Sunteți pe pagina 1din 14

Numerical Methods Problems

Epiphany Term
Numerical Dierentiation
Question 1
Use the forward, central and backward dierence formulae to complete the last row
of the table
x 0.5 0.6 0.7
f(x) 0.493 0.712 0.933
f

(x)
Solution:
We use the forward dierence formula to get the rst entry:
f

(0.5)
f(x + h) f(x)
h
=
0.712 0.493
0.1
= 2.190,
the central dierence formula to get the middle entry:
f

(0.6)
f(x + h) f(x h)
2h
=
0.933 0.493
0.2
= 2.200
and the backward dierence formula to get the last entry:
f

(0.7)
f(x) f(x h)
h
=
0.933 0.712
0.1
= 2.210.
Question 2
Recall the error term for the 2
nd
order central dierence formula is given by
1
6
f

(c)h
2
for some x h < c < x +h. For the function f(x) = x
3
+ 2x, how small a stepsize h
would this require to compute f

(0) to within an accuracy of 0.001 ?


Solution:
We use the error formula at x = 0. Since f

(c) = 6, we have |f

(c)| 6 for
h < c < h. Hence, then absolute error satises

1
6
f

(c)h
2

1
6
(6)h
2
= h
2
.
Thus we need h
2
0.001 and so h 0.03 will suce. (Note that we should round
down from

0.001 = 0.0316... here.)


1
Question 3
Apply the 2
nd
order central dierence formula with stepsizes h = 0.1, h = 0.2 and
h = 0.4 to estimate the derivative of x
4
at x = 1 and compare with the actual value.
A 4-point dierence formula for the derivative is given by
f

(x)
2f(x h) 3f(x) + 6f(x + h) f(x + 2h)
6h
Repeat your estimates with this formula and comment on the order of the error.
Solution:
Using the central dierence formula,
f

(1)
f(x + h) f(x h)
2h
=
_

_
1.1
4
0.9
4
2 0.1
= 4.04 when h = 0.1
1.2
4
0.8
4
2 0.2
= 4.16 when h = 0.2
1.4
4
0.6
4
2 0.4
= 4.64 when h = 0.4
Using the 4-point formula,
f

(1)
_

_
2 0.9
4
3 1
4
+ 6 1.1
4
1.2
4
6 0.1
= 3.998 when h = 0.1
2 0.8
4
3 1
4
+ 6 1.2
4
1.4
4
6 0.2
= 3.984 when h = 0.2
2 0.6
4
3 1
4
+ 6 1.4
4
1.8
4
6 0.4
= 3.872 when h = 0.4
The actual value of the derivative is f

(1) = 4.
For the central dierence formula, the error is 4h
2
each time conrming that it is
a 2
nd
order method.
For the 4-point formula the error is 2h
3
each time which suggests that it is a 3
rd
order method.
Question 4
Using the central dierence formula for the second derivative
f

(x)
f(x + h) 2f(x) + f(x h)
h
2
with a stepsize h = 0.1, estimate f

(2) when f(x) = xe


x
. Repeat with h = 0.2 and
h = 0.4 and comment on the order of the error.
Page 2
Solution:
Using the central dierence formula for f

(x),
f

(2)
_

_
2.1e
2.1
2 2e
2
+ 1.9e
1.9
0.1
2
= 29.593 when h = 0.1
2.2e
2.2
2 2e
2
+ 1.8e
1.8
0.2
2
= 29.704 when h = 0.2
2.4e
2.4
2 2e
2
+ 1.6e
1.6
0.4
4
= 30.152 when h = 0.4
The actual 2
nd
derivative is f

