x) = 0
cannot have three independent solutions.
 Hint. Assume a third solution and show that the Wronskian vanishes for all x.
7.6.11 Show the following when the linear second-order differential equation py’’+qy’+r y =
0 is expressed in self-adjoint form:
Ans :
(a) The Wronskian is equal to a constant divided by p:
C
W(x) =
p (x)
(b) A second solution y2(x) is obtained from a first solution y1(x) as
X
dt
y2(x) = C y1(x) ∫ ¿
❑ p (t) ¿ ¿

7.6.10 Assuming there to be three linearly independent solutions, construct their Wronskian.
It will be identically zero.
d d
7.6.11 From ( d (x )
p( x)
d ( x) )
q(x ) u = 0 we have

x
dW p
(a) ∫ =−∫ d (x)= ln 1/p + ln C = ln W.
W ❑ p (x )

W (a)
Hence W = with W(a) = C.
p( x )
x
ds 2 d y 1( x )
(b) y2 = W(a)y1 ∫ 2 follows from W(y1, y2) = y1
❑ p ( s ) y 1( s) dx y 2( x )

7.6.12 Transform our linear, second-order ODE

yn + P(x)y’ + Q(x)y = 0
by the substitution

x
y = z exp [
−1
∫ P ( t ) dt
2 ❑ ]
and show that the resulting differential equation for z is
zn + q(x)z = 0,
where q(x)= Q(x) – ½ P’
(x) – ¼ p2 (x)
Note. This substitution can be derived by the technique of Exercise 7.6.25.
 x
’ -1/2
Ans : Using y = z’E, E = e ∫ pdt ,
❑

z z
y’ = z’E–½ zP E, yn = znE − P z’E– P’E + P2E, we obtain
2 4
yn + Py’ + Q = E

z z
[ z n− p ' − p 2+Qz = 0.
2 4 ]
7.6.13 Use the result of Exercise 7.6.12 to show that the replacement of ϕ(r) by rϕ(r) may be
expected to eliminate the first derivative from the Laplacian in spherical polar
coordinates. See also Exercise 3.10.34.
∂2 2∂ L2
Ans : Since∇ = 2 +
2
+ we have
∂ r r ∂r r 2

r r x
dr -1/2 1
− ½ ∫ P dt = −∫ = − ln r, e ∫ pdt = , so that ϕ(r)
❑ ❑ r ❑
r

1 d2 L2
= ψ(r)/r. Equivalently∇ ϕ(r) = 2
(rϕ) + 2 ϕ.
r d r2 r

7.6.14 By direct differentiation and subsitution show that

s

y 2 ( x ) = y 1( x)∫
[
exp −∫ P ( t ) dt
❑
2
] ds
❑ [ y 1 (s )]
Satisfies, like y 1 ( x), the ODE
y ' '2 ( x )+ P ( x) y '2 +Q(x ) x y 2 (x)=0
Ans :
s

x −∫ Pdt x

Defining E =∫ e ds,
−∫ Pdt
and using
❑

1 2 E(X ) =e ❑

❑ y 1 ( s)

E PE
y 2= y 1 ( x )E1 , y ' 2= y ' 1 E 1 + , y ' '2 = y '' 1 E 1−
y1 y1
We obtain
y ' '2 + P y ' 2+Q y 2=E 1 ¿
7.6.15 In the equation
 s

y 2 ( x ) = y 1( x)∫
x [
exp −∫ P ( t ) dt
❑
2
] ds
❑ [ y 1 (s )]
y 1 ( x) satisfies
y ' '1 + P( x ) y ' 1 +Q(x) y 1=0
The function y 2 ( x ) is a linearly independent second solution of the same equetion.
Show that the inclusion of lower limits on the two integrals leads to nothing new, that
is, that it generates only an overall constant factor and a constant multiple of the
known solution y 1 ( x )
Ans :
Changing the lower limit from a to b changes the integral that multiplies y2 by a
constant :
s s b

