NUMERICAL SOLUTIONS OF ORDINARY

DIFFERENTIAL EQUATIONS:
SINGLE-STEP METHODS

Dr. Gabriel Obed Fosu

Department of Mathematics
Kwame Nkrumah University of Science and Technology
Google Scholar: https://scholar.google.com/citations?user=ZJfCMyQAAAAJ&hl=en&oi=ao
ResearchGate ID: https://www.researchgate.net/profile/Gabriel_Fosu2

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 1 / 49

Lecture Outline

1 Introduction to ODE

2 Single-Step Methods and Multi-Step Schemes

3 Solution To IVP’s Using Single-Step Methods

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 2 / 49

 Introduction to ODE

Introduction
1 The general form of an nt h order ODE is

f (x, y ′ , y ′′ , y ′′′ , · · · , y (n) ) = 0 (1)

2 The order of an ODE is the order of its highest derivative.

3 A differential equation together with some initial conditions is called an Initial
Value Problem(IVP).
4 A first order IVP can be written as:

y ′ = f (x, y); y(x 0 ) = y 0 (2)

5 The methods for solving this IVP’s can be classified into single-step methods
and multi-step methods.

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 4 / 49

 Introduction to ODE

Recall

1 Given an interval [x 0 , b] in which a solution is desired.

2 The interval is divided into finite number of sub-intervals by points

x0 < x1 < x2 < · · · < xn ; xn = b (3)

3 These points are called mesh points or grid points.
4 The spacing between the points are given by

h i = x i − x i −1 , i = 1, 2, 3, · · · , n (4)

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 5 / 49

 Single-Step Methods and Multi-Step Schemes

Single-Step Scheme

Definition
For single step method’s the solution at any point is obtained by using the solution
at the previous point.

Single-step methods can be classified as

Implicit single-step schemes
Explicit single-step schemes

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 7 / 49

 Single-Step Methods and Multi-Step Schemes

Implicit Single-Step Schemes

1 With this method, the solution at any point y i +1 is obtained by using the
solution at only the previous point y i and at the point itself.
A general single step implicit method can be written as
y i +1 = y i + h f (x i +1 , x i , y i +1 , y i , h) (5)
2 The function f is called the increment function.

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 8 / 49

 Single-Step Methods and Multi-Step Schemes

Explicit Single-Step Scheme

1 In the explicit case, the right-hand side does not depend on y i +1 . For the
explicit scheme eq. (5) reduces to:
y i +1 = y i + h f (x i , y i , h) (6)

2 Given an initial value y 0 , then the other values of y are computed successively
as
y 1 = y 0 + h f (x 0 , y 0 , h), when i = 0 (7)
y 2 = y 1 + h f (x 1 , y 1 , h), when i = 1 (8)
y 3 = y 2 + h f (x 2 , y 2 , h), when i = 2 (9)
.. .. ..
. . .
y n = y n−1 + h f (x n−1 , y n−1 , h), when i = n − 1 (10)
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 9 / 49
 Single-Step Methods and Multi-Step Schemes

Single-Step Scheme

1 The solution of y 1 requires only one previous point y 0 .

2 The solution of y 2 requires only one previous point y 1 .
3 The solution of y 3 requires only one previous point y 2 .
4 The solution of y n requires only one previous point y n−1 .

Note
All single-step methods are self starting, that is, they do not require values of y or
it’s derivatives beyond the immediate previous point.

