Scribd
Upload
156 views12 pages

2.3 Newton's Method and Its Extension

Newton's method is an iterative method for finding the roots of a function. It uses the tangent line of the function at the current point to get closer to the root during each iteration. The secant method approximates the derivative using two previous points. The method of false position is similar to the secant method but ensures the root remains bracketed between iterations.

Uploaded by

Lucas Santos
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
156 views12 pages

2.3 Newton's Method and Its Extension

Newton's method is an iterative method for finding the roots of a function. It uses the tangent line of the function at the current point to get closer to the root during each iteration. The secant method approximates the derivative using two previous points. The method of false position is similar to the secant method but ensures the root remains bracketed between iterations.

Uploaded by

Lucas Santos
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

2.

3 Newton’s Method and Its


Extension

1
Basic Idea
• Taylor’s Theorem Recap
Suppose 𝑓𝑓 ∈ 𝐶𝐶 2 [𝑎𝑎, 𝑏𝑏] and 𝑝𝑝0 ∈ [𝑎𝑎, 𝑏𝑏] approximates solution 𝑝𝑝
of 𝑓𝑓 𝑥𝑥 = 0 with 𝑓𝑓𝑓(𝑝𝑝0 ) ≠ 0. Expand 𝑓𝑓(𝑥𝑥) about 𝑝𝑝0 :
(𝑝𝑝 − 𝑝𝑝 )2
′ 0
𝑓𝑓 𝑝𝑝 = 𝑓𝑓 𝑝𝑝0 + 𝑝𝑝 − 𝑝𝑝0 𝑓𝑓 𝑝𝑝0 + 𝑓𝑓 ′′ 𝜉𝜉 𝑝𝑝
2
𝑓𝑓 𝑝𝑝 = 0, and assume (𝑝𝑝 − 𝑝𝑝0 )2 is negligible:
0 ≈ 𝑓𝑓 𝑝𝑝0 + 𝑝𝑝 − 𝑝𝑝0 𝑓𝑓 ′ 𝑝𝑝0
Solving for 𝑝𝑝 yields:
𝑓𝑓(𝑝𝑝0 )
𝑝𝑝 ≈ 𝑝𝑝1 ≡ 𝑝𝑝0 −
𝑓𝑓𝑓(𝑝𝑝0 )
This gives the sequence {𝑝𝑝𝑛𝑛 }∞ 𝑛𝑛=0 :
𝑓𝑓(𝑝𝑝𝑛𝑛−1 )
𝑝𝑝𝑛𝑛 = 𝑝𝑝𝑛𝑛−1 −
𝑓𝑓𝑓(𝑝𝑝𝑛𝑛−1 )
Remark: 𝑝𝑝𝑛𝑛 is an improved approximation. 2
Algorithm: Newton’s Method

INPUT initial approximation p0; tolerance TOL; maximum number of


iterations N0.
OUTPUT approximate solution p or message of failure.
STEP1 Set i = 1.
STEP2 While i ≤ N0 do STEPs 3-6
STEP3 Set p = p0 – f(p0)/f’(p0).
STEP4 If |p-p0| < TOL then
OUTPUT (p);
STOP.
STEP5 Set i = i + 1.
STEP6 Set p0 = p.
STEP7 OUTPUT(‘The method failed’);
STOP. 3
Geometric Interpretation
• Two steps of Newton’s method for solving 𝑓𝑓 𝑥𝑥 = 𝑥𝑥 3 + 4𝑥𝑥 2 −
10 = 0.

• Choose 𝑝𝑝0 = 1
𝑝𝑝03 +4𝑝𝑝02 −10
• 𝑝𝑝1 = 𝑝𝑝0 − = 1.4545454545
3𝑝𝑝02 +8𝑝𝑝0
𝑝𝑝13 +4𝑝𝑝12 −10
• 𝑝𝑝2 = 𝑝𝑝1 − = 1.3689004011
3𝑝𝑝12 +8𝑝𝑝1
4
Example 2.3.1 Consider the function 𝑓𝑓 𝑥𝑥 =
𝜋𝜋
cos 𝑥𝑥 − 𝑥𝑥 = 0, with 𝑥𝑥 ∈0, . Approximate a
2
root of 𝑓𝑓 using
(a) a fixed-point method, and
(b) Newton’s method.

5
About Newton’s Method
• Pros.
1. Fast convergence: Newton’s method converges fastest
among methods we explore (quadratic convergence)
• Cons.
1. 𝑓𝑓 ′ 𝑥𝑥𝑛𝑛−1 cause problems
Remark: Newton’s method works best if |𝑓𝑓 ′ | ≥ 𝑘𝑘 > 0
2. Expensive: Computing derivative in every iteration
• We assume |𝑝𝑝 − 𝑝𝑝0 | is small, then |𝑝𝑝 − 𝑝𝑝0 |2 ≪
|𝑝𝑝 − 𝑝𝑝0 |, and we can neglect the 2nd order term in
Taylor expansion.
Remark: In order for Newton’s method to converge we need
a good starting guess.
6
Relation to fixed-point iteration
Newton’s method is fixed-point iteration with
𝑓𝑓(𝑥𝑥)
𝑔𝑔 𝑥𝑥 = 𝑥𝑥 −
𝑓𝑓𝑓(𝑥𝑥)

