ENGR 3621 Advanced Engineering Mathematics: Higher Order Differential Equations

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 55

ENGR 3621

Advanced Engineering Mathematics

Higher Order Differential Equations


(chapter 3 in Zill)

1
➢ Higher-order Linear ODEs: IVP, BVP, (non)homogeneous.
Methods of solutions:
➢ Reduction of Order
➢ Homogeneous Equations with Constant Coefficients
➢ Undetermined Coefficients
➢ Variation of Parameters
➢ Cauchy-Euler Equations
➢ Nonlinear Equations
➢ Linear Models: Initial-Value Problem
➢ Linear Models: Boundary-Value Problems
➢ Greens’s Function
➢ Nonlinear Models
2
Overview
Theory of Linear Equations Cauchy-Euler
Concepts Equations

➢ IVP and BVP ➢ Linear Eq. with variable coeff.


➢ Superposition Principle ➢ Homogeneous Solutions
➢ Linear Dependence and Independence (Wronskian) ➢ Nonhomogeneous Solutions
➢ Homogeneous and Nonhomogeneous
Examples of
Linear Models
Reduction of Order
➢ Spring/Mass Systems
Homogeneous Solutions o Undamped Motion
o Free Damped Motion
➢ Linear Eq. with constant coeff. o Driven Motion
➢ nth degree polynomial equation
Green’s Functions
Nonhomogeneous Solutions
Nonlinear Functions
➢ Method of Undetermined Coefficients
➢ Variation of Parameters ➢ Linearization

3
3.1 Initial Value Problem

Concepts of initial-value problems (IVPs),


introduced in Section 1.2, are applicable to nth-
order linear DEs:
𝑎𝑛 𝑥 𝑦 (𝑛) + 𝑎𝑛−1 𝑥 𝑦 (𝑛−1) + ⋯ + 𝑎0 𝑥 𝑦 = 𝑔(𝑥)

Subject to 𝑦 𝑥0 = 𝑦0, 𝑦" 𝑥0 = 𝑦1 ….. y(n-1) 𝑥0 = 𝑦n-1


given
Existence and Uniqueness of solution of an IVP: If all coefficients
(an(x),..,a0(x)) and g(x) are continuous in x on some interval I, and
an(x)0 on any x in I, then a solution y(x) for the IVP defined for some
x0 in I exists and is unique.
4
3.1 Boundary-value problems

Boundary-value problems (BVPs) consist of


solving a linear DE of order 2 or greater in which
the dependent variable y or its derivatives are
specified at different points:
y ( a ) = y0 and y ( b ) = y1

Or 𝑦 𝑎 = 𝐴, 𝑦′ 𝑏 = 𝐵

And other combinations

5
3.1 Boundary-value problems

The Existence and Uniqueness condition introduced for


IVP is not sufficient for Boundary-value problems (we
may have multiple solutions, or no solutions at all, or
precisely 1 solution, even if the conditions on slide 4
are satisfied).

6
Homogeneous vs Nonhomogeneous
➢ A linear ODE is non-homogenous if it contains a term that
involves only the independent variable (t in the examples below)

𝑥 ′′ + 2𝑥 ′ + 𝑥 = 0 Homogenous

𝑥 ′′ + 2𝑥 ′ + 𝑥 = sin(𝑡) Non-homogenous

𝑥′ + 𝑡2𝑥 = 0 Homogenous

𝑥′ + 𝑡2𝑥 = 𝑡 + 𝑡2 Non-homogenous

➢ That term usually corresponds to a forcing term in the physical model

7
3.1 On solutions of homogeneous linear ODEs
The sum, or superposition, of two or more solutions of
a homogenous linear DE is also a solution.
Therefore,
A constant multiple of a solution is also a solution,
and
The trivial solution y=0 is always a solution.