∫ Pdt =¿∫ Pdt +¿∫ Pdt=¿ ¿ ¿ ¿

a b a

And via
s s s

s −∫ Pdt b
x −∫ Pdt b −∫ Pdt
−∫ Pdt
e
∫ ey 2 (s) ds+∫ ey 2(s) ds
a b a

∫ y 2 (s) ds=e a

a 1 b 1 a 1

Adds a constant to y2
7.6.16 Given that one solution of
'' 1 ' m2
R + R − 2 R=0
r r
Is R=r m, show that eq 7.67 predicts a second solution, R=r−m
Answer :
r
r r
ds −r −2m
Using −∫ dr =−lnr , e 1
−∫ Pdt
= , ❑
∫ s . s 2m 2 m , we have
=
❑ r r ❑

m r −2 m −1
y 2=−r =
2 m 2 mrm
7.6.17 Using y1(x) = X∞ n=0 (−1) n (2n + 1)! x 2n+1 as a solution of the linear oscillator
equation, follow the analysis that proceeds through Eq. (7.81) and show that in that
equation cn = 0, so that in this case the second solution does not contain a logarithmic
term.
Ans :
As P = 0, y1 = sin x, e − R x P dt =const. and y2 = sin x Z x ds sin2 s = sin x cot x =
cos x. Using the series expansions with p−1 = 0 = q−2 gives the indicial equation k(k
 − 1) = 0. Thus k = α = 1 = n and y2(x) = y1(x)  c1 ln x + X∞ j=0,j6=1 cj j − 1 x
j−1   . Substituting these y2, y0 2 , y00 2 into the classical harmonic oscillator
ODE yields 2y 0 2   c1 x + X∞ j=0,j6=1 cjx j−2   + y1  − c1 x 2 + X∞
j=0,j6=1 cj (j − 2)x j−3   = 0. The Taylor series for y1 = sin x gives 2c1 − c1 = c1
= 0 for the coefficient of 1/x. Thus, y2 does not contain a term proportional to ln x.
7.6.18 Show that when n is not an integer in Bessel’s ODE, Eq. (7.40), the second solution of
Bessel’s equation, obtained from Eq. (7.67), does not contain a logarithmic term.
Ans :
Since Bessel’s ODE is invariant under n → −n we expect and verify that J−n(x),
defined by its Taylor series, is a solution along with Jn(x). From the lowest power
series coefficients we obtain W(Jn, J−n) = An x = −2 sin πn πx 6= 0, so that they are
independent if n 6= integer. This is Eq. (14.67). The standard series y = X∞ j=0 ajx
j+k leads to the indicial equation [k(k − 1) + k − N 2 ]a0 = 0. For a0 6= 0 we obtain k
= ±N, N ≥ 0. The roots are α = N, n = 2N, consistent with pj = δj,−1, qj = δj0 + N 2 δj,
−2, n − 2α = p−1 − 1, α(α − n) = q−2. The second solution is y2 = y1(x) X∞ j=0 cj Z x
x j−n−1 1 dx1. If n 6=integer there is no ln x term in y2. Since n = 2N, if N is neither
an integer nor half of an odd integer, there is no logarithmic term in y2. It remains for
us to show that when N =half an odd integer there is no ln x term in y2. Since WN (x)
6= 0 for N 6=integer, this is clear from our first part.
7.6.19 (a) One solution of Hermite’s differential equation y 00 − 2x y0 + 2αy = 0 for α = 0 is
y1(x) = 1. Find a second solution, y2(x), using Eq. (7.67). Show that your second
solution is equivalent to yodd (Exercise 8.3.3).
(b) Find a second solution for α = 1, where y1(x) = x, using Eq. (7.67). Show that your
second solution is equivalent to yeven (Exercise 8.3.3).
Ans :
(a) If y1 = 1 for α = 0 then Z x P dt = −x 2 ,and y2 = Z x e s 2 ds, y0 2 = e x 2 , y00 2 =
2xex 2 . Hence y 00 2 − 2xy0 2 = 0. Integrating the power series for e s 2 yields y2(x)
= X∞ j=0 x 2j+1 (2j + 1)j! = X∞ j=0 a2jx 2j+1 with aj+2 aj = 2(j + 1) (j + 2)(j + 3), j
even, which is the recursion for the k = 1 case of Exercise 8.3.3 (a) for α = 0.
(b) If α = 1 then y1 = x is a solution of the ODE, as is easily verified, and y2 = x Z x e
s 2 s 2 ds.
Integrating the power series yields y2(x) = X∞ j=0 a2jx 2j = −1 +X∞ j=1 x 2j (2j −
1)j! with aj+2 aj = 2(j − 1) (j + 1)(j + 2), j even, which is the recursion for k = 0 of
Exercise 8.3.3 (a) for α = 1
7.6.20 One solution of Laguerre’s differential equation
xy ' ' +( 1−x) y ' +ny=0
for n = 0 is y1(x) = 1. Using Eq. (7.67), develop a second, linearly independent
solution. Exhibit the logarithmic term explicitly.
Ans :
x y ' ' + ( 1−x ) y ' +ny =0 : x
( 1−x ) ' ny ( 1−x ) 1
y' '+ y + =0 : x → P ( x )= = −1
x x x x
x2
x
−∫ P ( x 1 ) dx 1
y 1 ( x )∫ e ❑

