Week 2 Roots of Equation

1. Graphical Methods
Graphical methods provide visual insight into techniques.

Graphical approach.
c
gm  t
f (c )  (1  e m )  v, v  40 @ t  10 sec
c
9.8  68.1
c
 10
f (c )  (1  e 68.1 )  40
c

c 4 8 12 16 20

Function 34.11484 17.65343 6.066936 -2.26876 -8.40063

Value of function for different Drag Coefficient

40
35
30
25
20
Function

15
10
5
0
-5
-10
-15
4 6 8 10 12 14 16 18 20
Drag Coefficient

Root is between 14 & 16, roughly 14.75.

14.75
667.38 10
f (14.75)  (1  e 68.1
)  40  0.059 close to zero
14.75

2. Bracketing Methods.
Start with guesses that bracket, the root and then systematically reduce the width of the
bracket. It is based on the principle that Function changes its sign in the vicinity of root.
Both method assume Upper and lower value.

There are two methods.

 Bisection

 False Position.

Bisection Method
1) xL , xu are lower and upper bounds.

If f ( xL ) f ( xu )  0 At  least one Root between xL & xu

x L  xu
2) xc 
2

If f ( xL ) f ( xc )  0 Root between xL & xc

xu  xc Go To Step  2

If f ( xL ) f ( xc )  0 Root between xc & xu

xL  xc Go To Step  2

If f ( xL ) f ( xc )  0 Root  xc

Problem1

667.38
f (c )  (1  e 0.14684c )  40
c
xL  12 xu  16
f (12)  6.066936 and f (16)  2.26876
f (12) f (16)  0
12  16
xc   14, f (14)  1.568699
2
f (12). f (14)  0. Set xL  14.
14  16
xc   15 f (15)  0.42484
2
f (14) f (15)  0 Set xu  15.
14  15
xc   14.5 f (14.5)  0.552319
2
f (14). f (14.5)  0. Set xL  14.5
14.5  15
xc   14.75 f (14.75)  0.058954
2
f (14.5). f (14.75)  0. Set xL  14.75
14.75  15
xc   14.875 f (14.875)  0.18413
2
f (14.75). f (14.875)  0. Set xu  14.875
14.75  14.875
xc   14.8125 f (14.8125)  0.06288
2
f (14.75). f (14.8125)  0. Set xu  14.8125
14.75  14.78125
xc   14.8125 f (14.78125)  0.00204
2

True value is 14.7802734375

ITR XL XU XC f X C  a
1 12.00000 16.00000 14.00000 1.56870 100.00000
2 14.00000 16.00000 15.00000 -.42484 6.66667
3 14.00000 15.00000 14.50000 .55232 3.44828
4 14.50000 15.00000 14.75000 .05895 1.69492
5 14.75000 15.00000 14.87500 -.18413 .84034
6 14.75000 14.87500 14.81250 -.06288 .42194
7 14.75000 14.81250 14.78125 -.00204 .21142
8 14.75000 14.78125 14.76563 .02844 .10582
9 14.76563 14.78125 14.77344 .01320 .05288
10 14.77344 14.78125 14.77734 .00558 .02643
11 14.77734 14.78125 14.77930 .00177 .01322
12 14.77930 14.78125 14.78027 -.00014 .00661
Re ad xL & xu
Calculate f ( xL ) & f ( xu )
If f ( xL )  f ( xu )  0 Then  ( If N 0) 
 ( If Yes)
x  xu
xc  L , Calculate f ( xc )
2
If f ( xL )  f ( xc )  0 xu  xc
If f ( xL )  f ( xc )  0 xL  xc , f ( xL )  f ( xc )

Problem 2
f ( x)  x 3  4 x 2  10 has root in 1,2

ITR XL XU XC f X C  a
1 1.00000 2.00000 1.50000 2.37500 100.00000
2 1.00000 1.50000 1.25000 -1.79688 20.00000
3 1.25000 1.50000 1.37500 .16211 9.09091
4 1.25000 1.37500 1.31250 -.84839 4.76190
5 1.31250 1.37500 1.34375 -.35098 2.32558
6 1.34375 1.37500 1.35938 -.09641 1.14943
7 1.35938 1.37500 1.36719 .03236 .57143
8 1.35938 1.36719 1.36328 -.03215 .28653
9 1.36328 1.36719 1.36523 .00007 .14306

Root 1.36523438
False Position Method
In bisection method no account is taken of the magnitude of f ( xL ) & f ( xu ). From similar triangles
 f ( xL ) f ( xu )

xc  x L xu  x c
 f ( x L )( xu  xc )  f ( xu )( xc  x L )
 xu f ( x L ).  xc . f ( x L )  xc . f ( xu )  x L f ( xu )
xc { f ( x L )  f ( xu )}  . xu f ( x L )  x L f ( xu )
x . f ( x L )  x L f ( xu ) xu f ( x L )  xu f ( xu )  xu f ( xu )  x L f ( xu )
xc  u 
f ( x L )  f ( xu ) f ( xl )  f ( xu )
( x  x L ) f ( xu )
x c  xu  u
f ( xl )  f ( xu )
( xu  x L )
x c  xu  f ( xu )
f ( xu )  f ( x L )

f ( xu )( xu  x L )
For x L & xu find f ( x L ). & f ( xu ) Then x c  xu 
f ( xu )  f ( x L )

Problem 1
ITR XL XU XC f X C  a
1 12.00000 16.00000 14.91130 -.25428 100.00000
2 12.00000 14.91130 14.79419 -.02726 .79160
3 12.00000 14.79419 14.78170 -.00291 .08455
4 12.00000 14.78170 14.78036 -.00031 .00901

Problem 2
ITR XL XU XC f X C  a
1 1.00000 2.00000 1.26316 -1.60227 100.00000
2 1.26316 2.00000 1.33883 -.43036 5.65196
3 1.33883 2.00000 1.35855 -.11001 1.45144
4 1.35855 2.00000 1.36355 -.02776 .36677
5 1.36355 2.00000 1.36481 -.00698 .09229
6 1.36481 2.00000 1.36512 -.00176 .02320
7 1.36512 2.00000 1.36520 -.00044 .00583
PROGRAM ROOT
REAL*8 XL, XU, XC, FXL, FXU, FXC
OPEN(5,FILE='INPUT',STATUS='UNKNOWN')
OPEN(6,FILE='OUT',STATUS='UNKNOWN')
READ(5,*) MODE ! MODE = 1 Bisection, MODE = 2 False Position
READ(5,*) XL, XU
FXL = FUN(XL) ; FXU = FUN(XU); ITR = 0
PROD = FXL * FXU
IF ( PROD > 0.0 ) THEN
WRITE(*,*) 'TRY DIFFERENT SET'
STOP
ENDIF
10 ITR = ITR +1
IF ( MODE == 1 ) XC = (XL + XU)/2.0
IF ( MODE == 2 ) XC = XU - (XU - XL)*FXU /(FXU -FXL)