(x) = (x + 2)e
x
so f

(2) = 4e
2
= 29.556.
Hence the errors are -0.037, -0.148 and -0.595 respectively.
These are roughly 3.7h
2
in each case which suggests the method is 2
nd
order.
Numerical Integration
Question 5
Given data
x 1.0 1.2 1.4 1.6 1.8
f(x) 3.120 4.426 6.042 8.030 10.467
use the Left, Right and Midpoint rectangular rules to approximate
_
1.8
1
f(x) dx.
Solution:
For the Left and Right Endpoint Rules, we can take n = 4 subintervals of width
h = 0.2.
The Left Endpoint Rule gives:
_
1.8
1
f(x) dx h
n

i=1
f(x
i1
)
= 0.2(3.120 + 4.426 + 6.042 + 8.030) = 4.3236
The Right Endpoint Rule gives:
_
1.8
1
f(x) dx h
n

i=1
f(x
i
)
= 0.2(4.426 + 6.042 + 8.030 + 10.467) = 5.7930
To use the Midpoint Rule, since we need values of f(x) at midpoints of each subin-
terval, well need to take only n = 2 intervals of width h = 0.4. Then we get:
_
1.8
1
f(x) dx h
n

i=1
f(x
i
h/2)
= 0.4(4.426 + 8.030) = 4.9824
Page 3
Question 6
The arclength of a curve y = g(x) over a x b is given by the integral
_
b
a
_
1 + g

(x)
2
dx
Use the Midpoint rule with 3 panels on the integral when g(x) = sin(x), a = 0 and
b = /2 and hence estimate the length of a complete sine wave.
Given that the 2nd derivative of f(x) =
_
1 + g

(x)
2
satises
1
2

2 f

(x) 1,
nd a bound for the error in your estimate.
Solution:
Let f(x) =
_
1 + g

(x)
2
=

1 + cos
2
x.
Applying the Midpoint Rule with n = 3 intervals, h = /6 and x
i
= ih, we get
_
/2
0
f(x) dx h
n

i=1
f(x
i
h/2)
=

6
_
f
_

12
_
+ f
_
3
12
_
+ f
_
5
12
__
= 0.5236 [1.3903 + 1.2247 + 1.0330]
= 1.9102
From x = 0 to /2 is a quarter of a wave, hence we estimate 4 1.9102 = 7.6404
for the length of a full wave.
From the given information, we know |f

(c)| K = 1 for 0 < c < /2 hence the


Midpoint Rule error satises
|E
M
|
K(b a)h
2
24
=

3
288
= 0.1077
Hence after multiplying by 4, our estimate is bounded above by 0.4306.
(The actual error is less than 0.00001 ! )
Question 7
Use the Midpoint Rule to estimate
_
1
0
x
4
2x
3
dx to within 0.02 of the correct value.
Justify the number of panels you have used.
Solution:
The second derivative f

(x) = 12x
2
6 is increasing over the interval 0 x 1 so
takes its largest absolute values at the endpoints.
These are f

(0) = 6 and f

(1) = 6 so |f

(x)| K = 6 for 0 x 1.
Hence the error term satises |E
M
|
K(b a)
24n
2
=
1
4n
2
.
Page 4
We want
1
4n
2
0.02 and so n
2
> 12.5, i.e. we should use n 4 panels.
Hence use the Midpoint Rule with h = 0.25:
_
1
0
x
4
2x
3
dx 0.25 [f(0.125) + f(0.375) + f(0.625) + f(0.875)] = 0.295
Question 8
Given the table
x 1 1.2 1.4
f(x) 1.0 2.5 2.0
use the Trapezoidal rule with 2 panels to estimate
_
1.4
1
f(x) dx. If we are told that
f(1.1) = f(1.3) = d and that the Trapezium Rule with 4 panels gives the value 0.9,
calculate d.
Solution:
With 2 panels, the Trapezoidal Rule gives
_
1.4
1
f(x) dx
0.2
2
[1.0 + 2 2.5 + 2.0] = 0.8
With 4 panels, the Trapezoidal Rule give
_
1.4
1
f(x) dx
0.1
2
[1.0 + 2d + 2 2.5 + 2d + 2.0] = 0.9
Rearranging we nd d = 2.5.
Question 9
Compute the integral
_
0.4
0
sin(x
2
) dx correct to 3 decimal places by repeatedly apply-
ing the Trapezoidal rule with 1, 2, 4, 8, . . . panels until the sequence stabilises.
Solution:
Write T(h) for the Trapezoidal Method approximation with interval width h.
T(0.4) =
0.4
2
_
sin(0) + sin(0.4
2
)

