You are on page 1of 26

Infinite Series and Differential Equations

Nguyen Thieu Huy

Hanoi University of Science and Technology

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 1 / 10


SECOND-ORDER LINEAR DIFFERENTIAL EQUATIONS
1. Definitions and Notations
1.1 Definition
A second-order differential equation is called linear if it can be written in
the form
y 00 + p(t)y 0 + q(t)y = r (t). (1)

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 2 / 10


SECOND-ORDER LINEAR DIFFERENTIAL EQUATIONS
1. Definitions and Notations
1.1 Definition
A second-order differential equation is called linear if it can be written in
the form
y 00 + p(t)y 0 + q(t)y = r (t). (1)

Remark.
1 We often see the equation P(t)y 00 + Q(t)y 0 + G (t)y = R(t). Of

course, if P(t) 6= 0, we can divide this Eq. by P(t) and thereby obtain
Eq. (1) with p(t) = Q(t) G (t) R(t)
P(t) , q(t) = P(t) , r (t) = P(t) .

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 2 / 10


SECOND-ORDER LINEAR DIFFERENTIAL EQUATIONS
1. Definitions and Notations
1.1 Definition
A second-order differential equation is called linear if it can be written in
the form
y 00 + p(t)y 0 + q(t)y = r (t). (1)

Remark.
1 We often see the equation P(t)y 00 + Q(t)y 0 + G (t)y = R(t). Of

course, if P(t) 6= 0, we can divide this Eq. by P(t) and thereby obtain
Eq. (1) with p(t) = Q(t) G (t) R(t)
P(t) , q(t) = P(t) , r (t) = P(t) .
2 In trying to solve Eq. (1), we restrict ourselves to interval I in which

p, q, and r are continuous functions on I , say, I = (α, β). The cases


α = −∞, or β = ∞, or both, are included. The function p(t) and
q(t) are called the coefficients of the Eq. (1).

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 2 / 10


SECOND-ORDER LINEAR DIFFERENTIAL EQUATIONS
1. Definitions and Notations
1.1 Definition
A second-order differential equation is called linear if it can be written in
the form
y 00 + p(t)y 0 + q(t)y = r (t). (1)

Remark.
1 We often see the equation P(t)y 00 + Q(t)y 0 + G (t)y = R(t). Of

course, if P(t) 6= 0, we can divide this Eq. by P(t) and thereby obtain
Eq. (1) with p(t) = Q(t) G (t) R(t)
P(t) , q(t) = P(t) , r (t) = P(t) .
2 In trying to solve Eq. (1), we restrict ourselves to interval I in which

p, q, and r are continuous functions on I , say, I = (α, β). The cases


α = −∞, or β = ∞, or both, are included. The function p(t) and
q(t) are called the coefficients of the Eq. (1).
3 Eq. (1) is called homogeneous if the term r (t) = 0 ∀t. Otherwise, the

equation is called nonhomogeneous. Therefore, the term r (t) is called


the nonhomogeneous term.
Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 2 / 10
Examples:
1 y 00 + 4y = e −t sin t is nonhomogeneous second-order linear equation.
2 y 00 − 2ty 0 + 6(1 − t 2 )y = 0 is homogeneous one.

1.2 Definition: Solution


A solution of a second-order linear differential equation (1) on some
interval I is a function y = h(t) that is twice differentiable and satisfies
(1) for all t ∈ I .

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 3 / 10


Examples:
1 y 00 + 4y = e −t sin t is nonhomogeneous second-order linear equation.
2 y 00 − 2ty 0 + 6(1 − t 2 )y = 0 is homogeneous one.

1.2 Definition: Solution


A solution of a second-order linear differential equation (1) on some
interval I is a function y = h(t) that is twice differentiable and satisfies
(1) for all t ∈ I .

1.3 Existence and Uniqueness Theorem:


Consider the initial-value problem

y 00 + p(t)y 0 + q(t)y = r (t), y (t0 ) = y0 , y 0 (t0 ) = y1 ; (IVP)

where p, q, and r are continuous on an open interval I . Then there exists


a unique solution y = ϕ(t) of this problem, and the solution exists
throughout the interval I .