FXC = FUN(XC)
TERM = 100.0 * ABS( (XC-XCOLD)/XC )
WRITE(6,'(I3,5(1X,F10.5))')ITR, XL, XU, XC, FXC, TERM

IF ( (TERM < 0.01) .OR. ABS(FXC) < 0.0005 ) THEN

WRITE(6,'(A,F14.8)') 'ROOT ', XC
STOP
ENDIF
IF ( ITR > 20 ) THEN
WRITE(6,*) 'DOES NOT CONVERGE'
STOP
ENDIF

PROD = FXL * FXC ; XCOLD = XC

IF ( PROD < 0.0 ) THEN
XU = XC ; FXU = FXC
GO TO 10
ELSE
XL = XC ; FXL = FXC
GO TO 10
ENDIF

END PROGRAM ROOT

FUNCTION FUN(X)
REAL*8 X
FUN = 667.38 * ( 1.0 - EXP(-X/6.81) )/X - 40.0
RETURN
END
Open Methods Iteration X
Newton Raphson Method 1 0.5 1.3
2 51.65 1.17943
f ( x)
xi 1  xi  3 46.485 1.084131
f ' ( x) 4 41.8365 1.024053
F 5 37.65285 1.002389
F ( x)  e  x  x  e  x  1
x 6 33.88757 1.000025
7 30.49881 1
Iteration X F(X) F'(X)
8 27.44893 1
1 0 1 -2
9 24.70403
2 0.5 0.106531 -1.60653
10 22.23363
3 0.566311 0.001305 -1.56762
11 20.01027
4 0.567143 1.96E-07 -1.56714
12 18.00924
5 0.567143 4.44E-15 -1.56714
13 16.20832
6 0.567143 0 -1.56714
14 14.58749
7 0.567143 0 -1.56714
15 13.12874
16 11.81586
17 10.63428
F
F ( x)  x10  1  10 x 9 18 9.570849
x 19 8.613764
Newton Raphson Method 20 7.752388
X F(X) F'(X) 21 6.977149
1.3 12.7858492 106.045 22 6.279434
1.17943 4.20860657 44.16207 23 5.651491
1.084131 1.24293636 20.6888 24 5.086342
1.024053 0.2683079 12.38518 25 4.577708
1.002389 0.02415311 10.21712 26 4.119937
1.000025 0.00025472 10.00229 27 3.707944
1 2.9188E-08 10 28 3.33715
1 0 10 29 3.003437
Computation with Guess value 1.3 30 2.703098
31 2.432801
Determine the highest real root of 32 2.189555
f ( x)  x 3  6 x 2  11x  6.1 33 1.970686
Using the Newton Raphson method, perform four 34 1.77384
iteration with starting value 3.5. 35 1.597031
36 1.438808
f ( x)  x 3  6 x 2  11x  6.1 37 1.298711
f ' ( x)  3x 2  12 x  11 38 1.178355
39 1.08335
X F(X) F'(X) Error 40 1.023665
3.5 1.775 5.75 41 1.002316
3.191304 0.399402 3.257618 9.673025 42 1.000024
3.068699 0.05188 2.426352 3.995359 43 1
3.047317 0.001456 2.290617 0.701669
3.046681 1.27E-06 2.286624 0.020864
3.046681 9.68E-13 2.28662 1.82E-05
3.046681 0 2.28662 1.39E-11

The effect of starting value is clear from the next

trial. If same example is solved with starting
value of 0.5 , it requites 43 iteration as compare
to those 7 with 1.3 as starting value.
Multiple Roots
f ( x)  ( x  3)( x  1)( x  1)  x 3  5x 2  7 x  3
In multiple roots, function does change sign at
even multiple roots. Main problem is related to
the fact that not only f(x) but also f’(x) tends to
zero. A simple way to overcome this difficulty is
based on the fact that always f(x) reach zero
before f’(x).
Ralston and Rabinowitz
f ( x)
xl 1  xl  m m number of multiple
f ' ( x)
roots.
Another approach is assume function u(x), such
that u(x) has same root as f(x).
f ( x)
u ( x) 
f ' ( x)
u ( x)
xl 1  xl 
u ' ( x)
f ' ( x). f ' ( x)  f ( x). f " ( x)
u ' ( x) 
( f ' ( x)) 2
f ( x) ( f ' ( x)) 2
xl 1  xl  .
f ' ( x) ( f ' ( x)) 2  f ( x). f " ( x))
f ( x). f ' ( x)
xl 1  xl 
f ' ( x) 2  f ( x). f " ( x)

F ( x)  x 3  5 x 2  7 x  3
F 2F
 3x 2  10 x  7  6 x  10
x x 2
Compare two approach
x 3  5x 2  7 x  3
1. x New  x  ( )
3x 2  10 x  7
2.
( x 3  5 x 2  7 x  3)(3x 2  10 x  7)
x New  x 
(3x 2  10 x  7) 2  ( x 3  5 x 2  7 x  3)(6 x  10)
Iteration X App Error True Error Iteration X True Error
0 0 100 0 0 100
1 0.42857143 100 57.14286 1 1.105263 10.52632
2 0.68571429 37.5 31.42857 2 1.003082 0.308166
3 0.8328654 17.66805 16.71346 3 1.000002 0.000238
4 0.91332989 8.810014 8.667011 4 1 3.73E-09
5 0.95578329 4.441739 4.421671
6 0.9776551 2.23717 2.23449
7 0.98876617 1.12373 1.123383
8 0.99436744 0.5633 0.563256
9 0.99717977 0.282028 0.282023
10 0.99858889 0.141111 0.141111
11 0.9992942 0.07058 0.07058
12 0.99964704 0.035296 0.035296
13 0.9998235 0.01765 0.01765
14 0.99991175 0.008825 0.008825
15 0.99995587 0.004413 0.004413
16 0.99997794 0.002206 0.002206
17 0.99998897 0.001103 0.001103
18 0.99999448 0.000552 0.000552
19 0.99999724 0.000276 0.000276
20 0.99999862 0.000138 0.000138
21 0.99999931 6.89E-05 6.9E-05
22 0.99999966 3.45E-05 3.45E-05
Systems of Nonlinear Equation
Linear simultaneous equations in next
chapter.
f ( x)  a1 x1  a2 x2  ...  an xn  b  0
Where a & b are constants.
Algebraic and transcendental equations that
do not fit this formate are called non linear
equations.
f1 ( x1 , x 2 ,...., x n )  0
f 2 ( x1 , x 2 ,...., x n )  0
.
.
f n ( x1 , x 2 ,....., x n )  0
Locate the roots of a set simultaneous
equations.