= 0.03186
T(0.2) =
0.2
2
_
sin(0) + 2 sin(0.2)
2
+ sin(0.4
2
)

= 0.02393
T(0.1) =
0.1
2
_
sin(0) + 2 sin(0.1
2
) + 2 sin(0.2)
2
+ 2 sin(0.3
2
) + sin(0.4
2
)

= 0.02195
T(0.05) =
0.05
2
_
sin(0) + 2 sin(0.05
2
) + + 2 sin(0.35
2
) + sin(0.4
2
)

= 0.02146
T(0.025) =
0.025
2
_
sin(0) + 2 sin(0.025
2
) + + 2 sin(0.375
2
) + sin(0.4
2
)

= 0.02134
The sequence has stabilised at 0.021.
Page 5
Question 10
Use Simpsons Rule with n = 6 intervals to estimate
_
4
1

1 + x
3
dx.
Solution:
Using Simpsons Rule with h =
41
6
= 0.5,
_
4
1

1 + x
3
dx
0.5
3
_

1 + 1
3
+ 4

1 + 1.5
3
+ 2

1 + 2
3
+ 4

1 + 2.5
3
+2

1 + 3
3
+ 4

1 + 3.5
3
+

1 + 4
3
_
= 12.8718
Question 11
Approximate
_
2
0
x
2
e
x
2
dx using intervals of width h = 0.5 using
(a) the Composite Midpoint Rule,
(b) the Composite Trapezoidal Rule,
(c) the Composite Simpsons Rule.
Solution:
(a) With the Midpoint Rule,
_
2
0
x
2
e
x
2
dx 0.5
_
0.25
2
e
0.25
2
+ 0.75
2
e
0.75
2
+ 1.25
2
e
1.25
2
+ 1.75
2
e
1.75
2
_
= 0.42498
(b) With the Trapezoidal Rule,
_
2
0
x
2
e
x
2
dx
0.5
2
_
0
2
e
0
2
+ 2 0.5
2
e
0.5
2
+ 2 1
2
e
1
2
+ 2 1.5
2
e
1.5
2
+ 2
2
e
2
2
_
= 0.41818
(c) With Simpsons Rule,
_
2
0
x
2
e
x
2
dx
0.5
3
_
0
2
e
0
2
+ 4 0.5
2
e
0.5
2
+ 2 1
2
e
1
2
+ 4 1.5
2
e
1.5
2
+ 2
2
e
2
2
_
= 0.42274
(The actual value is 0.42272506...)
Page 6
Question 12
Determine the interval width h required to approximate
_
1
0
sin(3x) dx to within
0.0001 using
(a) the Composite Midpoint Rule,
(b) the Composite Trapezoidal Rule,
(c) the Composite Simpsons Rule.
Solution:
(a) Set f(x) = sin(3x). Then f