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 10 / 49

 Single-Step Methods and Multi-Step Schemes

Multi-Step Methods

With this method, the solution at point y i +1 is obtained using the solution at a
number of previous points, y i , y i −1 , y i −2 , y i −3 , · · · .
1 Two-step implicit depends on y i +1 , y i , y i −1
2 Two-step explicit depends on y i , y i −1
3 Four-step implicit depends on y i +1 , y i , y i −1 , y i −2 , y i −3
4 Four-step explicit depends on y i , y i −1 , y i −2 , y i −3

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 11 / 49

 Single-Step Methods and Multi-Step Schemes

Multi-Step Schemes
1 A classical example of a two-step implicit method can be written as
y i +1 = y i + h f (x i +1 , x i , x i −1 , y i +1 , y i , y i −1 , h) (11)
2 A classical example of a three-step explicit method can be written as
y i +1 = y i + h f ( x i , x i −1 , x i −2 , y i , y i −1 , y i −2 , h) (12)
3 A general k -step explicit method can be written as
y i +1 = y i + h f (x i −k+1 , · · · , x i −1 , x i , y i −k+1 , · · · , y i −1 , y i , h) (13)
4 A general k -step implicit method can be written as
y i +1 = y i + h f (x i −k+1 , · · · , x i −1 , x i , x i +1 , y i −k+1 , · · · , y i −1 , y i , y i +1 , h) (14)

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 12 / 49

 Single-Step Methods and Multi-Step Schemes

Note
The following are equivalent

y i +1 = y(x i +1 ), y i = y(x i ), ···

Note
The following are equivalent

f i = f [x i , y(x i )], f i +1 = f [x i +1 , y(x i +1 )], ···

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 13 / 49

 Solution To IVP’s Using Single-Step Methods

Solution To IVP’s Using Single-Step Methods

Some numerical techniques used for solving IVP’s include:
1 Euler or Taylor series Method
2 Backward Euler
3 Modified Euler or Midpoint Method
4 Trapezium Method
5 Heun’s Method or Euler-Cauchy Method
6 Runge-Kutta (RK) Methods

All these methods are derived using Taylor series. Given the Taylor
series:
y(x i +1 ) = y(x i ) + h f [(x i + θh), y(x i + θh)]; 0≤θ≤1 (15)
These schemes can be deduced from the above depending on the value of θ .
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 15 / 49
 Solution To IVP’s Using Single-Step Methods 1. Taylor Series of Order 1 or Euler Method

Taylor Series of Order 1 or Euler Method

1 Given the general Taylor series:

y(x i +1 ) = y(x i ) + h f [(x i + θh), y(x i + θh)]; 0≤θ≤1 (16)

2 The Euler Method is obtained from eq. (16) be letting

θ=0

3 The scheme is given by the formula

y(x i +1 ) = y(x i ) + h f [x i , y(x i )] (17)

4 This is an explicit scheme.

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 16 / 49

 Solution To IVP’s Using Single-Step Methods Backward Euler

2. Backward Euler
1 Given the general Taylor series:

y(x i +1 ) = y(x i ) + h f [(x i + θh), y(x i + θh)]; 0≤θ≤1 (18)

2 The Backward Euler is obtained from eq. (18) be letting

θ=1

.
3 The scheme is given by the formula

y(x i +1 ) = y(x i ) + h f [(x i + h), y(x i + h)] (19)

= y(x i ) + h f [x i +1 , y(x i +1 )] (20)
4 This is an implicit scheme.
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 17 / 49
 Solution To IVP’s Using Single-Step Methods Modified Euler or Midpoint Method

3. Modified Euler or Midpoint Method

1 Given the general Taylor series:
y(x i +1 ) = y(x i ) + h f [(x i + θh), y(x i + θh)]; 0≤θ≤1 (21)
2 The Modified Euler method is obtained from eq. (21) be letting
1
θ=
2
3 The scheme is given by the formula
h h
·µ ¶ µ ¶¸
y(x i +1 ) = y(x i ) + h f xi + , y xi + (22)
2 2
³ ´
4 However, x i + h2 is not a nodal point, hence we approximate y x i + h
2 using the
h
Euler method with spacing 2
.
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 18 / 49
 Solution To IVP’s Using Single-Step Methods Modified Euler or Midpoint Method

Modified Euler or Midpoint Method

1 The Euler approximation is given by eq. (23).

h h
µ ¶
y xi + = y(x i ) + f (x i , y i ) (23)
2 2

2 Substituting eq. (23) into eq. (22), we obtain the Modified Euler as
h h
·µ ¶ ¸
y(x i +1 ) = y(x i ) + h f xi + , y(x i ) + f (x i , y i ) (24)
2 2