Fixed Point Iteration

Condition for convergence
u v
 1
x x
u v
 1
y y

u  x 2  xy  10 v  y  3xy 2  57
57  y
x  10  xy y
3x

Iteration X Y
0 1.5 3.5
1 2.17944947 2.860506
2 1.94053388 3.049551
3 2.02045629 2.983405
4 1.99302813 3.005704
5 2.00238524 2.998054
6 1.99918491 3.000666
7 2.00027865 2.999773
8 1.99990475 3.000078
9 2.00003256 2.999973
10 1.99998887 3.000009
11 2.0000038 2.999997
12 1.9999987 3.000001
13 2.00000044 3
14 1.99999985 3
15 2.00000005 3
16 1.99999998 3
17 2.00000001 3
18 2 3
Newton Raphson Method
Taylor’s Series
u i u
u i 1  u i  ( xi 1  xi )  ( yi 1  yi ) i
x y
v v
vi 1  vi  ( xi 1  xi ) i  ( yi 1  yi ) i
x y
Root estimate corresponds to the value of x and y where ui 1 and vi 1 are equal to zero.
u l u u u
xi 1  l yi 1  u i  xi i  yi i
x y x y
vl v v v
xi 1  l yi 1  vi  xi i  yi i
x y x y
1
 u i u i   vi u 
 x  i
y  1  y y
    
 vi vi  u i vi u i vi  vi
 
u i 
 x y  x y y x  x x 
1 v u v u v u u v u v
xi 1  {ui i  xi i i  yi i i  vi i  xi i i  yi i i }
ui vi ui vi y x y y y y y x y y

x y y x
v u
u l 1  vl l
y y
xl 1  xl 
u l v1 u l v1

x y y x
1 v u v u v u u v u v
yl 1  [u l 1  xl l 1  yl l 1  vl l  xl l 1  yl l 1 ]
u l v1 u l v1 x x x y x x x x x y

x y y x
u v
vl 1  u l l
yl 1  yl  x x
u v u v
( l 1  l 1)
x y y x

v u u v
v u u
v
y y x x
xN  x  and yN  y 
u v u v u v u v
 
x y y x x y y x

u u v v
 2x  y x  3y2  1  6 xy
x y x y
X Y U V U'x U'y V'x V'y Determin.
1.5 3.5 -2.5 1.625 6.5 1.5 36.75 32.5 156.125
2.036029 2.8438751 -0.06437 -4.75621 6.915933 2.036029 24.26288 35.74127 197.7843
1.998701 3.00228856 -0.00452 0.049571 6.99969 1.998701 27.04121 37.00406 204.9696
2 2.99999941 -1.3E-06 -2.2E-05 6.999999 2 26.99999 36.99999 204.9999
2 3 0 2.24E-12 7 2 27 37 205
Polynomial Evaluation and Differentiation

Evaluation by Deflation
f ( x)  a0  a1 x  a 2 x 2  a3 x 3
f ( x)  a0  x[a1  x(a 2  xa 3 )]
f ' ( x)  a1  2a 2 x  3a3 x 2
f ' ( x)  a1  x(2a 2  3a3 x)
Do I  n,0,1
Df  Df * x  p
p  p * x  ai
EndDo

an
a n x  a n 1
a n x 2  a n 1 x  a n  2
Let n  3, P  a3 I 3
a3 x  a 2 I 2
a3 x 2  a 2 x  a1 I 1
a3 x  a 2 x  a1 x  a0
3 2
I 0

Df  0 I 3
Df  a3 I 2
Df  a3 x  a3 x  a 2  2a3 x  a 2 ; I  1
Df  (2a3 x  a 2 ) x  a3 x 2  a 2 x  a1
Df  3a3 x 3  2a 2 x  a1

Muller’s Method
Projects a parabola through three points.
f 2 ( x )  a ( x  x 2 ) 2  b( x  x 2 )  c
f ( x 0 )  a ( x 0  x 2 ) 2  b( x 0  x 2 )  c
f ( x1 )  a( x1  x 2 ) 2  b( x1  x 2 )  c
f ( x 2 )  a ( x 2  x 2 ) 2  b( x 2  x 2 )  c
c  f ( x2 )
f ( x 0 )  f ( x 2 )  a ( x 0  x 2 ) 2  b( x 0  x 2 )
f ( x1 )  f ( x2 )  a( x1  x 2 ) 2  b( x1  x2 )
f ( x0 )  f ( x 2 )
0   a ( x0  x 2 )  b
( x0  x 2 )
f ( x1 )  f ( x 2 )
1   a( x1  x 2 )  b
( x1  x 2 )
 f ( x0 )  f ( x 2 ) f ( x1 )  f ( x 2 )
a( x0  x1 )  
( x0  x 2 ) x1  x 2
x1 f ( x0 )  x1 f ( x 2 )  x 2 f ( x0 )  x 2 f ( x 2 )  x0 f ( x1 )  x0 f ( x 2 )  x 2 f ( x1 )  x 2 f ( x 2 )
a
( x0  x1 )( x0  x 2 )( x1  x 2 )
( x1  x 2 ) f ( x0 )  ( x 2  x0 ) f ( x1 )  ( x0  x1 ) f ( x 2 )
a
( x0  x1 )( x0  x 2 )( x1  x 2 )
a( x0  x1 )   0   1
 0  1 ( x0  x 2 )
a , b   0  a( x0  x 2 )   0  ( 0   1 )
x0  x1 ( x0  x1 )

 b  b 2  4ac  b b 2  4ac b 2  (b 2  4ac)