(x) = 9 sin(3x) and |f

(x)| K = 9 for all


0 < x < 1. So we need
|E
M
|
(b a)h
2
K
24
=
9h
2
24
0.0001
i.e. h 0.0163. Hence taking h = 1/62 = 0.0161 i.e. 62 panels will suce.
(b) We can take K = 9 again and need
|E
T
|
(b a)h
2
K
12
=
9h
2
12
0.0001
i.e. h 0.1154. Hence taking h = 1/87 = 0.01149 i.e. 87 panels will suce.
(c) Now f
(4)
(x) = 243 sin(x) and |f
(4)
(x)| K = 243. We need
|E
S
|
(b a)h
4
K
180
=
243h
4
180
0.0001
i.e. h 0.0927. Hence taking h = 1/12 = 0.0833 i.e. 12 panels will suce.
(Remember for Simpsons Rule we need n to be even.)
Question 13
(a) How many intervals are required to approximate
_
1
0
(x
3
+x
5
) dx using Simpsons
Rule to within 0.0001 of the actual answer ?
(b) Do the numerical integration with this value of n.
Solution:
(a) Let f(x) = x
3
+x
5
. Then f
(4)
(x) = 120x so for 0 < x < 1, |f
(4)
(x)| < K = 120.
Hence the bound for the error term is Hence the error term satises
|E
S
|
K(b a)
5
180n
4
=
2
3n
4
.
We want
2
3n
4
0.0001 and so n
4
> 6667, i.e. we should use n 9 panels.
Simpsons Rule must use an even number so we take n = 10 panels of width
h = 0.1.
Page 7
(b) With Simpsons Rule,
_
1
0
(x
3
+ x
5
) dx
0.1
3
_
(0
3
+ 0
5
) + 4(0.1
3
+ 0.1
5
) + 2(0.2
3
+ 0.2
5
) + 4(0.3
3
+ 0.3
5
)
+2(0.4
3
+ 0.4
5
) + 4(0.5
3
+ 0.5
5
) + 2(0.6
3
+ 0.6
5
)
+4(0.7
3
+ 0.7
5
) + 2(0.8
3
+ 0.8
5
) + 4(0.9
3
+ 0.9
5
) + (1
3
+ 1
5
)

= 0.4167
Question 14
Determine the value of n such that the Trapezoidal Rule approximates the value of
_
3
0
ln(x
2
+ 2) dx
to within 0.01. Compute the estimated value.
Solution:
Let f(x) = ln(x
2
+ 2). Then f

(x) =
2x
x
2
+ 2
and f

(x) =
2(2 x
2
)
(x
2
+ 2)
2
.
To apply the error formula for the Trapezoidal rule, we need to calculate the max-
imum value of |f

(x)| for 0 < x < 3. This value must be at either the ends of the
interval or at a turning point. Now f

(0) = 1 and f

(3) =
14
121
.
Also, the turning points of f

(x) are found by solving


f

(x) =
4x(x
2
6)
(x
2
+ 2)
3
= 0
i.e. when x = 0 or x =

6. The only new value to calculate is f

6) =
1
8
.
Comparing these values, we nd the maximum K = |f

(0)| = 1.
The error term for the Trapezoidal rule satises
|E
T
|
K(b a)
3
12n
2
=
1 (3 0)
3
12n
2
=
9
4n
2
and we require this to be 0.01.
Hence n
2

9
40.01
= 225, so taking n = 15 will suce.
Applying the Trapezoidal Rule with h =
ba
n
=
3
15
= 0.2,
_
3
0
ln(x
2
+ 2) dx
0.2
2
_
ln(0
2
+ 2) + 2 ln(0.2
2
+ 2) + ... + 2 ln(2.8
2
+ 2) + ln(3
2
+ 2)

= 4.39243
Page 8
Question 15
The following table shows the speed of a snail at various distances in a race of overall
distance 20cm.
Distance (in cm) 0 5 10 15 20
Speed (in cm/sec) 4 8 9 6 4
Use Simpsons Rules to estimate how long it took to nish.
Solution:
We want to integrate 1/speed against distance. To see this, let p(t) be the distance
travelled after t seconds and T be the overall time. Then dp = p

