Vous êtes sur la page 1sur 5

Regula Falsi (False-Position) Method

Introduction
The bisection method was described as one of the simple bracketing methods of solving
a nonlinear equation of the general form
0 ) ( = x f
(1)
Figure 1 False-Position ethod
The above nonlinear equation can be stated as finding the value of x such that !quation
(1) is satisfied"
In the bisection method# we identif$ proper values of
L
x (lower bound value) and U
x
(upper bound value) for the current bracket# such that

0 ) ( ) ( <
U L
x f x f
" (%)
The ne&t predicted'improved root
r
x can be computed as the midpoint between
L
x
and U
x
as
2
U L
r
x x
x
+
= (()
The new upper and lower bounds are then established# and the procedure is repeated
until the convergence is achieved (such that the new lower and upper bounds are
sufficientl$ close to each other)"
)owever# in the e&ample shown in Figure 1# the bisection method ma$ not be efficient
because it does not take into consideration that ) (
L
x f is much closer to the *ero of the
function
) (x f
as compared to
) (
U
x f
" In other words# the ne&t predicted root
r
x

would be closer to
L
x (in the e&ample as shown in Figure 1)# than the mid-point
between
L
x and U
x
" The false-position method takes advantage of this observation
mathematicall$ b$ drawing a secant from the function value at
L
x to the function
value at U
x
# and estimates the root as where it crosses the x-a&is"
False-Position ethod
+ased on two similar triangles# shown in Figure 1# one gets
U r
U
L r
L
x x
x f
x x
x f

) ( 0 ) ( 0
(,)
From !quation (,)# one obtains
( ) ( ) ( ) ( )
L U r U L r
x f x x x f x x =
( ) ( ) ( ) ( ) { }
U L r U L L U
x f x f x x f x x f x =
The above equation can be solved to obtain the ne&t predicted root m
x
as
( ) ( )
( ) ( )
U L
U L L U
r
x f x f
x f x x f x
x

=
" (-)
./
( ) ( )
( ) ( ) b f a f
b af a bf
x

=

False-Position 0lgorithm
The steps to appl$ the false-position method to find the root of the equation
( ) 0 = x f
are
as follows"
1" 1hoose
L
x and
U
x
as two guesses for the root such that
( ) ( ) 0 <
U L
x f x f
# or in other
words#
( ) x f
changes sign between
L
x and U
x
"
%" !stimate the root#
r
x of the equation
( ) 0 = x f
as
( ) ( )
( ) ( )
U L
U L L U
r
x f x f
x f x x f x
x

=
(" 2ow check the following
If ( ) ( ) 0 <
r L
x f x f # then the root lies between
L
x and
r
x 3 then
L L
x x = and r U
x x =
"
If ( ) ( ) 0 >
r L
x f x f # then the root lies between
r
x and U
x
3 then
r L
x x = and U U
x x =
"
If ( ) ( ) 0 =
r L
x f x f # then the root is
r
x " 4top the algorithm"
," Find the new estimate of the root
( ) ( )
( ) ( )
U L
U L L U
r
x f x f
x f x x f x
x

=
Find the absolute relative appro&imate error as
100

=
new
r
old
r
new
r
a
x
x x
where
new
r
x 5 estimated root from present iteration
old
r
x 5 estimated root from previous iteration
!&ample 1
Find the root between 6 and 6"11 for the equation 0 10 993 . 3 165 . 0
4 2 3
= +

x x b$
false-position method "
4olution
7et us assume
11 . 0 , 0 = =
U L
x x
1heck if the function changes sign between
L
x and
U
x
( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
4 4 2 3
4 4 2 3
10 662 . 2 10 993 . 3 11 . 0 165 . 0 11 . 0 11 . 0
10 993 . 3 10 993 . 3 0 165 . 0 0 0


= + = =
= + = =
f x f
f x f
U
L
)ence
( ) ( ) ( ) ( ) ( ) ( ) 0 10 662 . 2 10 993 . 3 11 . 0 0
4 4
< = =

f f x f x f
U L
Therefore# there is at least one root between
L
x and U
x
# that is between 6 and 6"11"
Iteration 1
The estimate of the root is
( ) ( )
( ) ( )
( )
( )
0660 . 0
10 662 . 2 10 993 . 3
10 662 . 2 0 10 993 . 3 11 . 0
4 4
4 4
=