2a  b b 2  4ac  2a(b  b 2  4ac )
 2c
x3  x 2 
b  b 2  4ac

Problem
f ( x)  x 3  13x  12
x0  4.5, x1  5.5, x 2  5
f ( x0 )  f (4.5)  20.625
f ( x1 )  f (5.5)  82.875
f ( x 2 )  f (5)  48
 0  a( x0  x 2 )  b
 1  a( x1  x 2 )  b
f ( x0 )  f ( x 2 ) 20.625  48
0  
x0  x 2 45  5.0
 270375
0   54.75
 0 .5

f ( x1 )  f ( x 2 ) 82.875  48 34.875
1   
x1  x 2 5.5  5 0.5
 1  69.75
54.75  a(4.5  5)  b
69.75  a(5.5  5)  b
( 0   1 )  a( x0  x1 ) a  15
69.75  7.5  b, b  62.25
c  48
15( x  x 2 ) 2  62.25( x  x 2 )  48  0
 2c
x3  5 
b  b 2  4ac
 96
 5  3.976
62.25  62.25 2  4  15  48
Repeat with 5.5, 5, 3.976
 X0 X1 X2 F(x0) F(x1) C=F(x2) D0 D1 a b
4.5 5.5 5 20.625 82.875 48 54.75 69.75 15 62.25
5.5 5 3.976487 82.875 48 -0.81633 54.93313 47.69488 14.47649 32.87801
5 3.97648704 4.00105 48 -0.81633221 0.036781 48.01366 34.73098 12.97754 35.04975
3.976487 4.0010505 4.000001 -0.816332211 0.036780702 2.47E-05 34.71841 35.01262 11.97754 35.00004
4.00105 4.00000071 4 0.036780702 2.46845E-05 1.74E-11 35.01261 35.00001 12.00105 35
4.000001 4 4 2.46845E-05 1.74083E-11 0 35.00001 35 11.99937 35
Bairstow’s Method r  1  C2  C3    b1 
  2  C  
f ( x)  a0  a1 x  a2 x      an x C 2   b0 
2 n.
s  C 2  C1C3  1
Divide by (x-t).
r  1 b0 C3  b1C 2 
f n1 ( x)  b1  b2 x  b3 x 2      bn x n1 .   2  
s  C 2  C1C3 b1C1  b0 C 2 
remainder R  b0
bn  a n bn  a n
bi  ai  bi 1t bn 1  a n 1  r bn
For i  n  1 to 0
C n  bn
To permit, Evaluation of complex roots,
C n 1  bn 1  r bn
Division by x 2  r x  s
do i  n  2, 0,  1
bi  ai  r bi 1  s bi  2
f n2 ( x)  b2  b3 x      bn1 x n3  bn x n2 . C i  bi  r Ci  2  s C i  2
End do.
Remainder R  b0  b1 ( x  r ) Problem
Synthetic division. f ( x)  x 5  3.5x 4  2.75x 3  2.125x 2  3.875x  1.25
bn  a n
a5 a4 a3 a2 a1 a0
bn1  a n 1  r bn .
1 -3.5 2.75 2.125 -3.875 1.25
bi  al  r bi 1  s bi  2 First Iteration
For i  n  1 to 0 Start with r  s  1,
For remainder to be zero b1 and b0 must be b5  1, b4  4.5, b3  6.25, b2  0.375,
zero. b1  10.5, b0  11.375
To modify our guess, use Taylor’s series. C5  1, C4 ,  5.5, C3  10.75, C2  4.875,
b b
b1 (r  r , s  s)  b1  1 r  1 s C1  16.375
r s
Det  199.7969, r  0.35583, s  1.138109
b b
b 0 (r  r , s  s)  b0  0 r  0 s.
r s
x 5  3.5 x 4  2.75 x 3  2.125 x 2  3.875  1.25  0
For exact guess, LHS is equal to zero.
b1 b ( x 2  0.5 x  0.5)( x 3  4 x 2  5.25 x  2.5)  0
r  1 s  b1 .
r s ( x 2  0.5 x  0.5)  0
b0 b  0.5  0.52  4  1  (0.5)  0.5  1.5
r  0 s  b0 . Roots  
r s 2 2
If the partial derivatives of the b’s can be  0.5,1
determined, then solve above two equations
can be solved simultaneously for r & s.
C n  bn
C n1  bn1  rCn .
Ci  bi  r Ci 1  s Ci  2
b0 b b b
 C1 , 0  C 2 , 1  C 2 , 1  C3 .
r s r s
C 2 r  C3 s  b1
C1r  C3 s  b0
C 2 C3  r    b1 
C   
 1 C 2  s   b0 
 f ( x)  x 5  3.5x 4  2.75x 3  2.125x 2  3.875x  1.25
f ( x)  x 5  3.5x 4  2.75x 3  2.125x 2  3.875x  1.25

R s b5 b4 b3 b2 b1 b0 c5 c4 c3 c2 c1 Det dr ds
-1 -11 -4.5 6.25 0.375 -10.5 11.375 1 -5.5 10.75 -4.875 -16.375 199.7969 0.35583 1.138109
-0.64417 0.1381091 -4.14417 5.557658 -2.02742 -1.80143 2.130423 1 -4.78834 8.780271 -8.34472 4.786621 27.60656 0.133057 0.331625
-0.51111 0.4697341 -4.01111 5.269866 -2.45265 -0.14598 0.172521 1 -4.52223 8.050969 -8.69185 8.078345 10.50975 0.011427 0.030469
-0.49969 0.5002021 -3.99969 5.248792 -2.4984 -0.00112 0.000855 1 -4.49937 7.99727 -8.74513 8.36895 9.548493 -0.00031 -0.0002
-0.5 0.51 -4 5.25 -2.5 4.88E-07 -4.6E-07 1 -4.5 8.000001 -8.75 8.375002 9.562502 6.53E-08 1.04E-08
-0.5 0.51 -4 5.25 -2.5 3.91E-14 -5.6E-14 1 -4.5 8 -8.75 8.375 9.5625 -1.1E-14 -1.7E-14

x 3  4 x 2  5.25x  2.5  0
a3 a2 a1 a0
1 -4 5.25 -2.5

R s b3 b2 b1 b0 c3 c2 c1 Det dr ds
-0.5 0.5 1 -4.5 8 -8.75 1 -5 11 14 2.232143 3.160714
1.732143 3.660714 1 -2.26786 4.982462 -2.17164 1 -0.53571 7.715242 -7.42825 -0.06698 -5.01834
1.665164 -1.35763 1 -2.33484 0.004486 0.677311 1 -0.66967 -2.46826 2.916716 0.233247 0.151713
1.898411 -1.20592 1 -2.10159 0.054404 0.137621 1 -0.20318 -1.53723 1.578508 0.094187 -0.03527
1.992598 -1.24118 1 -2.0074 0.008871 0.009231 1 -0.0148 -1.26181 1.262029 0.007418 -0.00876
2.000017 -1.24994 1 -1.99998 5.5E-05 -2.1E-05 1 3.32E-05 -1.24982 1.249824 -1.7E-05 -5.5E-05
2 -1.25 1 -2 2.75E-10 2.38E-09 1 -3.8E-09 -1.25 1.25 1.9E-09 -2.8E-10
2 -1.25 1 -2 0 0 1 0 -1.25 1.25 0 0
x 3  4 x 2  5.25 x  2.5  0
( x 2  2 x  1.25)( x  2)  0
x 2  2 x  1.25  0
2 45 2i
Roots    1  0.5 i
2 2
 LINEAR ALGEBRAIC EQUATION
Gauss Elimination Method (Direct method)