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 3 / 10


2. Theory for Solutions of Linear Homogeneous Equations:
Consider linear homogeneous equation on an interval I :
y 00 + p(t)y 0 + q(t)y = 0 (2)
with continuous coefficients p, q and initial conditions
y (t0 ) = y0 ; y 0 (t0 ) = y1 for given y0 , y1 , and some fixed t0 ∈ I .
2.1 Linearity Principle
If y1 and y2 are two solutions of the differential equation (2), then the
linear combination c1 y1 + c2 y2 is also a solution for any values of the
constants c1 and c2 .
PROOF. The assertion follows from the following direct substitution:
(c1 y1 (t) + c2 y2 (t))00 + p(t)(c1 y1 (t) + c2 y2 (t))0 + q(t)(c1 y1 (t) + c2 y2 (t)) =
c1 (y1 (t)00 +p(t)y1 (t)0 +q(t)y1 (t))+c2 (y2 (t)00 +p(t)y2 (t)0 +q(t)y2 (t)) = 0.

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 4 / 10


2. Theory for Solutions of Linear Homogeneous Equations:
Consider linear homogeneous equation on an interval I :
y 00 + p(t)y 0 + q(t)y = 0 (2)
with continuous coefficients p, q and initial conditions
y (t0 ) = y0 ; y 0 (t0 ) = y1 for given y0 , y1 , and some fixed t0 ∈ I .
2.1 Linearity Principle
If y1 and y2 are two solutions of the differential equation (2), then the
linear combination c1 y1 + c2 y2 is also a solution for any values of the
constants c1 and c2 .
PROOF. The assertion follows from the following direct substitution:
(c1 y1 (t) + c2 y2 (t))00 + p(t)(c1 y1 (t) + c2 y2 (t))0 + q(t)(c1 y1 (t) + c2 y2 (t)) =
c1 (y1 (t)00 +p(t)y1 (t)0 +q(t)y1 (t))+c2 (y2 (t)00 +p(t)y2 (t)0 +q(t)y2 (t)) = 0.
2.2 Definition: Linear Independence of Solutions:
The two solutions y1 and y2 are called linearly independent on I if
k1 y1 (t) + k2 y2 (t) = 0 ∀t ∈ I implies k1 = k2 = 0;
and we call y1 and y2 linearly dependent on I if there exist k1 , k2 , not both
zero, such that k1 y1 (t) + k2 y2 (t) = 0 ∀t ∈ I .
Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 4 / 10
In the case that y1 and y2 are linearly dependent (and only in this case),
we have that y1 and y2 are proportional, that is y1 = − kk21 y2 if k1 6= 0, or
y2 = − kk21 y1 if k2 6= 0.

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 5 / 10


In the case that y1 and y2 are linearly dependent (and only in this case),
we have that y1 and y2 are proportional, that is y1 = − kk21 y2 if k1 6= 0, or
y2 = − kk21 y1 if k2 6= 0.
Since, y1 and y2 are linearly independent if and only if they are not linearly
dependent, we obtain that, two solutions y1 and y2 are linearly
independent if and only if they are not proportional (i.e. yy12 is not a
constant, or yy21 is not a constant).

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 5 / 10


In the case that y1 and y2 are linearly dependent (and only in this case),
we have that y1 and y2 are proportional, that is y1 = − kk21 y2 if k1 6= 0, or
y2 = − kk21 y1 if k2 6= 0.
Since, y1 and y2 are linearly independent if and only if they are not linearly
dependent, we obtain that, two solutions y1 and y2 are linearly
independent if and only if they are not proportional (i.e. yy12 is not a
constant, or yy21 is not a constant).
Example. y1 (t) = e t and y2 (t) = e −2t are linearly independent (on R),
because they are not proportional, yy12 = e 3t , not a constant.

