Numerical Root-Finding:

Bracketing Methods

Non-linear Equations
What are Nonlinear Equations?
Linear Equations

Non-linear Equations

10 + 5 = 15

2 + 2 = 2

++1=0

= 0 cos 2
1 1 1
= +

= 2 3

ln =
+
= 0 (1 )

=
=
=
0
0 =

Root-finding
What is Root-finding?
Example: Find the roots of 2 + 6 + 5 = 0
Example: Find the roots of 2 2 + 2 = 0.
Note:
Solve for x:
2 = ln

is the
same as

Find the roots of:

Outline
Numerical Root-Finding Methods
1. Bisection Method
2. Regula-Falsi Method
Example: Solve for x in
2 + 2 = 2

Non-linear Equations
Example: Solve for x in 2 + 2 = 2
Lets do the Bisection Method!
1. Turn your equation into a function in x.

= 2 2 2
*Again, solving for x
means finding the root
of f(x), i.e. f(x) = 0.

Bisection Method
2. Plot . Find an interval [a, b] where 1 root exists.

Choose the
interval such that
is monotonic in
[a, b] and
< 0.

NO GOOD!

Lets say: [1, 3]

=
=

Bisection Method
2. Plot . Find an interval [a, b] where 1 root exists.

Choose the
interval such that
is monotonic in
[a, b] and
< 0.

=
Exactly 1
root exists!

Lets say: [4, 5]

=
=

Bisection Method
3. Find the midpoint of [a, b] and label it as c. Evaluate
, and ().
Bounds: 4, 5
=

=4

=5
+
=
2
=

. = .

= .

Bisection Method
4. If < 0, then set the new bounds to [a, c].
If < 0, then set the new bounds to [c, b].
=

= .

= 1.887

. = .

Bisection Method
Repeat step 3. Find the midpoint of [a, b] and label it
as c. Evaluate , and ().
Bounds: 4, 4.5
=4
= 4.5
=

. = .

. = .

= .

Bisection Method
4. If < 0, then set the new bounds to [a, c].
If < 0, then set the new bounds to [c, b].

= 2.070

= .
=

. = .

. = .

Bisection Method
How do you know when to stop?

Ans. Set a tolerance!

Bisection Method
5. If the interval is less than a tolerance, then STOP.
Report = as your answer. Otherwise, go to step 3.
Tolerance: = 0.01
Previous c:
4.25
Current c:
4.375
Is 4.375 4.25 <

No! Repeat step 3.

. = .

. = .

Bisection Method
Lets look at what really happens

= 2 2 2

Bisection Method
Lets look at what really happens = 2 2 2
Iter

f(a)

f(b)

f(c)

4.5

-2

0.377417

4.5

4.25

-2

0.377417

-1.03519

4.25

4.5

4.375

-1.03519

0.377417

-0.39119

4.375

4.5

4.4375

-0.39119

0.377417

-0.02332

4.4375

4.5

4.46875

-0.02332

0.377417

0.172832

4.4375

4.46875

4.453125

-0.02332

0.172832

0.073717

4.4375

4.453125

4.445313

-0.02332

0.073717

0.024941

At Iter = 6, since 4.453125 4.445313 < 0.01, then we stop!

Bisection Method
Lets look at what really happens = 2 2 2
Iter

f(A)

f(B)

f(C)

14

4.441284

4.441345

4.441314697

-6.6E-06

0.00037

0.000182

15

4.441284

4.441315

4.441299438

-6.6E-06

0.000182

8.77E-05

16

4.441284

4.441299

4.441291809

-6.6E-06

8.77E-05

4.06E-05

17

4.441284

4.441292

4.441287994

-6.6E-06

4.06E-05

1.7E-05

18

4.441284

4.441288

4.441286087

-6.6E-06

1.7E-05

5.23E-06

19

4.441284

4.441286

4.441285133

-6.6E-06

5.23E-06

-6.6E-07

For a tolerance of 0.000001

Bisection Method
Summary of steps:
1. Turn your equation into a function of x. Set a tolerance value
(typically 0.0001).
2. Plot . Find an interval [a, b] where 1 root exists.
3. Find the midpoint of [a, b] and label it as c. Evaluate
, and ().
4. If < 0, then set the new bounds to [a, c]. If
< 0, then set the new bounds to [c, b].
5. If < , then STOP. Report = as your
answer. Otherwise, go to step 3.

Bisection Method

Alternatively,
instead of
having a
tolerance, you
can also set the
maximum
number of
iterations.
N=4

Bisection Method
YOUR TASK:
Develop a program that carries out the bisection
method. Prompt the user for the initial guess
interval [a, b], and the desired tolerance.
Output the final root, and the number of
iterations needed to reach that final value.

Regula-Falsi Method
Example: Solve for x in 2 + 2 = 2
Lets do the Regula-falsi Method!
1. Turn your equation into a function in x.

= 2 2 2
*Again, solving for x
means finding the root
of f(x), i.e. f(x) = 0.

Regula-Falsi Method
2. Plot . Find an interval [a, b] where 1 root exists.

Choose the
interval such that
is monotonic in
[a, b] and
< 0.
From before:
[a, b] = [4, 5]

=
Exactly 1
root exists!

Regula-Falsi Method
3. Find c using the following formula. Evaluate ().
+
=
2

()( )
=
()

. = .

(5)(5 4)
=5
5 (4)
5 30
=5 =
= 4.286
7
7
=

Regula-Falsi Method
Why =

()()
()

?
From similar triangles:

bc
f(b)

()

=
()
f(b) f(a)

ba
a c

Regula-Falsi Method
4. If < 0, then set the new bounds to [a, c].
If < 0, then set the new bounds to [c, b].
=

= 1.726

= .

. = .

Regula-Falsi Method
Summary of steps:
1. Turn your equation into a function of x. Set a tolerance value
(typically 0.0001).
2. Plot . Find an interval [a, b] where 1 root exists.
3. Find =

()()
. Evaluate
()

().

4. If < 0, then set the new bounds to [a, c]. If

Regula-Falsi Method
YOUR TASK:
Develop a program that carries out the regula-falsi
(false-position) method. Prompt the user for the
initial guess [a, b], and the desired tolerance.
Output the final root, and the number of
iterations needed to reach that final value.

Bracketing Methods
Problems in the
Bracketing Methods:

- Knowing where to start

Practice
Solve for the only positive value of x that satisfies:

1
x
x 1 1 e
2
x

Use the following methods. Which converges fastest?

1. Bisection Method
2. Regula-Falsi Method
Use a tolerance of 0.00001.