Consider the system of Equations as below

7 x1  2 x 2  3x3  12
2 x1  5 x 2  3x3  20
x1  x 2  6 x3  26
Example 1

Use Gauss Elimination to solve the System of Equations

7 x1  2 x 2  3x3  12
2 x1  5 x 2  3x3  20
x1  x 2  6 x3  26

Gauss Elimination

7 2 -3 -12
2 5 -3 -20
1 -1 -6 -26

7 2 -3 -12
0 4.428571 -2.14286 -16.5714
0 -1.28571 -5.57143 -24.2857

7 2 -3 -12
0 4.428571 -2.14286 -16.5714
0 0 -6.19355 -29.0968

X3= 4.697917
X2= -1.46875
X1= 0.71875

LU Decomposition method

Example 1

Find the LU decomposition of the matrix

 25 5 1
A   64 8 1
144 12 1
Solution

A  LU 
1 0 0 u11 u12 u13 
  21 1 0  0 u22 u23 

  
 31  32 1  0 0 u33 
The U  matrix is the same as found at the end of the forward elimination of Naï
ve
Gauss elimination method, that is
 25 5 1 

U    0  4.8  1.56
 0 0 0.7 
To find  21 and  31 , find the multiplier that was used to make the a 21 and a 31
elements zero in the first step of forward elimination of the Naï ve Gauss elimination
method. It was
64
 21 
25
 2.56
144
 31 
25
 5.76
To find  32 , what multiplier was used to make a32 element zero? Remember a 32
element was made zero in the second step of forward elimination. The A matrix at
the beginning of the second step of forward elimination was
25 5 1 
 0  4.8  1.56 
 
 0  16.8  4.76
So
 16.8
 32 
 4.8
 3.5
Hence
 1 0 0

L  2.56 1 0
5.76 3.5 1
Confirm LU   A .
 1 00 25 5 1 
LU   2.56 1  
0  0  4.8  1.56
5.76 3.5 1  0 0 0.7 
 25 5 1
  64 8 1
144 12 1

Example 2
Use the LU decomposition method to solve the following simultaneous linear
equations.
 25 5 1  a1  106.8 
 64 8 1 a   177.2 
   2  
144 12 1  a3  279.2
Solution
Recall that
AX   C 
and if
 A  LU 
then first solving
LZ   C 
and then
U X   Z 
gives the solution vector X  .
Now in the previous example, we showed
A  LU 
 1 0 0 25 5 1 
 2.56 1 0  0  4.8  1.56
  
5.76 3.5 1  0 0 0.7 
First solve
LZ   C 
 1 0 0  z1  106.8 
2.56 1 0  z   177.2 
  2   
5.76 3.5 1  z 3  279.2
to give
z1  106.8
2.56z1  z2  177.2
5.76z1  3.5z2  z3  279.2
Forward substitution starting from the first equation gives
z1  106.8
z2  177.2  2.56z1
 177.2  2.56 106.8
 96.208
z3  279.2  5.76z1  3.5z2
 279.2  5.76 106.8  3.5   96.208
 0.76
Hence
 z1 
Z    z 2 
 z 3 
 106.8 
  96.208
 0.76 
This matrix is same as the right hand side obtained at the end of the forward
elimination steps of Naï ve Gauss elimination method. Is this a coincidence?
Now solve
U X   Z 
25 5 1   a1   106.8 
 0  4.8  1.56 a    96.208
   2  
 0 0 0.7   a3   0.76 
25a1  5a2  a3  106.8
 4.8a2  1.56a3  96.208
 0.7a3  0.76
From the third equation
0.7a3  0.76
0.76
a3 
0.7
 1.0857
Substituting the value of a3 in the second equation,
 4.8a2  1.56a3  96.208
 96.208  1.56a3
a2 
 4.8
 96.208  1.56  1.0857

 4.8
 19.691
Substituting the value of a 2 and a3 in the first equation,
25a1  5a2  a3  106.8
106.8  5a 2  a3
a1 
25
106.8  5  19.691  1.0857

25
 0.29048
Hence the solution vector is
 a1  0.29048
a    19.691 
 2  
 a3   1.0857 

How do I find the inverse of a square matrix using LU decomposition?

A matrix B  is the inverse of A if

AB  I   BA.
How can we use LU decomposition to find the inverse of the matrix? Assume the first
column of B  (the inverse of A ) is
[b11 b12 ... ... bn1 ]T
Then from the above definition of an inverse and the definition of matrix multiplication
b11  1
b  0
A 21    
   
   
bn1  0
Similarly the second column of B  is given by
 b12  0
b  1
A 22    
   
   
bn 2  0
Similarly, all columns of B  can be found by solving n different sets of equations
with the column of the right hand side being the n columns of the identity matrix.
Example 3
Use LU decomposition to find the inverse of
 25 5 1
A   64 8 1
144 12 1
Solution
Knowing that
A  LU 
 1 0 0 25 5 1 
 2.56 1 0  0  4.8  1.56
5.76 3.5 1  0 0 0.7 
We can solve for the first column of [ B]  A by solving for
1

 25 5 1 b11  1
 64 8 1 b   0
   21   
144 12 1 b31  0
First solve
LZ   C  ,
that is
 1 0 0  z1  1
2.56 1 0  z   0
  2   
5.76 3.5 1  z 3  0
to give
z1  1
2.56 z1  z 2  0
5.76 z1  3.5z 2  z3  0
Forward substitution starting from the first equation gives
z1  1
z2  0  2.56z1
 0  2.561
 2.56
z3  0  5.76 z1  3.5z 2
 0  5.761  3.5 2.56
 3.2
Hence
 z1 
Z    z 2 
 z 3 
 1 
  2.56
 3.2 
Now solve
U X   Z 
that is
25 5 1  b11   1 
 0  4.8  1.56 b    2.56
   21   
 0 0 0.7  b31   3.2 
25b11  5b21  b31  1
 4.8b21  1.56b31  2.56
0.7b31  3.2
Backward substitution starting from the third equation gives
3.2
b31 
0.7
 4.571
 2.56  1.56b31
b21 
 4.8
 2.56  1.56(4.571)