=

=


U L
U L L U
r
x f x f
x f x x f x
x
( ) ( )
( ) ( ) ( )
5
4 2 3
10 1944 . 3
10 993 . 3 0660 . 0 165 . 0 0660 . 0
0660 . 0

=
+ =
= f x f
r
( ) ( ) ( ) ( ) ( ) ( ) 0 0660 . 0 0 < + = = f f x f x f
r L
)ence# the root is bracketed between
L
x and
r
x # that is# between 6 and 6"6886" 4o# the
lower and upper limits of the new bracket are
0660 . 0 , 0 = =
U L
x x
# respectivel$"
Iteration %
The estimate of the root is
( ) ( )
( ) ( )
( )
( )
0611 . 0
10 1944 . 3 10 993 . 3
10 1944 . 3 0 10 993 . 3 0660 . 0
5 4
5 4
=


=

=


U L
U L L U
r
x f x f
x f x x f x
x
The absolute relative appro&imate error for this iteration is
% 8 100
0611 . 0
0660 . 0 0611 . 0

=
a

( ) ( )
( ) ( ) ( )
5
4 2 3
10 1320 . 1
10 993 . 3 0611 . 0 165 . 0 0611 . 0
0611 . 0

=
+ =
= f x f
r
( ) ( ) ( ) ( ) ( ) ( ) 0 0611 . 0 0 > + + = = f f x f x f
r L
)ence# the lower and upper limits of the new bracket are
0660 . 0 , 0611 . 0 = =
U L
x x
#
respectivel$"
Iteration (
The estimate of the root is
( ) ( )
( ) ( )
( )
( )
0624 . 0
10 1944 . 3 10 132 . 1
10 1944 . 3 0611 . 0 10 132 . 1 0660 . 0
5 5
5 5
=


=

=


U L
U L L U
r
x f x f
x f x x f x
x
The absolute relative appro&imate error for this iteration is
% 05 . 2 100
0624 . 0
0611 . 0 0624 . 0

=
a
( )
7
10 1313 . 1

=
r
x f
( ) ( ) ( ) ( ) ( ) ( ) 0 0624 . 0 0611 . 0 < + = = f f x f x f
r L
)ence# the lower and upper limits of the new bracket are
0624 . 0 , 0611 . 0 = =
U L
x x
0ll iterations results are summari*ed in Table 1"
Table 1 /oot of ( ) 0 10 993 . 3 165 . 0
4 2 3
= + =

x x x f for false-position method"
Iteration
L
x
U
x
r
x %
a

( )
m
x f
1 6"6666 6"1166 6"6886 ----
5
10 1944 . 3


% 6"6666 6"6886 6"6811 9"66
5
10 1320 . 1


( 6"6811 6"6886 6"68%, %"6-
7
10 1313 . 1


!&ample %
Find the root of ( ) ( ) ( ) 0 2 4
2
= + = x x x f # using the initial guesses of 5 . 2 =
L
x and
, 0 . 1 =
U
x
and a pre-specified tolerance of
% 1 . 0 =
s
"
4olution
/oot of ( ) ( ) ( ) 0 2 4
2
= + = x x x f for false-position method"
Iteration
L
x
U
x
( )
L
x f ( )
U
x f
r
x %
a

( )
m
x f
1 -%"- -1 -%1"1( %-"66 -1"91( 2'0 8"(1:
% -%"- -1"91( -%1"1( 8"(1: -1":;1 9"6%, 1"6%9
( -%"- -1":;1 -%1"1( 1"6%9 -1"::8 1"%%: 6"1-,%
, -%"- -1"::8 -%1"1( 6"1-,% -1"::: 6"19%9 6"6%%98
- -%"- -1"::: -%1"1( 6"6%%98 -%"666 6"6%;68 6"66((9(
)ence the root of the equation is -%"666

Vous aimerez peut-être aussi