8
3.1 On solutions of homogeneous linear ODEs

Two functions are linearly dependent if one is a constant


multiple of the other (otherwise, they are linearly independent).
We say that n functions y1(x), y2(x),…, yn(x) are linearly
dependent if there are constants c1, c2,…, cn (not all zero) such
that
c1y1+c2y2+…cnyn=0.
Otherwise, linearly independent.
Any set of n linearly independent solutions y1, y2,…yn of the
homogenous nth order linear DE on interval I is a fundamental
set of solutions.
Relevance: Any solution to that ODE can be expressed as some
linear combination of the fundamental solutions
y1, y2,…yn

9
If all functions in some set  f1 ( x ) , f 2 ( x ) , , f n ( x )
have at least 𝑛 − 1 derivatives, the Wronskian
function is defined as the determinant:

f1 f2 fn
f1 f 2 f n
W ( f1 , , fn ) =
( n −1) ( n −1) ( n −1)
f1 f2 fn

10
Test for linear independence of n solutions of a linear
homogeneous nth order ODE

If 𝑦1, … , 𝑦𝑛 𝑥 are any n solutions of an nth order ODE on some interval I,


then this set is linearly independent on I if and only if the corresponding Wronskian
is not equal to 0 on I :

𝑦1 ⋯ 𝑦𝑛
det ⋮ ⋱ ⋮ 0 for any xI
(𝑛−1) (𝑛−1)
𝑦1 ⋯ 𝑦𝑛

This set of solutions is known as the Fundamental Set of Solutions

11
Homogeneous
➢ If the ODE is Homogeneous, then the General Solution of a
homogeneous linear ODE is
𝑦 = 𝑐1 𝑦1 𝑥 + 𝑐2 𝑦2 𝑥 + ⋯ + 𝑐𝑛 𝑦𝑛 𝑥
Which is a linear combination of all linearly independent solutions.
(superposition principle)

This function also serves as part of the


solution of a non-homogeneous linear ODE,
where it is known as the Complementary Function 𝑦𝑐

12
Non-Homogeneous Linear ODE
➢ If the ODE is Nonhomogeneous, then the General Solution is

𝑦 = 𝑐1 𝑦1 𝑥 + 𝑐2 𝑦2 𝑥 + ⋯ + 𝑐𝑛 𝑦𝑛 𝑥 + 𝑦𝑝

➢ Where 𝑦𝑝 is any particular solution (note: 𝑦𝑝 has no free


parameters).
➢ The General Solution of
𝑎𝑛 𝑥 𝑦 (𝑛) + 𝑎𝑛−1 𝑥 𝑦 (𝑛−1) + ⋯ + 𝑎0 𝑥 𝑦 = 𝑔(𝑥)

Can be written as 𝑦 = 𝑦𝑐 + 𝑦𝑝

Complementary Particular
Function Solution
13
3.1 To summarize ...
General solution of a homogeneous DE is
() () ()
y = c1 y1 x + c2 y2 x + … + cn yn x
Where y1, y2, …yn is a fundamental set of solutions and
ci, I =1,2,…, n are arbitrary constants (free parameters).
General solution of non-homogeneous DE is
() () ()
y = c1 y1 x + c2 y2 x + … + cn yn x + y p
Where y1, y2,…yn is a fundamental set of solutions, yp is
a particular solution, and ci, I = 1,2,…, n are arbitrary
constants

14
Methods of Solution
(for linear ODEs)

15
3.2 Reduction of Order
➢ Assume the following 2nd Order Homogeneous Linear DE

𝑎2 𝑥 𝑦 ′′ + 𝑎1 𝑥 𝑦 ′ + 𝑎0 𝑥 𝑦 = 0

Where y1 and y2 are


➢ General Solution: 𝑦 = 𝑐1 𝑦1 𝑥 + 𝑐2 𝑦2 𝑥 Fundamental solutions
(linearly independent)

➢ Suppose that we want to know 𝑦 but we only have 𝑦1 (obtained


in some other way, e.g., method of undetermined constants (TBD)

How can we obtain 𝑦2 so 𝑦1 and 𝑦2 are linear Independent ?