 4.8
 0.9524
1  5b21  b31
b11 
25
1  5(0.9524)  4.571

25
 0.04762
Hence the first column of the inverse of A is
b11   0.04762 
b    0.9524
 21   
b31   4.571 
Similarly by solving
 25 5 1 b12  0 b12   0.08333
 64 8 1 b   1 gives b    1.417 
   22     22   
144 12 1 b32  0 b32    5.000 
and solving
 25 5 1 b13  0 b13   0.03571 
 64 8 1 b   0 gives b    0.4643
   23     23   
144 12 1 b33  1 b33   1.429 
Hence
 0.04762  0.08333 0.03571 
A   0.9524 1.417  0.4643
1

 4.571  5.000 1.429 

Can you confirm the following for the above example?
AA1  I   A1 A

Gauss-Seidel Method(iterative method)

Example 1

The first step is to solve the first equation for x1, the second for x2, and the
third for x3 when the system becomes:
 .

An initial solution is now assumed; we shall use

xl = 0, x2 = 0 and x3 = 0.

Inserting these values into the right-hand side of Equation (1) yields xl = 1.3.
This value for xl is used immediately together with the remainder of the initial
solution (i.e., x2 = 0 and x3 = 0) in the right-hand side of Equation (2) and
yields x2 =1.3 - 0.2 x 1.3 - 0 = 1.04. Finally, the values xl = 1.3 and x2 = 1.04
are inserted into Equation (3) to yield x3 = 0.936. This second approximate
solution (1.3, 1.04, 0.936) completes the first iteration.

Beginning with this second approximation, we repeat the process to obtain a

third approximation, etc. Under certain conditions relating to the coefficients
of the system, this sequence will converge to the exact solution.

Convergence

The sequence of solutions produced by the iterative process for the above
numerical example are shown in the table:

The student should check that the exact solution for this system is (1,1,1). It
is seen that the Gauss-Seidel solutions are rapidly approaching these values;
in other words, the method is converging.

Example 2

10x1 − x2 + 2x3 =6
− x1 + 11x2 − x3 + 3x4 = 25
2x1 − x2 + 10x3 − x4 = − 11
3x2 − x3 + 8x4 = 15

Solving for x1, x2, x3 and x4 gives:

x1 = x2 / 10 − x3 / 5 + 3 / 5
x2 = x1 / 11 + x3 / 11 − 3x4 / 11 + 25 / 11
x3 = − x1 / 5 + x2 / 10 + x4 / 10 − 11 / 10
 x4 = − 3x2 / 8 + x3 / 8 + 15 / 8

Suppose we choose (0,0,0,0) as the initial approximation, then the first approximate
solution is given by

x1 = 3 / 5 = 0.6,
x2 = (3 / 5) / 11 + 25 / 11 = 3 / 55 + 25 / 11 = 2.3272,
x3 = − (3/5)/5 + (2.3272) /10 − 1/10 = − 3 /25+0.23272−1.1= − 0.9873,
x4 = − 3(2.3272) / 8 + ( − 0.9873) / 8 + 15 / 8 = 0.8789.

Using the approximations obtained, the iterative procedure is repeated until the desired
accuracy has been reached. The following are the approximated solutions after four
iterations.

x1 x2 x3 x4

Initial
value 0.0 0.0 0.0 0.0

1 0.6 2.32727 − 0.987273 0.878864

2 1.03018 2.03694 − 1.01446 0.984341

3 1.00659 2.00356 − 1.00253 0.998351

4 1.00086 2.0003 − 1.00031 0.99985

The exact solution of the system is (1, 2, −1, 1).

Cramer’s Rule
Determinants can be used to solve a linear system of equations using Cramer’s Rule.

Cramer’s Rule for Two Equations in Two Variables

Given the system

This system has the unique solution

where

When solving a system of equations using Cramer’s Rule, remember the following:

1. Three different determinants are used to find x and y. The determinants in the
denominators are identical.

2. The elements of D, the determinant in the denominator, are the coefficients of the
variables in the system; coefficients of x in the first column and coefficients of y in the
second column.

3. , the determinant in the numerator of x, is obtained by replacing the x-

coefficients, , in D with the constants from the right sides of the equations,
.

4. , the determinant in the numerator for y, is obtained by replacing the y-

coefficients, , in D with the constants from the right side of the equation,
.

Example. Use Cramer’s Rule to solve the system:

5x – 4y = 2

6x – 5y = 1
Solution. We begin by setting up and evaluating the three determinants
:

From Cramer’s Rule, we have

The solution is (6,7).

Cramer’s Rule does not apply if D = 0. When D = 0, the system is either inconsistent or
dependent. Another method must be used to solve it.

Example. Solve the system:

3x + 6y = -1

2x + 4y = 3

Solution. We begin by finding D:

Since D = 0, Cramer’s Rule does not apply. We will use elimination to solve the
system.

3x + 6y = -1

2x + 4y = 3

2(3x + 6y) = 2(-1) Multiply both sides of equation 1

by 2 and both sides of equation 2
-3(2x + 4y) = -3(3) by –3 to eliminate x

6x + 12y = -2 Simplify

-6x – 12y = -9

0 = -11 Add the equations

The false statement, 0 = -11, indicates that the system is inconsistent and has no
solution.

Cramer’s Rule can be generalized to systems of linear equations with more than two
variables. Suppose we are given a system with the determinant of the coefficient matrix
D. Let denote the determinant of the matrix obtained by replacing the column
containing the coefficients of "n" with the constants from the right sides of the
equations. Then we have the following result:

If a linear system of equations with variables x, y, z, . . . has a unique solution given by

the formulas

Example. Use Cramer’s Rule to solve the system:

4x - y + z = -5

2x + 2y + 3z = 10

5x – 2y + 6z = 1

Solution. We begin by setting up four determinants: :

D consists of the coefficients of x, y, and z from the three equations

is obtained by replacing the x-coefficients in the first column of D with the

constants from the right sides of the equations.

is obtained by replacing the y-coefficients in the second column of D with the

constants from the right sides of the equations

is obtained by replacing the z-coefficients in the third column of D with the

constants from the right sides of the equations
Next, we evaluate the four determinants:

= 4(12 – (-6)) + 1(12 – 15) + 1(-4 – 10)

= 4(18) + 1(-3) + 1(-14)

= 72 – 3 – 14

= 55

= -5(12 – (-6)) + 1(60 – 3) + 1(-20 – 2)

= -5(18)+1(57) + 1(-22)

= -90 + 57 – 22

= -55

= 4(60 – 3) + 5(12 – 15) + 1(2 – 50)

= 4(57) + 5(-3) + 1(-48)

= 228 - 15 – 48

= 165

= 4(2 – (-20)) + 1(2 – 50) – 5(-4 – 10)

= 4(22) + 1(-48) – 5(-14)

= 88 – 48 + 70

= 110

Substitute these four values into the formula from Cramer’s Rule:

The solution is (-1, 3, 2).