3 This is an explicit scheme

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 19 / 49

 Solution To IVP’s Using Single-Step Methods Trapezium Method

4. Trapezium Method

1 If the continuously varied slope in x i and x i +1 is approximated by the mean of

the slope, then the trapezium method is deduced as
h© ª
y(x i +1 ) = y(x i ) + f [x i , y(x i )] + f [x i +1 , y(x i +1 )] (25)
2
h
= yi + [ f i + f i +1 ] (26)
2
2 This is an implicit scheme.
3 When this is converted to an explicit scheme we obtain the Heun’s method.

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 20 / 49

 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

5. Heun’s Method or Euler-Cauchy Method

1 This is the explicit form of the trapezium method. This conversion is made
possible by using the approximation

y(x i +1 ) = y(x i ) + h f [x i , y(x i )] (27)

2 Substituting eq. (27) into eq. (25) we obtain the Euler-Cauchy iterative scheme
as:
h© ª
y(x i +1 ) = y(x i ) + f [x i , y(x i )] + f [x i +1 , y(x i +1 )] (28)
2
h© ª
= y(x i ) + f [x i , y(x i )] + f [x i +1 , y(x i ) + h f [x i , y(x i )]] (29)
2

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 21 / 49

 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Example
Solve the following IVP

y y ′ = x, y(0) = 1, 0 ≤ x ≤ 0.6, h = 0.2

using
1 Euler method
2 Modified Euler method
3 Euler-Cauchy method
In each case compute the absolute error at x = 0.6

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 22 / 49

 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Euler Method
Given the step size h = 0.2. Then the x values are given by the interval table

x 0 = 0, x 1 = 0.2, x 2 = 0.4, x 3 = 0.6 (30)

The Euler formula is given as

y(x i +1 ) = y(x i ) + h f [x i , y(x i )] (31)
y i +1 = y i + h f (x i , y i ) (32)

From eq. (2), y ′ = f (x, y). Therefore making y ′ the subject from the question we
have
x
y ′ = f (x, y) = (33)
y
Iteratively, xi
f (x i , y i ) = (34)
yi
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 23 / 49
 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Euler Method: Iteration 1: when i=0

The formula reduces to
y(x 1 ) = y(x 0 ) + h f [x 0 , y(x 0 )] (35)
y 1 = y 0 + h f (x 0 , y 0 ) (36)
From the question, that is y(0) = 1, the initial conditions can be deduced as
x 0 = 0, and y 0 = 1
Therefore x0 0
f (x 0 , y 0 ) = = =0
y0 1
Hence
y 1 = y 0 + h f (x 0 , y 0 ) (37)
= 1 + 0.2(0) (38)
=1 (39)
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 24 / 49
 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Euler Method: Iteration 2: when i=1

The formula reduces to
y 2 = y 1 + h f (x 1 , y 1 )
From the interval table (30), x 1 = 0.2, and from the previous solution y 1 = 1
Therefore x 1 0.2
f (x 1 , y 1 ) = = = 0.2
y1 1
Hence
y 2 = y 1 + h f (x 1 , y 1 ) (40)
= 1 + 0.2(0.2) (41)
= 1.04 (42)

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 25 / 49

 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Euler Method: Iteration 3: when i=2

The formula reduces to
y 3 = y 2 + h f (x 2 , y 2 )
From the interval table (30), x 2 = 0.4, and from the previous solution y 2 = 1.04
Therefore
y 3 = y 2 + h f (x 2 , y 2 ) (43)
µ ¶
0.4
= 1.04 + 0.2 (44)
1.04
= 1.117 (45)

Therefore, the nodal points are:

(x 0 , y 0 ) = (0, 1); (x 1 , y 1 ) = (0.2, 1); (x 2 , y 2 ) = (0.4, 1.04); (x 3 , y 3 ) = (0.6, 1.117)
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 26 / 49
 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Euler vs Analytical Solution