If we succeed, we have found the fundamental solutions.

16
3.2 Reduction of Order
Reduction of order can be used to reduce a linear
second-order DE with known solution y1 into a linear
first-order DE, which can be solved for a second
solution y2.

Applying reduction of order to the standard form of


a linear homogenous DE
y + P ( x ) y + Q ( x ) y = 0

gives
e 
− P ( x ) dx

y2 = y1 ( x )  2 dx
y1 ( x )
17
3.3 Homogeneous Linear Equations with Constant Coefficients

𝑎𝑛 𝑦 𝑛 + 𝑎𝑛−1 𝑦 𝑛−1 + ⋯ + 𝑎1 𝑦 ′ + 𝑎0 𝑦 = 0

where 𝑎𝑖 , i = 0, 1, ..., n are real constants and 𝑎𝑛 ≠ 0

LCCDEs are solved by first finding solutions of the associated


nth degree algebraic (polynomial) equation
𝑎𝑛 𝑚𝑛 + 𝑎𝑛−1 𝑚𝑛−1 + ⋯ + 𝑎2 𝑚2 + 𝑎1 𝑚 + 𝑎0 = 0

Auxiliary Equation
Find the Roots

18
3.3 Homogeneous Linear Equations with Constant Coefficients

Here is the rationale: consider for example a 2nd order homogeneous LCCDE

Side note:

We already know one specific solution to this


equation:

Trivial Solution is a solution of


any linear homogeneous ODE
(not only CCDE)
3.3 Homogeneous Linear Equations with Constant Coefficients

Assume a solution is of the form:

Why can we guess that an exponential


is a solution? Think of the solution
Then: to a 1st order LCCDE, y’=ay (a=const).
We know (even without solving) that
the solutions take the form
y(t) = C*exp(a*t) for arbitrary constant
C
Then substitute into the ode:
3.3 Homogeneous Linear Equations with Constant Coefficients
3.3 Homogeneous Linear Equations with Constant Coefficients
3.3 Homogeneous Linear Equations with Constant Coefficients
3.3 Homogeneous Linear Equations with Constant Coefficients

So, starting from the LCCDE

And assuming the solution form

We obtained

Factoring out:

One of these two must be 0


3.3 Homogeneous Linear Equations with Constant Coefficients

Factoring out:

Can’t be 0
3.3 Homogeneous Linear Equations with Constant Coefficients

Factoring out:

Must be 0
3.3 Homogeneous Linear Equations with Constant Coefficients

Factoring out:

So

This is called the characteristic (or auxiliary) equation.


3.3 Solving...

Find the characteristic equation (aka auxiliary equation)

Find
(Either factor or use the quadratic equation)

Write the general solution:


3.3 Possible cases
1) Two distinct real values (r1 and r2). The
general solution of ODE is

2) A repeated real value r. The general solution of ODE is 𝑦 = 𝐶1𝑒 𝑟𝑡 + 𝐶2𝑡𝑒 𝑟𝑡

note

2) A complex conjugate 𝑟 = 𝛼 ± 𝑖𝛽, 𝛽 > 0. The general solution of ODE is

𝑦 = 𝐶1𝑒 𝑡 𝑐𝑜𝑠𝑡 + 𝐶2𝑒 𝑡 𝑠𝑖𝑛𝑡


3.3 Specific Solutions to IVPs...
… are obtained from the family of solutions (the general solution) by substituting
given initial conditions:

For first order equations, need 1 initial condition

For second order equations, need 2 initial conditions

For order N equations, need N initial conditions


3.3 Higher order linear ODEs

• Higher Order Equations are more difficult to


solve this way. But if all n solutions of the
auxiliary equation are real and distinct, then
the general solution is
𝑦 = 𝐶1𝑒 𝑟1𝑡 + 𝐶2𝑒 𝑟2𝑡 + ⋯ + 𝐶𝑛𝑒 𝑟𝑛𝑡