Interpolation
Given the data
x 1 2 3 5 6
f (x) 4.75 4 5.25 19.75 36

Calculate f (4) using Lagrange Polynomial of order 1 through 3.

Lagrange Polynomial

First Order
( x  x1 ) ( x  x0 )
f 1 ( x)  f ( x0 )  f ( x1 )
( x0  x1 ) ( x1  x 0 )
(4  5) (4  3)
f 1 (4)  5.25  19.75  12.5
(3  5) (5  3)

Second Order
( x  x1 )( x  x 2 ) ( x  x0 )( x  x 2 ) ( x  x0 )( x  x1 )
f 2 ( x)  f ( x0 )  f ( x1 )  f ( x2 )
( x0  x1 )( x0  x 2 ) ( x1  x0 )( x1  x 2 ) ( x 2  x0 )( x 2  x1 )
(4  3)(4  5) (4  2)(4  5) (4  2)(4  3)
f 2 (4)  4.0  5.25  19.75
(2  3)(2  5) (3  2)(3  5) (5  2)(5  3)
4 19.75
f 2 (4)    5.25   10.5
3 3
Third Order
( x  x1 )( x  x2 )( x  x3 ) ( x  x0 )( x  x2 )( x  x3 )
f 3 ( x)  f ( x0 )  f ( x1 ) 
( x0  x1 )( x0  x2 )( x0  x3 ) ( x1  x0 )( x1  x2 )( x1  x3 )
( x  x0 )( x  x1 )( x  x3 ) ( x  x0 )( x  x1 )( x  x2 )
f ( x2 )  f ( x3 )
( x2  x0 )( x2  x1 )( x2  x3 ) ( x3  x0 )( x3  x1 )( x3  x2 )
(4  3)(4  5)(4  6) (4  2)(4  5)(4  6)
f 3 (4)  4.0  5.25 
(2  3)(2  5)(2  6) (3  2)(3  5)(3  6)
(4  2)(4  3)(4  6) (4  2)(4  3)(4  5)
19.75  36
(5  2)(5  3)(5  6) (6  2)(6  3)(6  5)
2 39.5
f 3 (4)    3.5   6  10
3 3
 Numerical Integration
Trapezoidal Rule
ba
I  f (a)  2 f (a1 )  2 f (a2 )  2 f (a3 )  ...  2 f (an1 )  f (b)
2n
Simpson’s Rule
ba
I  f (a)  4 f (a1 )  2 f (a2 )  4 f (a3 )  ...  2 f (an1 )  f (b)
3n
Gauss-Legendre Formula
x  a0  a1 x d
ba ba
a0  and a1 
2 2
ba
dx  dx d
2
Two Point
1 1
w0  w1  1 x0   and x1 
3 3
Three Point
5 8
w0  w2  w1  x0   0.6 x1  0 and x2  0.6
9 9
Four Point
w0  w3  0.3478548 w1  w2  0.6521452
x0  0.861136312 x1  0.339981044
x2  0.339981044 x3  0.861136312

Theorem (Gauss-Legendre Quadrature). An approximation to the integral

is obtained by sampling at the unequally spaced

abscissas , where the corresponding weights
are .
The abscissa's and weights for Gauss-Legendre quadrature are often expressed in
decimal form.

n=2 Rule
where

n=3 Rule
where
n=4 Rule
where

n=5 Rule
where

Problem
Integrate the following function both analytically and numerically. Use both the
trapezoidal and Simpson’s 1/3 rules to numerically integrate the function. For
both cases, use the multiple application version, with n = 4.
3

 xe
2x
dx
0

Compute percent relative errors for the numerical results.

Obtain estimate of integral using 2, 3, 4 point Gauss-Legendre Formula.
Analytical Solution
3
I   xe 2 x dx
0
3 3
 e 2x  3 2x
e  e 2x e 2x 
I   x    dx   x  
 2 0 0 2  2 4  0
1
 3 1
  
I  (2 x  1)e 2 x 0  5e 6  1  504.5359919
4 4

Numerical Solution Multiple application with n=4 and h=0.75

Trapezoidal Rule
I
0.75
 f (0)  2 f (0.75)  2 f (1.5)  2 f (2.25)  f (0.75)
2
  
I  0.75 0  2 0.75e1.5  1.5e 3  2.25e 4.5  3e 6 
I  630.8784809
 504.5359919  630.8784809
Error   100  25.04132333
504.5359919
Simpson’s Rule
I
3
 f (0)  4 f (0.75)  2 f (1.5)  4 f (2.25)  f (0.75)
12
  
I  0.25 0  4 0.75e1.5  2.25e 4.5  2  1.5e 3  3e 6 
I  523.53556
504.5359919  523.53556
Error   100  3.765750797
504.5359919
Gauss-Legendre Formula
x  a 0  a1 x d
b a 30 ba 30
a0    1.5 a1    1.5
2 2 2 2
x  1.5  1.5 x d dx  1.5dx d
1
I   (1.5  1.5 x d )e (33 xd ) 1.5dx d
1
1
I  2.25  (1  x d )e 3(1 xd ) dx d
1

Two Point
1 1
c 0  c1  1 x0   and x1 
3 3
  1 
1  3 1 3   1  3 1 3  
 1 

I  2.25 1  e  1  e  406.2950216
 3   3  
 
504.5359919  406.2950216
Error   100  19.47154848
504.5359919
Three Point
5 8
c0  c 2  c1  x 0   0.6 x1  0 and x 2  0.6
9 9
5
   

I  2.25 1  0.6 e 31 0.6   e 3  1  0.6 e 31 0.6    495.8203637
8 5
9 9 9 
504.5359919  495.8203637
Error   100  1.727454197
504.5359919
Four Point
c0  c3  0.3478548 c1  c2  0.6521452
x0  0.861136312 x1  0.339981044
x2  0.339981044 x3  0.861136312

I  2.25 c0 1  x0 e 31 x0   c1 1  x1 e 31 x1   c2 1  x2 e 31 x2   c3 1  x3 e 31 x3  
I  504.1305
504.5359919  504.1305
Error   100  0.080369264
504.5359919
 Ordinary Differential Equation

A solution of an Ordinary Differential Equation (ODE) is a specific

function of the independent variable and parameter that satisfies the
original differential equation. To specify the solution completely a
differential equation is usually accompanied by auxiliary conditions. For
the First order ODEs, a type of auxiliary condition called an initial value is
required to determine the constant and obtain a unique solution. For
example, the initial condition could be value of y at x equal to zero.

 f  x, y 
dy
dx

x is the independent variable

y is the dependent variable

When the function involves one independent variable the equation is called ODE
When it involves two or more independent variable the equation is called PDE

Differential equation are also classified as to their order. The above equation is called
a first order equation because the highest derivative is the first derivative.