y 2 ( x )= ❑
2
[ y 1( x 2) ]
x2
x 1
−∫ −1 dx1
❑ x
1∫ e dx 2
y 2 ( x )= ❑

12
x
−lnx+ x
1∫ e dx 2
y 2 ( x )= ❑
2
1
x
y 2 ( x )=∫ e−lnx . e x dx 2
❑
x x
ex ex
y 2 ( x )=∫ dx 2=∫ x dx 2=li ( x )+ c
❑ e lnx ❑

7.6.21 For Laguerre’s equation with n = 0,

x
es
y 2 ( x )=∫ ds
❑ s
(a) Write y2(x) as a logarithm plus a power series.
(b) Verify that the integral form of y2(x), previously given, is a solution of Laguerre’s
equation (n = 0) by direct differentiation of the integral and substitution into the
differential equation.
(c) Verify that the series form of y2(x), part (a), is a solution by differentiating the
series and substituting back into Laguerre’s equation.
Ans :
x ∞
es x xn
a) y 2 ( x )=∫ ds=ln x +1+ +…=ln x +∑
❑ s 2 n=1 n. n !

x s s
b) y 2 ( x )=∫ e ds= y 2' = e
❑ s s
 −e s e s −1 e s −1
s
''
y2 = 2 + =
s s
+1 =
s s
+1 y 2
'
( ) ( )
( −1s +1) y 2
y 2' ' = '

−1
y 2 −( +1 ) y 2 =0
'' '
s
1
y 2 +( −1 ) y 2 =0
'' '
s

x ∞
es x xn
c) y 2 ( x )=∫ ds=ln x +1+ +…=ln x +∑
❑ s 2 n=1 n. n !

∞ ∞
ex x n−1 −1 x n−2
y 2' = =∑ , y 2 ' ' = 2 +∑
x n=0 n ! x n=2 ( n−2 ) ! n
∞
1 x n−2 1 1 1
y 2' ' + ( )
s
−1 y 2' =∑
n=2 ( n−2 ) ! n
+
n(n−1)
−
n−1
=0 ( )
7.6.21 One solution of the Chebyshev equation
( 1−x 2 ) y ' ' −x y ' + n2 y=0
for n = 0 is y1 = 1.
(a) Using Eq. (7.67), develop a second, linearly independent solution.
(b) Find a second solution by direct integration of the Chebyshev equation.
Hint. Let v= y 0
and integrate. Compare your result with the second solution given in
Section 18.4.
Ans :
a) ( 1−x 2 ) y ' ' −x y ' + n2 y=0 : ( 1−x 2 )

'' x y' ' n2 y −x

y − 2
+ 2
=0 → P ( x )=
( 1−x ) ( 1−x ) ( 1−x 2 )
x2
x
−∫ P ( x 1) dx 1
y 1 ( x )∫ e ❑

y 2 ( x )= ❑
2
[ y 1( x 2)]
x2
x −x
−∫ dx1
(1− x2 )
1∫ e ❑

y 2 ( x )= ❑
dx 2
12
 x2
−x 2 du
−∫ dx 1 →misal u=1−x sehingga =−2 x
❑
2
( 1−x ) dx 1
x2
−x du
−∫
❑ u −2 x
x2
−1 1 −1 1
∫ du= ln u=¿− ln ( 1−x2 ) ¿
2 ❑ u 2 2
x −1 2
ln ( 1−x )
2
1∫ e dx 2 x x −1
2 −1/2

y 2 ( x )= ❑
=∫ e ln (1− x ) dx 2=∫(1−x 2) 2 dx 2
12 ❑ ❑

x
1 dx 2
y 2 ( x )=∫ dx 2→ misal x =sin u , =cos u
❑ √ 1−x 2 du
x x x
cos u du cos u du
y 2 ( x )=∫ 2
=∫ 2
=∫ du=u ,
❑ √ 1−sin u ❑ √ cos u ❑

x=sin u sehinggau=arc sinx=sin −1 x

y 2 ( x )=u=arc sin u=sin −1 x

b) y ' =v , → y '' =v '