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 5 / 10


In the case that y1 and y2 are linearly dependent (and only in this case),
we have that y1 and y2 are proportional, that is y1 = − kk21 y2 if k1 6= 0, or
y2 = − kk21 y1 if k2 6= 0.
Since, y1 and y2 are linearly independent if and only if they are not linearly
dependent, we obtain that, two solutions y1 and y2 are linearly
independent if and only if they are not proportional (i.e. yy12 is not a
constant, or yy21 is not a constant).
Example. y1 (t) = e t and y2 (t) = e −2t are linearly independent (on R),
because they are not proportional, yy12 = e 3t , not a constant.

2.3 Definition
The Wronski Determinant (or Wronskian) of the two solutions y1 and y2
y1 y2
of Eq. (2) is defined by W (y1 , y2 ) = 0 = y1 y 0 − y 0 y2 .
y1 y20 2 1

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 5 / 10


2.4. Theorem
Let y1 and y2 be two solutions of the equation (2) on an interval I . Then,
the following assertions are equivalent.
(i) y1 and y2 are linearly independent.
(ii) ∃t0 ∈ I such that W (y1 , y2 )(t0 ) 6= 0.
(iii) W (y1 , y2 )(t) 6= 0 ∀t ∈ I .

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 6 / 10


2.4. Theorem
Let y1 and y2 be two solutions of the equation (2) on an interval I . Then,
the following assertions are equivalent.
(i) y1 and y2 are linearly independent.
(ii) ∃t0 ∈ I such that W (y1 , y2 )(t0 ) 6= 0.
(iii) W (y1 , y2 )(t) 6= 0 ∀t ∈ I .
 0
y y 0 −y 0 y W (y1 ,y2 )
Proof. We have yy12 = 1 2y 2 1 2 = .
So, y1 and y2 are linearly
y12
1  0
independent if and only if yy21 6= constant ⇐⇒ yy12 6≡ 0 ⇐⇒ ∃t0 ∈ I such
that W (y1 , y2 )(t0 ) 6= 0. Therefore (i)⇐⇒ (ii).

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 6 / 10


2.4. Theorem
Let y1 and y2 be two solutions of the equation (2) on an interval I . Then,
the following assertions are equivalent.
(i) y1 and y2 are linearly independent.
(ii) ∃t0 ∈ I such that W (y1 , y2 )(t0 ) 6= 0.
(iii) W (y1 , y2 )(t) 6= 0 ∀t ∈ I .
 0
y y 0 −y 0 y W (y1 ,y2 )
Proof. We have yy12 = 1 2y 2 1 2 = y12
.
So, y1 and y2 are linearly
1  0
independent if and only if yy21 6= constant ⇐⇒ yy12 6≡ 0 ⇐⇒ ∃t0 ∈ I such
that W (y1 , y2 )(t0 ) 6= 0. Therefore (i)⇐⇒ (ii).
Clearly (iii)⇒ (ii). Now, we prove (ii)⇒ (iii). Indeed, firstly we have y1 , y2
satisfy equation y 00 + p(t)y 0 + q(t)y = 0. Then we compute
W (y1 , y2 ) = y1 y20 − y10 y2 . So,
W (y1 , y2 )0 = y10 y20 + y1 y200 − y100 y2 − y10 y20 = y1 y200 − y100 y2 = y1 (−p(t)y20 −
q(t)y2 ) − (−p(t)y10 − q(t)y1 )y2 = −p(t)(y1 y20 − y10 y2 ) = −p(t)W (y1 , y2 ).
Therefore,
W (y1 , y2 )0 = −p(t)W (y1 , y2 ) ⇔ W (y1 , y2 )0 + p(t)W (y1 , y2 ) = 0.
Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 6 / 10
R
This is linear equation, and we have W (y1 , y2 )(t) = Ce − p(t)dt . Since
W (y1 , y2 )(t0 ) 6= 0, itR follows that C 6= 0. Thus,
W (y1 , y2 )(t) = Ce − p(t)dt 6= 0 ∀t ∈ I . So (ii)⇒ (iii). Proof is finished.

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 7 / 10


R
This is linear equation, and we have W (y1 , y2 )(t) = Ce − p(t)dt . Since
W (y1 , y2 )(t0 ) 6= 0, itR follows that C 6= 0. Thus,
W (y1 , y2 )(t) = Ce − p(t)dt 6= 0 ∀t ∈ I . So (ii)⇒ (iii). Proof is finished.