• If one root has multiplicity s (s<n) then the


general solution will contain the terms
𝑒 𝑟1𝑡 , 𝑡𝑒 𝑟1𝑡 , … , 𝑡 𝑠−1 𝑒 𝑟𝑛𝑡
3.3 Nonhomogeneous Linear Equations with Constant Coefficients

𝑎𝑛 𝑦 𝑛 + 𝑎𝑛−1 𝑦 𝑛−1 + ⋯ + 𝑎1 𝑦 ′ + 𝑎0 𝑦 = 𝑔(𝑥)

where 𝑎𝑖 , i = 0, 1, ..., n are real constants and 𝑎𝑛 ≠ 0

➢ Recall General Solution:


𝑦 = 𝑦𝑐 + 𝑦𝑝

Complementary Particular
Function Solution

32
Nonhomogeneous Linear Equations with Constant Coefficients

𝑎𝑛 𝑦 𝑛 + 𝑎𝑛−1 𝑦 𝑛−1 + ⋯ + 𝑎1 𝑦 ′ + 𝑎0 𝑦 = 𝑔(𝑥)

where 𝑎𝑖 , i = 0, 1, ..., n are real constants and 𝑎𝑛 ≠ 0

1. The Complementary Function 𝑦𝑐 𝑦 = 𝑦𝑐 + 𝑦𝑝

is the general solution of the associated Homogeneous DE

𝑎𝑛 𝑦 𝑛 + 𝑎𝑛−1 𝑦 𝑛−1 + ⋯ + 𝑎1 𝑦 ′ + 𝑎0 𝑦 = 0

33
Nonhomogeneous Linear Equations with Constant Coefficients

𝑎𝑛 𝑦 𝑛 + 𝑎𝑛−1 𝑦 𝑛−1 + ⋯ + 𝑎1 𝑦 ′ + 𝑎0 𝑦 = 𝑔(𝑥)

where 𝑎𝑖 , i = 0, 1, ..., n are real constants and 𝑎𝑛 ≠ 0

2. Find the Particular Solution 𝑦 = 𝑦𝑐 + 𝑦𝑝


Find Particular Solution

Method of Variation of
Undetermined Coeff. Parameters

34
Nonhomogeneous Linear Equations with Constant Coefficients

𝑎𝑛 𝑦 𝑛 + 𝑎𝑛−1 𝑦 𝑛−1 + ⋯ + 𝑎1 𝑦 ′ + 𝑎0 𝑦 = 𝑔(𝑥)

where 𝑎𝑖 , i = 0, 1, ..., n are real constants and 𝑎𝑛 ≠ 0

2. Find the Particular Solution 𝑦 = 𝑦𝑐 + 𝑦𝑝


Find Particular Solution

Method of
Undetermined Coeff.

35
3.4 Method of Undetermined Coefficients
➢ This method is a conjecture, an educated guess, about
the form of 𝑦𝑝 motivated by the kind of functions that make up
the input function 𝑔(𝑥);

➢ The general method is limited to nonhomogeneous linear DE


where:
o The coefficients 𝑎𝑖 , i = 0, 1, .., n are constant and

o Where 𝑔(𝑥) is either a constant, a polynomial function,


an exponential function 𝑒 𝛼𝑥 , sine or cosine functions
𝑠𝑖𝑛𝛽𝑥 or cos𝛽𝑥, or sums or products of these
functions

36
3.4 Method of Undetermined Coefficients
➢ Some examples:
𝑔(𝑥) = 10 𝑔 𝑥 = 𝑥 2 − 5𝑥 𝑔 𝑥 = 10𝑥 − 6 + 8𝑒 −𝑥