( 1−x 2 ) y ' ' −x y ' =0 → ( 1−x 2 ) v ' −xv=0
( 1−x 2 ) dv −xv=0
dx

( 1−x 2 ) dv =xv → dv = x 2 dx
dx v 1−x
dv x du
∫ =∫ dx misal u=1−x 2 , =−2 x
v 1−x 2
dx
dv x du
∫ v
=∫
u (−2 x )
dv −1 du
∫ v
=
2 ∫ u
−1
−1
ln v= ln u → ln v=¿ ln(1−x 2) 2 ¿
2
−1 −1
v=(1−x 2) 2 → ∫ v =∫ (1−x2 ) 2 dx=sin−1 x

7.6.23 One solution of the Chebyshev equation

( 1−x 2 ) y ' ' −x y ' + n2 y=0
 For n=1is y 1 ( x )=x . Set up the Wronsjian doulbe integral solution and derive a second
solution, y 2 ( x )
Ans : The value of exp(∫ Pdx ¿ ¿ has the same value as in Exercise 7.6.22, namely
−1
( 1−x 2 ) 2 . Therefore our solution y 2 ¿is
x
du 2 1/ 2
y 2=x ∫ =−( 1−x )
2 1/ 2
u2 ( 1−u )
7.6.24 The radial Schrödinger wave equation for a spherically symmetric potential can be
writen in the form
−h2 d 2 h2
[ 2 m d r2
+l ( l+1 )
2 mr 2 ]
+V ( r ) y ( r ) =Ey(r)

The potensial energy V ( r ) may be expanded about the origin as

b−1
V ( r )= +b 0+ b1 r +…
r
(a)Show that there is one (regular) solution y 1 (r ) starting with r l +1.
(b)From Eq.(7.69) show that the irregular solution y 2 (r ) diverges at the origin as r −1 .
Ans : (a) Rescale the ODE by multiplying by 2 m/h2 so that
' 2
E' =2 mE /h2, b−1=2 mb 1 /h , etc.
(b) The indicial equation has roots
2
√
−( P−1−1 ) ∓ ( P−1−1 ) −4 q−2
2
With the P−1=0 and q−2=−l ( l+ 1 ) .The root for the regular solution is α 1=l+1 and that
of the irregular solution is α 2=−l . Since P(r )=0 we have
r
ds
y 2 ( r ) = y 1 ( r )∫ 2
y1 ( s)
This leads to y 2 (r ) r −1 [ 1+O(r ) ] as well.

7.6.25 Show that if a second solution, y 2 , is assumed to be related to the first solution, y 1,
according to y 2 ( x ) f ( x), substitution back into the original equation
y '2' + P ( x ) y '2 +Q ( x ) y 2=0
Leads to
 s

f ( x )=∫
x [
exp −∫ P ( t ) dt
] ds ,
[ y 1 ( s )2 ]
in agreement with Eq. (7.67)
Ans : y '2= y 1 f implies

y '2= y 1 f + y 1 f ', y '2' = y '1' f +2 y '1 f ' + y 1 f ' '

And so
y '2' + P y '1 +Q y2 =f ( y '1' + P y '1 +Q y 1 ) + P y 1 f ' +2 y '1 f ' + y 1 f ' ' =0

Thus f ' ' y 1+ f ' ¿.

Separating variables and integrating yields
x
x
−∫ Pdt
1
¿ f ' =−2 ln y 1−∫ Pdt , f ' = 2
e , and f as given.
y1 ( x )

7.6.26 (a) Show that

n 1−a2
y + y=0
4 x2
Has two solutions :
y 1 ( x )=a0 x(1+a)/ 2 ,
y 2 ( x ) =a0 x(1+a)/ 2 .
(b) For α =0 the two linearly independent solutions of part (a) reduce to the single
solution y 1 ' =a0 x(1+a )/ 2. Using Eq. (7.68) derive a second solution,
y 2 ' ( x )=a 0 x (1 +a)/2 ln x .
Verify that y 2 ' is indeed a solution.
(c) Show that the second solution from part (b) may be obtained as a limiting case
from the two solutions of part (a) :
y 1− y 2
y 2 ( x )=lim (
' ).
α →0 α
Ans :
(a) From y 1 ( x )=a0 x(1+a)/ 2 , we have
(α −1)
a0
y ' 1= ( 1+ α ) x 2 ,
2
y } rsub {1} = {{a} rsub {0}} over {4} left ({α} ^ {2} - 1 right ) {x} ^ {{( α - 3)} over {2}} ¿
 Hence
y } rsub {1} + {1 - {α} ^ {2}} over {4 {x} ^ {2}} {y} rsub {1} = {{a} rsub {0}} over {4} left ({α
Similarly,