(t)dt so
T =
_
T
0
dt =
_
20
0
1
p

(t)
dp.
(Also notice this has the right dimensions: distance/speed=time.)
Applying Simpsons Rule with h = 5, we nd
T
0.5
3
_
1
4
+ 4
1
8
+ 2
1
9
+ 4
1
6
+
1
4
_
= 3
4
27
sec.
To check that our estimate makes sense, notice that if the snail moved at constant
speed 4cm/sec it would nish in 5 seconds and if it moved at constant speed 9cm/sec
it would nish in about 2 seconds. Its a very fast snail !
Numerical Solutions of ODEs
Question 16
Apply Eulers Method to the initial value problem y

= t y, y(1) = 2 to estimate
y(1.1) and y(1.2) using stepsizes h = 0.1 and 0.05.
Repeat using the 2
nd
order Taylor Series Method and compare your results with the
exact solution.
Solution:
Set f(t, y) = t y. For Eulers method with h = 0.1,
y
i+1
= y
i
+ hf(t
i
, y
i
) = 0.9y
i
+ 0.1t
i
t
0
= 1, y
0
= y(1) = 2
t
1
= 1.1, y
1
= 0.9 2 + 0.1 1 = 1.9 y(1.1) 1.9
t
2
= 1.2, y
2
= 0.9 1.9 + 0.1 1.1 = 1.82 y(1.2) 1.82
Page 9
For Eulers method with h = 0.05,
y
i+1
= y
i
+ hf(t
i
, y
i
) = 0.95y
i
+ 0.05t
i
t
0
= 1, y
0
= y(1) = 2
t
1
= 1.05, y
1
= 0.95 2 + 0.05 1 = 1.95
t
2
= 1.1, y
2
= 0.95 1.95 + 0.05 1.05 = 1.905 y(1.1) 1.905
t
3
= 1.15, y
3
= 0.95 1.905 + 0.05 1.1 = 1.86475
t
4
= 1.2, y
4
= 0.95 1.86475 + 0.05 1.15 = 1.82901 y(1.2) 1.82901
For the 2
nd
order Taylor series method with h = 0.1, we need the derivative f

(t, y) =
1 y

= 1 t + y. Then
y
i+1
= y
i
+ hf(t
i
, y
i
) +
1
2
h
2
f

(t
i
, y
i
)
= y
i
+ 0.1(t
i
y
i
) + 0.005(1 t
i
+ y
i
)
= 0.905y
i
+ 0.095t
i
+ 0.005
t
0
= 1, y
0
= y(1) = 2
t
1
= 1.1, y
1
= 0.905 2 + 0.095 1 + 0.005 = 1.91 y(1.1) 1.91
t
2
= 1.2, y
2
= 0.905 1.91 + 0.095 1.1 + 0.005 = 1.83805 y(1.2) 1.83805
For the 2
nd
order Taylor series method with h = 0.05,
y
i+1
= y
i
+ hf(t
i
, y
i
) +
1
2
h
2
f

(t
i
, y
i
) = 0.95125y
i
+ 0.04875t
i
+ 0.00125
t
0
= 1, y
0
= y(1) = 2
t
1
= 1.05, y
1
= 0.95125 2 + 0.04875 1 + 0.00125 = 1.9525
t
2
= 1.1, y
2
= 0.95125 1.9525 + 0.04875 1.05 + 0.00125 = 1.90975
y(1.1) 1.90975
t
3
= 1.15, y
3
= 0.95125 1.905 + 0.04875 1.1 + 0.00125 = 1.87152
t
4
= 1.2, y
4
= 0.95125 1.86475 + 0.04875 1.15 + 0.00125 = 1.8376
y(1.2) 1.8376 .
The equation y

+ y = t is linear with integrating factor = e


t
and
(e
t
y)