The differential equation is solved using separation of variables
dy x
Z Z
y = ′
= =⇒ dyy = xd x =⇒ y 2 = x 2 + c
dx y
Implementing the initial condition to find c
12 = 02 + c =⇒ c = 1
Therefore the analytical solution is
p
y= x2 + 1
Hence at point x 3 = 0.6
p
y 3 = y(0.6) = 0.62 + 1 = 1.166
The absolute error
AE = |E S − AS| = |1.166 − 1.117| = 0.049
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 27 / 49
 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Modified Euler
The formula is given as
h h
¶
·µ ¸
y(x i +1 ) = y(x i ) + h f x i + , y i + f (x i , y i ) (46)
2 2
h h
· ¸
y i +1 = y i + h f x i + , y i + f (x i , y i ) (47)
2 2

Again xi
f (x i , y i ) = (48)
yi
The initial conditions can be deduced as

x 0 = 0, and y 0 = 1

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 28 / 49

 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Modified Euler Iteration 1: when i=0

The formula reduces to
h h
· ¸
y 1 = y 0 + h f x 0 + , y 0 + f (x 0 , y 0 ) (49)
2 2
· µ ¶¸
0.2 0.2 0
= 1 + 0.2 f 0 + , 1+ (50)
2 2 1
= 1 + 0.2 f (0.1, 1) (51)
µ ¶
0.1
= 1 + 0.2 (52)
1
= 1.02 (53)

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 29 / 49

 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Modified Euler Iteration 2: when i=1

From the interval table (30), x 1 = 0.2, and from the previous solution y 1 = 1.02
The solution is as follows
h h
· ¸
y 2 = y 1 + h f x 1 + , y 1 + f (x 1 , y 1 ) (54)
2 2
· µ ¶¸
0.2 0.2 0.2
= 1.02 + 0.2 f 0.2 + , 1.02 + (55)
2 2 1.02
= 1.02 + 0.2 f (0.3, 1.04) (56)
µ ¶
0.3
= 1.02 + 0.2 (57)
1.04
= 1.0777 (58)

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 30 / 49

 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Modified Euler Iteration 3: when i=2

From the interval table (30), x 2 = 0.4, and from the previous solution y 2 = 1.0777
The solution is as follows
h h
· ¸
y 3 = y 2 + h f x 2 + , y 2 + f (x 2 , y 2 ) (59)
2 2
· µ ¶¸
0.2 0.2 0.4
= 1.0777 + 0.2 f 0.4 + , 1.03 + (60)
2 2 1.0777
= 1.0777 + 0.2 f (0.5, 1.1148) (61)
µ ¶
0.5
= 1.0777 + 0.2 (62)
1.1148
= 1.1674 (63)

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 31 / 49

 Solution To IVP’s Using Single-Step Methods Heun’s Method or Euler-Cauchy Method

Modified Euler vs Analytical Solution

Therefore the nodal points are:

(x 0 , y 0 ) = (0, 1); (x 1 , y 1 ) = (0.2, 1.02); (x 2 , y 2 ) = (0.4, 1.0777); (x 3 , y 3 ) = (0.6, 1.1148)

Hence
AE = |E S − AS| = |1.167 − 1.1148| = 0.0522

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 32 / 49

 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

6. Runge-Kutta Methods

1 This is also a single-step method used for solving IVPs.

2 A Runge-Kutta method of second-order uses two slopes, that is

k 1 and k 2 (64)
3 whereas the fourth-order Runge-Kutta uses four slopes;

k 1 , k 2 , k 3 , and k 4 (65)

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 33 / 49

 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Second-Order Runge-Kutta

1 A general second-order Runge-Kutta (RK2) is of the form

k2
µ¶
1
y i +1 = y i + 1 − k1 + (66)
2θ 2θ

where
k 1 = h f (x i , y i ) and
k 2 = h f (x i + θh, y i + θk 1 )
2 The value of θ is arbitrary such that 0 ≤ θ ≤ 1.
3 This lead to a myriad number of solution schemes