𝑔 𝑥 = sin 3𝑥 − 5𝑥𝑐𝑜𝑠(2𝑥) 𝑔 𝑥 = 𝑥𝑒 𝑥 sin(𝑥) + 3𝑥 2 − 1 𝑒 −4𝑥

➢ That is, 𝑔(𝑥) is a linear combination of functions of type:

𝑃 𝑥 = 𝑎𝑛 𝑥 𝑛 + 𝑎𝑛−1 𝑥 𝑛−1 + ⋯ + 𝑎1 𝑥 + 𝑎0

𝑃 𝑥 𝑒 𝛼𝑥 𝑃 𝑥 𝑒 𝛼𝑥 sin(𝛽𝑥) 𝑃 𝑥 𝑒 𝛼𝑥 cos(𝛽𝑥)
where n is a nonnegative integer and 𝛼 and 𝛽 are real numbers

37
3.4 Method of Undetermined Coefficients

38
3.4 Method of Undetermined Coefficients
➢ This method is not applicable to equations when

𝑔 𝑥 = ln(𝑥) 1
𝑔 𝑥 =
𝑥

𝑔 𝑥 = tan(𝑥) 𝑔 𝑥 = 𝑠𝑖𝑛−1 (𝑥)

➢ For these type of equations we can use the other method


(Variation of Parameters)

39
3.3 Nonhomogeneous Linear Equations with Constant Coefficients

𝑎𝑛 𝑦 𝑛 + 𝑎𝑛−1 𝑦 𝑛−1 + ⋯ + 𝑎1 𝑦 ′ + 𝑎0 𝑦 = 𝑔(𝑥)

where 𝑎𝑖 , i = 0, 1, ..., n are real constants and 𝑎𝑛 ≠ 0

2. Find the Particular Solution 𝑦 = 𝑦𝑐 + 𝑦𝑝


Find Particular Solution

Variation of
Parameters

40
3.5 Variation of Parameters
𝑎2 (𝑥)𝑦 ′′ + 𝑎1 (𝑥)𝑦 ′ + 𝑎0 (𝑥)𝑦 = 𝑔(𝑥)
General Solution
Put the ode into the Standard Form:
𝑦 = 𝑦𝑐 + 𝑦𝑝
𝑦 ′′ + 𝑃 (𝑥)𝑦 ′ + 𝑄 (𝑥)𝑦 = 𝑓(𝑥)

➢ Method:
1. Find Complementary Solution: 𝑦𝑐 = 𝑐1 𝑦1 + 𝑐2 𝑦2

2. Compute Wronskian: 𝑊 𝑦1 𝑥 , 𝑦2 (𝑥)

41
3.5 Variation of Parameters
➢ Method:
3. Find 𝑢1 and 𝑢2 defined as

𝑊1 0 𝑦2
𝑢1′ = 𝑊1 =
𝑓(𝑥) 𝑦2′
𝑊
𝑊2 𝑦1 0
𝑢2′ = 𝑊2 = ′
𝑊 𝑦1 𝑓(𝑥)
𝑦1 𝑦2
𝑊 = 𝑦′ 𝑦2′
1

42
3.5 Variation of Parameters
➢ Method:
4. Particular Solution: 𝑦𝑝 = 𝑢1 𝑦1 + 𝑢2 𝑦2

5. General Solution: 𝑦 = 𝑦𝑐 + 𝑦𝑝

43
3.5 Variation of Parameters
➢ Standard Form of a higher order non-homog. LCCDE

𝑦 𝑛 + 𝑃𝑛−1 (𝑥)𝑦 𝑛−1 + ⋯ + 𝑃1 (𝑥)𝑦 ′ + 𝑃0 (𝑥)𝑦 = 𝑓(𝑥)