= e
t
y

+ e
t
y = te
t
.
Integrating by parts gives e
t
y =
_
te
t
dt = e
t
(t 1) + C.
Since y(1) = 2, this means C = 2e and so y(t) = t 1 + 2e
1t
.
Hence y(1.1) = 0.1 + 2e
0.1
= 1.90967 and y(1.2) = 0.2 + 2e
0.2
= 1.83746.
Page 10
Question 17
Given
dy
dt
= 2ty and y(1) = 1, estimate y(1.2) using step size h = 0.1 with
(a) the Euler Method
(b) the Modied Euler Method
(c) the 4
th
order Runge-Kutta Method.
Solution:
(a) Set f(t, y) = 2ty. For Eulers method with h = 0.1,
y
i+1
= y
i
+ hf(t
i
, y
i
) = y
i
+ 0.2t
i
y
i
t
0
= 1, y
0
= y(1) = 1
t
1
= 1.1, y
1
= 1 + 0.2 1 1 = 1.2
t
2
= 1.2, y
2
= 1.2 + 0.2 1.1 1.2 = 1.464 y(1.2) 1.464
(b) For the Modied Eulers method with h = 0.1,
y
i+1
= y
i
+ hf(t
i
, y
i
) = y
i
+ 0.2t
i
y
i
y
i+1
= y
i
+
1
2
h(f(t
i
, y
i
) + f(t
i+1
, y
i+1
)) = y
i
+ 0.1t
i
y
i
+ 0.1t
i+1
y
i+1
t
0
= 1, y
0
= y(1) = 1
t
1
= 1.1, y
1
= 1 + 0.2 1 1 = 1.2
y
1
= 1 + 0.1 1 1 + 0.1 1.1 1.2 = 1.232
t
2
= 1.2, y
2
= 1.232 + 0.2 1.1 1.232 = 1.50304
y
2
= 1.232 + 0.1 1.1 1.232 + 0.1 1.2 1.50304 = 1.54788
y(1.2) 1.54788
(c) For the 4
th
order Runge-Kutta Method with h = 0.1,
k
1
= hf(t
i
, y
i
) = 0.2t
i
y
i
k
2
= hf(t
i
+ h/2, y
i
+ k
1
/2) = 0.2(t
i
+ 0.05)(y
i
+ 0.5k
1
)
k
3
= hf(t
i
+ h/2, y
i
+ k
2
/2) = 0.2(t
i
+ 0.05)(y
i
+ 0.5k
2
)
k
4
= hf(t
i+1
, y
i
+ k
3
) = 0.2t
i+1
(y
i
+ k
3
)
y
i+1
= y
i
+
1
6
(k
1
+ 2k
2
+ 2k
3
+ k
4
)
t
0
= 1, y
0
= y(1) = 1
t
1
= 1.1, k
1
= 0.2 1 1 = 0.2
k
2
= 0.2 1.05 (1 + 0.5 0.2) = 0.231
k
3
= 0.2 1.05 (1 + 0.5 0.231) = 0.234255
Page 11
k
4
= 0.2 1.1 (1 + 0.234255) = 0.271536
y
1
= 1 +
1
6
(0.2 + 2 0.231 + 2 0.234255 + 0.271536) = 1.23367
t
2
= 1.2, k
1
= 0.2 1.1 1.23162 = 0.271408
k
2
= 0.2 1.15 (1.23367 + 0.5 0.271408) = 0.314957
k
3
= 0.2 1.15 (1.23367 + 0.5 0.314957) = 0.319965
k
4
= 0.2 1.2 (1.23367 + 0.319965) = 0.372873
y
2
= 1.23367+
1
6
(0.271408+20.314957+20.319965+0.372873) = 1.55270
y(1.2) 1.55270
Question 18
Given y