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 34 / 49

 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Second-Order Runge-Kutta
1 When θ = 1
1 1
y i +1 = y i + k 1 + k 2 (67)
2 2
where k1 = h f (x i , y i ) and
k 2 = h f (x i + h, y i + k 1 )
This is the same as the Heun’s method.
2 When θ = 1/2

y i +1 = y i + k 2 (68)
where k1 = h f (x i , y i ) and
k 2 = h f (x i + h/2, y i + k 1 /2)
This is the same as the Modified Euler
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 35 / 49
 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Fourth-Order Runge-Kutta
In the case of RK4, the iterative scheme is given by
1
y i +1 = y i + (k 1 + 2k 2 + 2k 3 + k 4 ) (69)
6
where
1 k 1 = h f (x i , y i ),
h k1
µ ¶
2 k2 = h f xi + , y i +
2 2
h k2
µ ¶
3 k3 = h f xi + , y i +
2 2
¡ ¢
4 k 4 = h f x i + h, y i + k 3

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 36 / 49

 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Example
Solve the following IVP

y y ′ = x, y(0) = 1, 0 ≤ x ≤ 0.6, h = 0.2

using Runge-Kutta fourth-order scheme.

Hence determine the absolute error at x = 0.6

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 37 / 49

 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Solution
Given the step size h = 0.2. Then the x values are given by the interval table

x 0 = 0, x 1 = 0.2, x 2 = 0.4, x 3 = 0.6 (70)

We know that
x
y ′ = f (x, y) = (71)
y
Iteratively, xi
f (x i , y i ) = (72)
yi
The initial conditions can be deduced as

x 0 = 0, and y 0 = 1

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 38 / 49

 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Iteration 1: when i=0

The formula reduces to
1
y 1 = y 0 + (k 1 + 2k 2 + 2k 3 + k 4 ) (73)
6
where
µ ¶
0
k 1 = h f (x 0 , y 0 ) = 0.2 =0 (74)
1
h k1
µ ¶ µ ¶
0.2 0
k2 = h f x0 + , y 0 + = 0.2 f 0 + , 1+
2 2 2 2
= 0.2 f (0.1, 1) (75)
µ ¶
0.1
= 0.2 = 0.02
1

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 39 / 49

 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Iteration 1: when i=0

h k2
µ ¶ µ ¶
0.2 0.02
k3 = h f x0 + , y 0 + = 0.2 f 0 + , 1+
2 2 2 2
= 0.2 f (0.1, 1.01)
µ ¶ (76)
0.1
= 0.2
1.01
= 0.02
¡ ¢
k 4 = h f x 0 + h, y 0 + k 3 = 0.2 f (0 + 0.2, 1 + 0.02)
= 0.2 f (0.2, 1.02)
(77)
µ ¶
0.2
= 0.2
1.02
= 0.04
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 40 / 49
 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Iteration 1: when i=0

Now let substitute eqs. (74) to (77) into the main formula eq. (73). Hence
1
y 1 = y 0 + (k 1 + 2k 2 + 2k 3 + k 4 ) (78)
6
1
= 1 + [0 + 2(0.02) + 2(0.02) + 0.04] (79)
6
= 1.02 (80)