One simple approach to solve the given differential equation by numerical technique is
to predict new value based on slope f x, y  , old value and step size.

New Value  Old Value  Slope  Step Size

In mathematical form

yi 1  yi   h where   f x, y   Slope

This formula can be applied step by step to compute the trajectory of the solution.

Euler’s Method

The first derivative provides a direct estimate of the slope at xi , yi  .

  f xi , yi 
where, f xi , yi  is the differential equation evaluated at xi , yi  .
yi 1  yi   h where   f xi , yi   Slope

Problem 1
dy
 f ( x, y)  2 x 3  12 x 2  20 x  8.5, y(0)  1 Compute y(o.5)
dx
Analytical Solution y  0.5x 4  4 x 3  10 x 2  8.5x  1

xi  0 yi  1 h  0.5
  f xi , yi   f 0,1  8.5
y 0.5  1  8.5  0.5  5.25
  f xi , yi   f 0.5,5.25  1.25
y1  5.25  1.25  0.5  5.875

X f xi , yi  Yeuler Ytrue

0.0 8.5 1.00000 1.00000
0.5 1.25 5.25000 3.21875
1.0 -1.5 5.87500 3.00000
1.5 -1.25 5.12500 2.21875
2.0 0.5 4.50000 2.00000

Classical Fourth- order Runge-Kutta method

1
yi 1  yi  (k1  2k 2  2k 3  k 4 ) h
6
1
yi 1  yi   h where   (k1  2k 2  2k 3  k 4 )
6
k1  f ( x i , y i )
h 1
k 2  f ( xi  , y i  k 1 h)
2 2
h 1
k 3  f ( x i  , y i  k 2 h)
2 2
k 4  f ( x i  h, y i  k 3 h)

Problem 2
dy
 (1  x) y , y (0)  1
dx
Analytical Solution y
16

1 2
x  2x  4 
2

xi  0 yi  1 h  0.5
k1  f (0,1)  1
This value is used to calculate y and slope at midpoint.
1
y 0.25  1   1  0.5  1.25
2
k 2  f (0.25,1.25)  (1  0.25) 1.25  1.397542486
This slope in turn is used to compute another value of y and another slope at
midpoint.
1
y 0.25  1   1.397542486  0.5  1.349385621
2
k 3  f (0.25,1.349385621)  (1  0.25) 1.349385621  1.452038234
This slope is used to compute the value of y and slope at the end of interval.
y 0.5  1  1.452038234  0.5  1.726019117
k 4  f (0.5,1.726019117)  (1  0.25) 1.726019117  1.970670701
Finally four slope estimates are combined to yield an average slope. This
average slope is used to make the final prediction at the end of interval.
 1
  (k1  2k 2  2k 3  k 4 )
6
1
  (1  2(1.397542486  1.452038234)  1.970670701)  1.444972024
6

y0.5  1  1.444972024  0.5  1.722486012

Other values are computed in similar approach and are reported in following
Table.

X Y k1 k2 k3 k4 Y-Analytical Error
0 1 1 1.397542 1.452038 1.970671 1
0.5 1.722486 1.968653 2.604297 2.696114 3.504593 1.72265625 0.009882
1 3.061992 3.499709 4.464376 4.599082 5.788746 3.0625 0.016601
1.5 5.346606 5.780682 7.166785 7.34734 9.010132 5.34765625 0.019641
2 8.998195 8.999097 10.89985 11.12771 13.35609 9 0.020059
Problem 3
dy
 f ( x, y )  4e0.8 x  0.5 y,
dx
y (0)  2 Compute y (o.5)

Analytical Solution y 
1
13

40e 0.8 x  14e 0.5 x 
xi  0 yi  2 h  0.5
k1  f (0,2)  4  1  0.5  2  3
This value is used to calculate y and slope at midpoint.
1
y0.25  2   3  0.5  2.75
2
k2  f (0.25,2.75)
k2  4  e0.80.25  0.5  2.75  3.510611033
This slope in turn is used to compute another value of y and another slope at midpoint.
1
y0.25  2   3.510611033  0.5  2.877652758
2
k3  f (0.25,2.877652758)
k3  4  e0.80.25  0.5  2.877652758  3.446784654

This slope is used to compute the value of y and slope at the end of interval.
y0.5  2  3.446784654  0.5  3.723392327
k4  f (0.5,3.723392327)
k4  4  e0.80.5  0.5  3.723392327
k4  4.105602627

Finally four slope estimates are combined to yield an average slope. This average slope is used
to make the final prediction at the end of interval.
1
  (k1  2k 2  2k 3  k 4 )
6
1
  (3  2(3.510611033  3.446784654)  4.105602627)  3.503399
6

y0.5  2  3.503399  0.5  3.7516995

True value y0.5  3.751521303
Problem 4
dy
 f ( x, y )  2 x3  12 x 2  20 x  8.5,
dx
y (0)  1 Compute y (o.5)
Analytical Solution y  0.5x 4  4 x 3  10 x 2  8.5x  1
xi  0 yi  1 h  0.5
k1  f ( xi , yi )  f (0,1)  8.5
h 1 1
k 2  f ( xi  , yi  k1h)  f (0.25, yi  k1h)  4.21875
2 2 2
h 1 1
k 3  f ( xi  , yi  k 2 h)  f (0.25, yi  k 2 h)  4.21875
2 2 2
k 4  f ( xi  h, yi  k 3 h)  f (0.5, yi  k 3 h)  1.25
Finally four slope estimates are combined to yield an average slope. This average slope is used
to make the final prediction at the end of interval.
1
  (k1  2k 2  2k 3  k 4 )
6
1
  (8.5  2(4.21875  4.21875)  1.25)  4.4375
6

y0.5  1  4.4375  0.5  3.21875