−(α +1)
(1−a )/ 2 a0
y 2=a0 x , y ' 2= ( 1−α ) x 2 , y } rsub {2} = {{a} rsub {0}} over {4} left ({α} ^ {2} - 1
2
Hence

y } rsub {2} + {1 - {α} ^ {2}} over {4 {x} ^ {2}} {y} rsub {2} = {{a} rsub {0}} over {4} {x} ^ {- {( α -
(1−α 2 )
Alternatively, a solution y x p leads to p ( p−1 ) + =0 with the roots
4
( 1± α )
p= .
2
x
x
1 /2 −∫ Pdt
(b) y 10=a 0 x , P=0give ∫ Pdt =0 , e ❑
=1. Hence
❑
x
ds 1 1 /2
y 20=a 0 x 1 /2∫ 2
= x ln x .
❑ a0 s a0
(c) L’H^o pital’s rule gives
(α +1)/ 2
−x (−α +1)/2
y 1− y 2 lim x
α →0 1
lim = = x1 /2 ln x .
α →0 α α 2

7.7.1 If our linear, second-order ODE is inhomogeneous, that is, of the form of Eq.
(7.94), the most general solution is
y ( x) = y1 ( x ) + y2 ( x ) + y p ( x) ,
Where y 1 and y 2 are independent solutions of the homogeneous equation.
Show that
x x
y 1 ( s ) F ( s ) ds y 2 ( s ) F ( s ) ds
y p ( x ) = y 2 ( x )∫ − y1 ( x)∫ ,
❑ W { y 1 ( s ) , y 2 ( s) } ❑ W { y 1 ( s) , y2 ( s) }
With W{ y 1 ( x ) , y 2 ( x ) } the Wronskian of y 1 ( s )∧ y 2 ( s ) .
Find the general solutions to the following inhomogeneous ODEs:
Ans :
Denoting
x x
y 1 Fds y 2 Fds
E1=∫ , E 2=∫
❑ W ( y 1 , y 2) ❑ W ( y1 , y2 )
 We check that
' y2 y1 F y1 y2 F ' ' ''
y ' p= y ' 2 E 1− y 1 E2 + − = y 2 E1− y 1 E2 , y p=¿
W W
F '
y ' '2 E1− y ' ' 1 E2 + ( y y − y y ' )= y '' 2 E1− y ' '1 E2 + F
W 2 1 2 1
Hence
y ' ' p + P y ' p +Q y p=E1 ( y ' '2 + P y ' 2+Q y 2 ) −E2 ( y ' ' 1 + P y' 1 +Q y 1) + F=F .
This is the generalization of the variation of the constant method of solving
inhomogeneous first-order ODEs to second-order ODEs. Find the general
solutions to the following inhomogeneous ODEs:
7.7.2 y n + y=1
Ans :
We need the general solution to the related homogeneous equation and a particular
integral of the complete inhomogeneous ODE. The homogeneous equation y n + y=0
has solution y 1 ¿ cos x and y 2¿ sin x. We might be able to guess a particular integral
( y=1 ) but we can also use the method of variation of parameters. This method
assumes a particular integral of the form y ( x ) =u 1 ( x ) y 1 ( x ) +u 2 ( x ) y 2(x), and leads to
the two equations
u' 1 y 1+ u' 2 y 2=u' 1 cos x +u' 2sin x=0
u' 1 y ' 1=−u' 1 sin x+ u' 2 cos x=1
These equation have solution u' 1=−sin x , u' 1=cos x ; these can be integrated to obtain
u 1=cos x, u 2=sin x. Inserting these into the expression for y(x), we get
y ( x ) =cos2 x+ sin2 x=1 . The general solution to the original ODE is therefore
c 1cos x + c 2 sin x +1.
7.7.3 y n +4 y=e x
Ans :
Following the strategy and notation of the answer to Exercise 7.7.2, we find

(2 x ) (2 x )
y 1=cos 2 x, y 2=sin 2 x, from which we find u' 1=−e x sin ' x
and u 2=e sin .
2 2

ex ex
we integrate these expressions to find u 1= ( )
10
¿, u 2= ( )
10
¿, so
 ex
y=u 1 y 1+u 2 y 2= . The original ODE has general solution
5

ex
c 1cos 2 x +c 2sin 2 x+ .
5
7.7.4 y } - 3 {y} ^ {'} +2 y = sin {x ¿
Ans :
Following the strategy and notation of the answer to exercise 7.7.2, we find

y 1=e x , y 2=e 2 x ,from which we find u ' 1=−e−x sin x and u ' 2=e−2 x sin x .We integrate

e−x 3 cosx + sinx

these expressions to find u ' 1=( ) ¿ so y=u1 y 1+u 2 y 2= . The original
2 10

x 2x 3 cosx+ sinx
ODE has general solution c 1 e + c 2 e +
10
7.7.5 xy } - left (1+ x right ) {y} ^ {'} + y = {x} ^ {2 ¿