Convergence
Theorem 2.6
Let 𝑓𝑓 ∈ 𝐶𝐶 2 [𝑎𝑎, 𝑏𝑏] and 𝑝𝑝 ∈ [𝑎𝑎, 𝑏𝑏] is that 𝑓𝑓 𝑝𝑝 = 0 and
𝑓𝑓𝑓(𝑝𝑝) ≠ 0, then there exists a 𝛿𝛿 > 0 such that
Newton’s method generates a sequence {𝑝𝑝𝑛𝑛 }∞ 𝑛𝑛=1
converging to 𝑝𝑝 for any initial approximation 𝑝𝑝0 ∈
𝑝𝑝 − 𝛿𝛿, 𝑝𝑝 + 𝛿𝛿 . 7
The Secant Method
• Approximate the derivative:
𝑓𝑓 𝑝𝑝𝑛𝑛−1 − 𝑓𝑓(𝑝𝑝𝑛𝑛−2 )
𝑓𝑓𝑓(𝑝𝑝𝑛𝑛−1 ) ≈
𝑝𝑝𝑛𝑛−1 − 𝑝𝑝𝑛𝑛−2
to get
𝑓𝑓 𝑝𝑝𝑛𝑛−1 𝑝𝑝𝑛𝑛−1 − 𝑝𝑝𝑛𝑛−2
𝑝𝑝𝑛𝑛 = 𝑝𝑝𝑛𝑛−1 − (2.12)
𝑓𝑓 𝑝𝑝𝑛𝑛−1 − 𝑓𝑓 𝑝𝑝𝑛𝑛−2

8
Summary of Secant method algorithm:
1. Make two initial guesses: 𝑝𝑝0 and 𝑝𝑝1
2. Use Eq. (2.12) to construct 𝑝𝑝2 , 𝑝𝑝3 , 𝑝𝑝4 … till
accuracy is met.

Exercise 2.3.6 Consider the function 𝑓𝑓 𝑥𝑥 =


𝑒𝑒 𝑥𝑥 + 2−𝑥𝑥 + 2𝑐𝑐𝑐𝑐𝑐𝑐 𝑥𝑥 − 6. Solve 𝑓𝑓 𝑥𝑥 = 0 using the
Secant method for 1 ≤ 𝑥𝑥 ≤ 2.

9
Algorithm: The Secant Method
INPUT initial approximation p0, p1; tolerance TOL; maximum number of
iterations N0.
OUTPUT approximate solution p or message of failure.
STEP1 Set i = 2;
q0 = f(p0);
q1 = f(p1);
STEP2 While i ≤ N0 do STEPs 3-6
STEP3 Set p = p1 –q1(p1-p0)/(q1-q0).
STEP4 If |p-p1| < TOL then
OUTPUT (p);
STOP.
STEP5 Set i = i + 1.
STEP6 Set p0 = p1;
q0 = q1;
p1 = p;
q1 = f(p).
STEP7 OUTPUT(‘The method failed’);
STOP. 10
The Method of False Position
• The bisection method iterations satisfy:
1
𝑝𝑝𝑛𝑛 − 𝑝𝑝 < 𝑎𝑎𝑛𝑛 − 𝑏𝑏𝑛𝑛 , which means the root lies
2
between 𝑎𝑎𝑛𝑛 and 𝑏𝑏𝑛𝑛 .
• Root bracketing is not guaranteed for either
Newton’s method or Secant method.
• Method of false position: generate
approximations in the same manner as the
Secant method, but also includes a test to ensure
that the root is always bracketed between
successive iterations.

11
• Start with two points 𝑎𝑎𝑛𝑛 , 𝑏𝑏𝑛𝑛 which bracket the
root, i.e, 𝑓𝑓 𝑎𝑎𝑛𝑛 � 𝑓𝑓 𝑏𝑏𝑛𝑛 < 0. Let 𝑝𝑝𝑛𝑛+1 be the zero-
crossing of the secant line:
𝑓𝑓 𝑏𝑏𝑛𝑛 (𝑎𝑎𝑛𝑛 − 𝑏𝑏𝑛𝑛 )
𝑝𝑝𝑛𝑛+1 = 𝑏𝑏𝑛𝑛 −
𝑓𝑓 𝑎𝑎𝑛𝑛 − 𝑓𝑓(𝑏𝑏𝑛𝑛 )
• Update as in the bisection method:
If 𝑓𝑓 𝑎𝑎𝑛𝑛 � 𝑓𝑓 𝑝𝑝𝑛𝑛+1 > 0, then 𝑎𝑎𝑛𝑛+1 = 𝑝𝑝𝑛𝑛+1 , 𝑏𝑏𝑛𝑛+1 = 𝑏𝑏𝑛𝑛
If 𝑓𝑓 𝑎𝑎𝑛𝑛 � 𝑓𝑓 𝑝𝑝𝑛𝑛+1 < 0, then 𝑎𝑎𝑛𝑛+1 = 𝑎𝑎𝑛𝑛 , 𝑏𝑏𝑛𝑛+1 = 𝑝𝑝𝑛𝑛+1

12

You might also like