2.5 Theorem (Existence of Linearly Independent Solutions)


Consider Eq. (2) with continuous coefficients p, q on I . Then there exist
two linearly independent solutions y1 , y2 on I of Eq. (2).

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 7 / 10


R
This is linear equation, and we have W (y1 , y2 )(t) = Ce − p(t)dt . Since
W (y1 , y2 )(t0 ) 6= 0, itR follows that C 6= 0. Thus,
W (y1 , y2 )(t) = Ce − p(t)dt 6= 0 ∀t ∈ I . So (ii)⇒ (iii). Proof is finished.

2.5 Theorem (Existence of Linearly Independent Solutions)


Consider Eq. (2) with continuous coefficients p, q on I . Then there exist
two linearly independent solutions y1 , y2 on I of Eq. (2).

PROOF. By Theorem 1.3, there exists solution y1 of Eq. (2) satisfying


y1 (t0 ) = 1, y10 (t0 ) = 0 for some t0 in I . Also, there exists solution y2 of Eq.
(2) satisfying y2 (t0 ) = 0, y20 (t0 ) = 1. Therefore, W (y1 , y2 )(t0 ) = 1 6= 0.
Hence, y1 and y2 are linearly independent.

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 7 / 10


R
This is linear equation, and we have W (y1 , y2 )(t) = Ce − p(t)dt . Since
W (y1 , y2 )(t0 ) 6= 0, itR follows that C 6= 0. Thus,
W (y1 , y2 )(t) = Ce − p(t)dt 6= 0 ∀t ∈ I . So (ii)⇒ (iii). Proof is finished.

2.5 Theorem (Existence of Linearly Independent Solutions)


Consider Eq. (2) with continuous coefficients p, q on I . Then there exist
two linearly independent solutions y1 , y2 on I of Eq. (2).

PROOF. By Theorem 1.3, there exists solution y1 of Eq. (2) satisfying


y1 (t0 ) = 1, y10 (t0 ) = 0 for some t0 in I . Also, there exists solution y2 of Eq.
(2) satisfying y2 (t0 ) = 0, y20 (t0 ) = 1. Therefore, W (y1 , y2 )(t0 ) = 1 6= 0.
Hence, y1 and y2 are linearly independent.
2.6 Theorem (Structure of Solutions to Homogeneous Equations)
Consider Eq. (2) with continuous coefficients p, q on I. Let y1 and y2 be
two linearly independent solutions of Eq. (2) on I . Then the general
solution (the set of all solutions) of Eq. (2) is

{c1 y1 + c2 y2 | c1 and c2 are arbitrary constants} (3)


Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 7 / 10
PROOF. Clearly, for any fixed constants c1 and c2 , the formula in (3)
represents a solution of Eq. (2) on I . Let now Y be an arbitrary solution of
(2) on I . Put Y (t0 ) = k1 , Y 0 (t0 ) = k2 for some fixed t0 in I . Consider
y = c1 y1 + c2 y2 . We will find c1 , c2(such that y (t0 ) = k1 , y 0 (t0 ) = k2 .
c1 y1 (t0 ) + c2 y2 (t0 ) = k1
These conditions are equivalent to
c1 y10 (t0 ) + c2 y20 (t0 ) = k2 .
This system has a solution (c10 , c20 ) since its determinant of coefficients is
W (y1 , y2 )(t0 ) 6= 0 (because y1 , y2 are linearly independent). Theorem 1.3
now yields that Y (t) = c10 y1 (t) + c20 y2 (t) for all t in I . Therefore, Y has
a form represented in the set (3).