➢ Apply Variation of Parameters Method
0 𝑦2 𝑦3
𝑊𝑘 𝑊1 = 0 𝑦2′ 𝑦3′
𝑢𝑘′ = k = 1, 2, … n
𝑓(𝑥) 𝑦2′′ 𝑦3′′
𝑊
𝑊1 𝑊2 𝑊3 𝑦1 0 𝑦3
𝑢1′ = 𝑢2′ = 𝑢3′ = 𝑊2 = 𝑦1′ 0 𝑦3′
𝑊 𝑊 𝑊
𝑦1′′ 𝑓(𝑥) 𝑦3′′
𝑦1 𝑦2 𝑦3
′ 𝑦1 𝑦2 0
𝑊 = 𝑦1 𝑦2′ 𝑦3′
𝑦1′′ 𝑦2′′ 𝑦3′′ 𝑊3 = 𝑦1′ 𝑦2′ 0
𝑦1′′ 𝑦2′′ 𝑓(𝑥)

44
Overview
Theory of Linear Equations Cauchy-Euler
Concepts Equations

➢ IVP and BVP ➢ Linear Eq. with variable coeff.


➢ Differential Operator ➢ Homogeneous Solutions
➢ Superposition Principle ➢ Nonhomogeneous Solutions
➢ Linear Dependence and Independence (Wronskian)
➢ Homogeneous and Nonhomogeneous Examples of
Linear Models
Reduction Order
➢ Spring/Mass Systems
Homogeneous Solutions o Undamped Motion
o Free Damped Motion
➢ Linear Eq. with constant coeff. o Driven Motion
➢ nth degree polynomial equation
Green’s Functions
Nonhomogeneous Solutions
Nonlinear Functions
➢ Method of Undetermined Coefficients
➢ Variation of Parameters ➢ Linearization

45
3.5 Cauchy-Euler Equations

𝑎𝑛 𝑥 𝑛 𝑦 𝑛 + 𝑎𝑛−1 𝑥 𝑛−1 𝑦 𝑛−1 + ⋯ + 𝑎1 𝑥𝑦 ′ + 𝑎0 𝑦 = 𝑔(𝑥)

where 𝑎𝑛 , 𝑎𝑛−1 , ... 𝑎0 are constants


Cauchy-Euler ODE
➢ General Solution (Similar Approach)
𝑦 = 𝑦𝑐 + 𝑦𝑝

Complementary Particular
Function Solution

46
3.5 Cauchy-Euler Equations
𝑦 = 𝑦𝑐 + 𝑦𝑝
(Homogeneous Eq.)

𝑎𝑛 𝑥 𝑛 𝑦 𝑛 + 𝑎𝑛−1 𝑥 𝑛−1 𝑦 𝑛−1 + ⋯ + 𝑎1 𝑥𝑦 ′ + 𝑎0 𝑦 = 0

➢ Solution of the form: 𝑦 = 𝑥𝑚 Instead of nth degree


polynomial equation
➢ Find the Auxiliary Equation
𝑎𝑚(𝑚 − 1) + 𝑏𝑚 + 𝑐 = 0 (2nd Oder Example)

𝑎𝑚2 + (𝑏 − 𝑎)𝑚 + 𝑐 = 0

Find the Roots

47
3.5 Cauchy-Euler Equations
𝑎𝑚2 + (𝑏 − 𝑎)𝑚 + 𝑐 = 0 (2nd Oder Example)

➢ Case I: Distinct Real Roots


𝑦 = 𝑐1 𝑥 𝑚1 + 𝑐2 𝑥 𝑚2

➢ Case II: Repeated Real Roots


𝑦 = 𝑐1 𝑥 𝑚1 + 𝑐2 𝑥 𝑚1 ln(𝑥)

➢ Case III: Conjugate Complex Roots

𝑦 = 𝑥 𝛼 𝑐1 cos 𝛽 ln 𝑥 + 𝑐2 sin(𝛽ln(𝑥))