Ans :
Following the strategy and notation of the answer to exercise 7.7.2, we find y 1=x +1 ;

using te wronskian method we get the second solution y 2=e2 . Remembering that the
inhomogeneous term is to be determined when the original ODE is in standard form
(coefficient of y ¿ equal to 1) we set up the equations for the u ' i and find

u ' 1=−1, u' 2=( x +1 ) e−x , so u1=−x and u2=−( x+2 ) e− x . Thus, y=

u1 y 1+u 2 y 2=− ( x2 +2 x+ 2 ) .We can, without generating an error, remove from y the
2 x+2 since it is just 2 y 1. Thus, the original ODE has general solution

c 1 ( x +1 ) +c 2 e2 x −x 2.
7.8.1 Consider the Riccati equation y ´ = y 2 − y − 2. A particular solution to this equation is
y = 2. Find a more general solution.
Ans :
A more general solution to this Riccati equation is y = 2 + u, where u is a general
solution to the Bernoulli equation u´ = 3u + u2 . See Eq. (7.104). In the notation of Eq.
(7.101), p = 3, q = 1, and n = 2, and the Bernoulli equation has solution u = 1/v, where

1
v is a solution of v´ +3v = −1, namely v = Ce−3 x + . Therefore u = 3/(Ce−3 x − 1) and
3
y = 2 + 3/(Ce−3 x − 1).

7.8.2 A particular solution to y ´ = y 2/ x 3 − y/x +2x is y = x 2 . Find a more general solution.

 Ans :
A more general solution to this Riccati equation is y = x 2 + u, where u is a general
solution to the Bernoulli equation u´ = u2/ x 3 + u/x. See Eq. (7.104). In the notation of
Eq. (7.101), p = 1/x, q = 1/ x 3 , and n = 2, and the Bernoulli equation has solution u =
1/v, where v is a solution of v´ + v = −1/ x 3 , namely v = (Cx + 1)/x2 . Therefore u = x 2

2 x2 C x3
(Cx + 1) andy = x + =
C x +1 Cx+1

7.8.3 Solve the Bernouli equation y ' + xy=xy 3.

Ans : This ODE corresponds to Eq. (7.101) with p=−x , q+ x , and n=3.Thus, with
u= y−2, Eq. (7.102) becomes u' −2 xu=−2 x . The homoge- neous equation for u has
solution e, and from the method of variation of parameters or by inspection, a
particular integral of the inhomogeneous equation is u = 1. Thus the general solution
2 2
for u is u = Ceu=C e x +1. Since y=u−1/ 2, the general solution for y is y=1 / √C e x +1

7.8.4 ODEs of the form y=x y ' + f ( y ' ) are known as Clairaut equations. The first step in
solving an equation of this type is to differentiate it, yielding
y ' = y ' + x y ' ' + f ' ( y ' ) y ' ' ,∨ y ' ' ( x +f ' ( y ' ) )=0
Solutions may therefore be obtained both from y ' ' =0and from f ' ( y ' )=−x . The so
called general solution comes from y ' ' =0
for f ( y ' ) =( y ' )2,
(a) Obtain the general solution (note that it contains a single constant).
(b) Obtain the so-called singular solution contains no adjustable constans.
Note. The singular solution is the envelope of the general solutions
Ans :
(a) The general solution comes from y" = 0, and therefore has the form y = ax + b.
However, not all values of a and b lead to solutions of the original Clairaut equation.
Substituting into y=x y ' +( y ' )2, we find ax + b = xa + a², which shows that y is a
solution only if b = a².
(b) The singular solution comes from 2y' = -x, which integrates to y = -x2/4 + C.
Substituting into y = xy' + (y')², we get -x²/4 + C = x(-x/2) + x² /4, which shows that
this y is a solution only if C = 0. The singular and a general solution coincide only if
−x 20 / 4=αxo+a 2, the solution to which is xo=−2 a. At xo, both solutions have slope a,
so the singular solution is tangent to each instance of the general solution and is
therefore referred to as their envelope.