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 8 / 10


PROOF. Clearly, for any fixed constants c1 and c2 , the formula in (3)
represents a solution of Eq. (2) on I . Let now Y be an arbitrary solution of
(2) on I . Put Y (t0 ) = k1 , Y 0 (t0 ) = k2 for some fixed t0 in I . Consider
y = c1 y1 + c2 y2 . We will find c1 , c2(such that y (t0 ) = k1 , y 0 (t0 ) = k2 .
c1 y1 (t0 ) + c2 y2 (t0 ) = k1
These conditions are equivalent to
c1 y10 (t0 ) + c2 y20 (t0 ) = k2 .
This system has a solution (c10 , c20 ) since its determinant of coefficients is
W (y1 , y2 )(t0 ) 6= 0 (because y1 , y2 are linearly independent). Theorem 1.3
now yields that Y (t) = c10 y1 (t) + c20 y2 (t) for all t in I . Therefore, Y has
a form represented in the set (3).
2.7 Definition
A basis (or a fundamental set) of solutions of Eq. (2) on an interval I is a
pair y1 , y2 of linearly independent solutions of Eq. (2) on I .

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 8 / 10


Remark.
1 To solve the homogeneous equation (2) is to find a basic of solutions
y1 , y2 of Eq. (2). Then, the general solution is y = c1 y1 + c2 y2 ∀
constants c1 , c2 .
2 However, for most problems of the form (2), it is not possible to write
down a useful expression for the solution. This is a major difference
between first order and second order linear equations.

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 9 / 10


Remark.
1 To solve the homogeneous equation (2) is to find a basic of solutions
y1 , y2 of Eq. (2). Then, the general solution is y = c1 y1 + c2 y2 ∀
constants c1 , c2 .
2 However, for most problems of the form (2), it is not possible to write
down a useful expression for the solution. This is a major difference
between first order and second order linear equations.
Example. Solve
y 00 − 3y 0 + 2y = 0. (4)
To solve this equation, we remember that a first-order linear differential
equation y 0 + ky = 0 with constant coefficient k has an exponential
function as a solution, y = e −kt . This gives us the idea to try as a solution
of (4) the function y = e ct . Substituting into (4) we obtain
(c 2 − 3c + 2)e ct = 0, this is equivalent to c = 1 or c = 2. We then obtain
the two following linearly independent solutions of (4): y1 = e t and
y2 = e 2t . Therefore, the general solution of (4) is y = c1 e t + c2 e 2t .

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 9 / 10


2.8. Reduction of Order
Consider
y 00 + p(t)y 0 + q(t)y = 0. (2)
If a nontrivial solution y1 of (2) is known, then we can find the solution y2
linearly independent with y1 by the following way. Looking back at R
− p(t)dt
, so W (y1y,y2 2 )(t) = Ce y 2
R
Wronskian we have W (y1 , y2 )(t) = Ce − p(t)dt
 0 R 1 1
y2 Ce − p(t)dt y2 R Ce − R p(t)dt
⇔ y1 = y2
. Thus, y1 = y2
dt. Choosing C = 1 we have
1 1
R
e− p(t)dt
Z
y2 = y1 dt (5)
y12
(Liouville formula for reduction of order).

Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 10 / 10


2.8. Reduction of Order
Consider
y 00 + p(t)y 0 + q(t)y = 0. (2)
If a nontrivial solution y1 of (2) is known, then we can find the solution y2
linearly independent with y1 by the following way. Looking back at R
− p(t)dt
, so W (y1y,y2 2 )(t) = Ce y 2
R
Wronskian we have W (y1 , y2 )(t) = Ce − p(t)dt
 0 R 1 1
y2 Ce − p(t)dt y2 R Ce − R p(t)dt
⇔ y1 = y2
. Thus, y1 = y2
dt. Choosing C = 1 we have
1 1
R
e− p(t)dt
Z
y2 = y1 dt (5)
y12
(Liouville formula for reduction of order).
Example: Solve (t 2 − 1)y 00 − 2ty 0 + 2y = 0 given a solution y1 = t. To
use the formula (5) we write the equation in the standard form
y 00 − (t 22t−1) y 0 + (t 22−1) y = 0. Applying (5) we have
R e (t 22t−1) dt
R

y2 = t t2
dt = t 2 + 1 for |t| > 1. Therefore, the general solution
of given equation is y = c1 t + c2 (t 2 + 1).
Nguyen Thieu Huy (HUST) Infinite Series and Diff. Eq. 10 / 10

You might also like