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 41 / 49

 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Iteration 2: when i=1

From the interval table (70), x 1 = 0.2, and from the previous solution y 1 = 1.02
The formula reduces to
1
y 2 = y 1 + (k 1 + 2k 2 + 2k 3 + k 4 ) (81)
6
where
¶ µ
0.2
k 1 = h f (x 1 , y 1 ) = 0.2 = 0.04 (82)
1.02
h k1
µ ¶ µ ¶
0.2 0.04
k2 = h f x1 + , y 1 + = 0.2 f 0.2 + , 1.02 +
2 2 2 2
= 0.2 f (0.3, 1.04) (83)
µ ¶
0.3
= 0.2 = 0.06
1.04
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 42 / 49
 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Iteration 2: when i=1

h k2
µ ¶ µ ¶
0.2 0.06
k3 = h f x1 + , y 1 + = 0.2 f 0.2 + , 1.02 +
2 2 2 2
= 0.2 f (0.3, 1.05)
µ ¶ (84)
0.3
= 0.2
1.05
= 0.06
¡ ¢
k 4 = h f x 1 + h, y 1 + k 3 = 0.2 f (0.2 + 0.2, 1.02 + 0.06)
= 0.2 f (0.4, 1.08)
(85)
µ ¶
0.4
= 0.2
1.08
= 0.079
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 43 / 49
 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Iteration 2: when i=1

Now let substitute eqs. (82) to (85) into the main formula eq. (81). Hence
1
y 2 = y 1 + (k 1 + 2k 2 + 2k 3 + k 4 ) (86)
6
1
= 1.02 + [0.04 + 2(0.06) + 2(0.06) + 0.079] (87)
6
= 1.079 (88)

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 44 / 49

 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Iteration 3: when i=2

From the interval table (70), x 2 = 0.4, and from the previous solution y 2 = 1.079
The formula reduces to
1
y 3 = y 2 + (k 1 + 2k 2 + 2k 3 + k 4 ) (89)
6
where
µ¶
0.4
k 1 = h f (x 2 , y 2 ) = 0.2 = 0.074 (90)
1.079

h k1
µ ¶ µ ¶
0.2 0.074
k2 = h f x2 + , y 2 + = 0.2 f 0.4 + , 1.079 +
2 2 2 2
= 0.2 f (0.5, 1.116) (91)
µ ¶
0.5
= 0.2 = 0.09
1.116
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 45 / 49
 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Iteration 3: when i=2

h k2
µ ¶ µ ¶
0.2 0.09
k3 = h f x2 + , y 2 + = 0.2 f 0.4 + , 1.079 +
2 2 2 2
= 0.2 f (0.5, 1.125)
µ ¶ (92)
0.5
= 0.2
1.125
= 0.089
¡ ¢
k 4 = h f x 2 + h, y 2 + k 3 = 0.2 f (0.4 + 0.2, 1.079 + 0.089)
= 0.2 f (0.6, 1.168)
(93)
µ ¶
0.6
= 0.2
1.168
= 0.103
Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 46 / 49
 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Iteration 3: when i=2

Now let substitute eqs. (90) to (93) into the main formula eq. (89). Hence
1
y 3 = y 2 + (k 1 + 2k 2 + 2k 3 + k 4 ) (94)
6
1
= 1.079 + [0.07 + 2(0.09) + 2(0.089) + 0.103] = 1.168 (95)
6

Therefore the nodal points are;

(x 0 , y 0 ) = (0, 1); (x 1 , y 1 ) = (0.2, 1.02); (x 2 , y 2 ) = (0.4, 1.079); (x 3 , y 3 ) = (0.6, 1.168)

Hence AE = |E S − AS| = |1.167 − 1.168| = 0.001 (96)

Comparing the absolute error, we may conclude that, the fourth-order Runge-Kutta
scheme gives a more approximate solution than the other single-step schemes.

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 47 / 49

 Solution To IVP’s Using Single-Step Methods Runge-Kutta Methods

Exercise

Solve the initial value problem

y ′ = 2x + 3y, y(0) = 1, x ∈ [0, 0.4], h = 0.1

using
1 Euler method, hence determine the relative error at x = 0.1
2 Modified Euler method, hence determine the relative error at x = 0.2
3 Euler-Cauchy method, hence determine the relative error at x = 0.3
4 Fourth-Order Runge-Kutta, hence determine the relative error at x = 0.4

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 48 / 49

 END OF LECTURE
THANK YOU

Dr. Gabby (KNUST-Maths) ODE: Single Step Methods 49 / 49