= sin(ty) and y(0) = 1 estimate y(0.2) using the Midpoint Method with
stepsize h = 0.2 and h = 0.1.
Writing M(h) for the estimate of y(0.2) with stepsize h, use Richardsons extrapola-
tion i.e. compute
4M(h/2) M(h)
3
to get a better estimate.
Solution:
For the Midpoint method with h = 0.2,
y
i+
1
2
= y
i
+
1
2
hf(t
i
, y
i
) = y
i
+ 0.1 sin(t
i
y
i
)
y
i+1
= y
i
+ hf
_
t
i+
1
2
, y
i+
1
2
_
= y
i
+ 0.2 sin
_
t
i+
1
2
y
i+
1
2
_
t
0
= 0, y
0
= y(0) = 1
t1
2
= 0.1, y1
2
= 1 + 0.1 sin(0 1) = 1
t
1
= 0.2, y
1
= 1 + 0.2 sin(0.1 1) = 1.01997 y(0.2) M(0.2) = 1.01997
For the Midpoint method with h = 0.1,
y
i+
1
2
= y
i
+
1
2
hf(t
i
, y
i
) = y
i
+ 0.05 sin(t
i
y
i
)
y
i+1
= y
i
+ hf
_
t
i+
1
2
, y
i+
1
2
_
= y
i
+ 0.1 sin
_
t
i+
1
2
y
i+
1
2
_
t
0
= 0, y
0
= y(0) = 1
t1
2
= 0.05, y1
2
= 1 + 0.05 sin(0 1) = 1
t
1
= 0.1, y
1
= 1 + 0.1 sin(0.05 1) = 1.00500
t3
2
= 0.15, y3
2
= 1.00500 + 0.05 sin(0.1 1.00500) = 1.01001
t
2
= 0.2, y
2
= 1.00500 + 0.1 sin(0.15 1.01001) = 1.02009
y(0.2) M(0.1) = 1.02009
Page 12
Applying Richardsons Extrapolation, a better estimate is
y(0.2)
4M(0.1) M(0.2)
3
= 1.02013
Question 19
Use the Modied Euler Method with steplength 0.1 to estimate y(0.2), given that
dy
dt
= sin(y
2
) and y(0) = 1.
Keep 5 decimal places in your calculations.
Solution:
For the Modied Eulers method with h = 0.1,
y
i+1
= y
i
+ hf(t
i
, y
i
) = y
i
+ 0.1 sin(y
2
i
)
y
i+1
= y
i
+
1
2
h(f(t
i
, y
i
) + f(t
i+1
, y
i+1
)) = y
i
+ 0.05 sin(y
2
i
) + 0.05 sin( y
2
i+1
)
t
0
= 0, y
0
= y(0) = 1
t
1
= 0.1, y
1
= 1 + 0.1 sin(1
2
) = 1.08415
y
1
= 1 + 0.05 sin(1
2
) + 0.05 sin(1.08415
2
) = 1.04669
t
2
= 0.2, y
2
= 1.04669 + 0.1 sin(1.04669
2
) = 1.13561
y
2
= 1.04669 + 0.05 sin(1.04669
2
) + 0.05 sin(1.13561
2
) = 1.13898
y(0.2) 1.13898
Question 20
Given the system of 1
st
order equations
x

= 2tx + 3y
2
x(0) = 1
y

= 3x
2
(1 y) y(0) = 2
use Eulers Method with stepsize h = 0.1 to compute approximations for x(t) and
y(t) at time t = 0.1 and 0.2.
Question 21
Write y

= ty as a pair of equations of 1
st
order. Hence apply Eulers Method with
stepsizes h = 0.1 and h = 0.05 to estimate y(0.2) given that y(0) = y

(0) = 1.
Question 22
Consider the equation for a simple harmonic oscillator i.e.
y

= y, y(0) = 0, y

(0) = 1
Find an estimate for y(0.4) by writing the equation as a system and then applying
the Modied Euler Method with stepsize h = 0.2.
Solve the equation exactly and compare your estimate with the actual value.
Page 13
Question 23
Repeat the previous question using the 4
th
order Runge-Kutta Method with a single
step of stepsize h = 0.4.
Question 24
Consider a driven simple harmonic oscillator equation
y

= 4y 2y

+ sin(t), y(0) = 0, y

(0) = 1
Estimate y(0.4) by applying the Euler Method with stepsize h = 0.2.
Page 14

S-ar putea să vă placă și