48
3.5 Cauchy-Euler Equations
𝑦 = 𝑦𝑐 + 𝑦𝑝

➢ Apply Variation of Parameters Method


Standard Form
𝑦 ′′ + 𝑃 (𝑥)𝑦 ′ + 𝑄 (𝑥)𝑦 = 𝑓(𝑥)

1. Find Complementary Solution: 𝑦𝑐 = 𝑐1 𝑦1 + 𝑐2 𝑦2


2. Compute the Wronskian: 𝑊 𝑦1 𝑥 , 𝑦2 (𝑥)
3. Find 𝑢1 and 𝑢2
4. Particular Solution: 𝑦𝑝 = 𝑢1 𝑦1 + 𝑢2 𝑦2
5. General Solution: 𝑦 = 𝑦𝑐 + 𝑦𝑝

49
50
Nonlinear ODEs
➢ More difficult to solve than linear ODEs.
➢ Do not share the property of linear homogeneous DEs that a linear
combination of solutions is also a solution.
Some simple NL cases can be handled by methods such as solution
by substitution, exact differential, or by recasting ODE as a
Bernoulli equation.

51
NL ODEs: Riccati equation

52
NL ODEs: Taylor series approximation
Example: 𝑦 ′′ = 𝑥 + 𝑦 2 , 𝑦 0 = 1, 𝑦 ′ 0 = 1.
Assuming that a solution exists and is analytic at x=0:
1 ′′
𝑦 𝑥 = 𝑦 0 + 𝑦 𝑥 + 𝑦 0 𝑥2 + ⋯
′ 0
2
Substitute 𝑦 0 = 1, 𝑦 ′ 0 = 1:
𝑦 ′′ 0 = 0 + 𝑦 2 0 = 1.
Now compute higher derivatives:
′′′
𝑑
𝑦 0 = 𝑥 + 𝑦 2 ቤ 𝑥 = 0, 𝑦 = 1 = 1 + 2𝑦𝑦 ′ ֜ 𝑦 ′′′ 0 = 3
𝑑𝑥
(4)
𝑑
𝑦 0 = 1 + 2𝑦𝑦 ′ ቤ 𝑥 = 0, 𝑦, 𝑦′ = 1 = 2 𝑦 ′ 2 + 2𝑦𝑦 ′′ = 4
𝑑𝑥

Continue with as many terms as desired/needed for better
accuracy.
53
NL ODEs: Reduction of order
Example: 𝑦 ′′ = 1 + 𝑦′2 .
Substitution: let 𝑢 = 𝑦′. The ODE becomes 𝑢′ = 1 + 𝑢2
Which is a separable ODE.
𝑑𝑢
2
= 𝑑𝑥 ֜ 𝑢 = tan(𝑥 + 𝐶1)
𝑢 +1

𝑢𝑦 = − ln cos(𝑥 + 𝐶1) + C2

This is useful in the cases when
1) Variable y is missing, that is, ODE is of the form F(x,y’,y”)=0
2) Variable x is missing, that is, ODE is of the form F(y,y’,y”)=0

The substitution above reduces the ODE to a 1st order NL ODE.

54
NL ODEs: Reduction of order: Example 2
Example: Curvature of a function 𝑦 = 𝑓(𝑥) is defined as
k=
𝑦′′
′ 2 3ൗ
1+ 𝑦 2
Find the function f(x) whose curvature is equal to 1.
let 𝑢 = 𝑦′.
2 3ൗ
1 + 𝑢 2 = 𝑢′
Which is a separable ODE.
𝑑𝑢 2 3Τ
= 1 + 𝑢 2 ֜ separable form; Use 𝑢 = 𝑡𝑎𝑛.
𝑑𝑥
sec 2  sec 2 
 1 + tan 2  3/2 d = x   sec3  d = x   cos d = x
( )
u y x
 sin  = x  =x = x  y = ,x  0
1+ u 2
1 + y 2
1− x 2

 y( x) = − 1 − x 2 55

You might also like