Newton–Raphson Method

Introduction

Definition
The Newton–Raphson Method is an open method used to solve nonlinear equations of the
form
f (x) = 0
by using the tangent line at a point on the curve to generate successive approximations of the
root.

Key Idea

Unlike bracketing methods, Newton–Raphson does not require an interval containing the root. It
starts with a single initial guess and improves that guess iteratively using the derivative of the
function.

Open Method Concept

Why it is called an open method

Bracketing methods always keep the root within an interval and therefore are generally convergent.
In contrast, open methods use one or two starting values that do not necessarily bracket the root.
So they may sometimes diverge, but when they converge, they are usually faster than bracketing
methods.

Basic Formula

Newton–Raphson Formula

f (xi )
xi+1 = xi −
f ′ (xi )

Meaning
Each new estimate xi+1 is obtained from the current estimate xi by subtracting the ratio of the
function value to its derivative at xi .

Geometrical Interpretation

Geometric Meaning

At the point xi , draw the tangent to the curve y = f (x). The x-coordinate where this tangent
line cuts the x-axis becomes the next estimate xi+1 .

1
 f (x)

x
xi xi+1

Figure 1: Tangent line interpretation of the Newton–Raphson method.

Derivation

Derivation using tangent slope

From the tangent geometry,

AB
tan(α) =
AC
and the slope of the tangent at xi is

f (xi )
f ′ (xi ) = .
xi − xi+1

Now rearranging,
f (xi )
xi − xi+1 =
f ′ (xi )
so,
f (xi )
xi+1 = xi −
f ′ (xi )

2
Algorithm

Step-by-step Procedure

1. Choose an initial guess x0 .

2. Compute the derivative f ′ (x) either mathematically or symbolically.

3. Use the iterative formula

f (xi )
xi+1 = xi − .
f ′ (xi )

4. Compute the absolute relative approximate error:

xi+1 − xi
εa = × 100%.
xi+1

5. Compare |εa | with the specified tolerance εs .

6. If the tolerance is satisfied, stop; otherwise, repeat using the new estimate.

Error Formula

Absolute Relative Approximate Error

xi+1 − xi
εa = × 100%
xi+1

Stopping Criterion

The iterations are terminated when

|εa | ≤ εs
or when the maximum number of iterations is reached.

Illustrative Example

Example Function

Given
f (x) = e−x − x

Step 1: Differentiate

d −x
f ′ (x) = (e − x) = −e−x − 1
dx

3
 Step 2: Use Newton–Raphson Formula

Starting from an initial guess xi , compute

f (xi )
xi+1 = xi − .
f ′ (xi )

Worked Example: Floating Ball Problem

Problem Statement
A floating ball has specific gravity 0.6 and radius 5.5 cm. The equation giving the submerged
depth x in meters is
f (x) = x3 − 0.165x2 + 3.993 × 10−4 .
Use the Newton–Raphson method to estimate the root.

Step 1: Derivative

f (x) = x3 − 0.165x2 + 3.993 × 10−4

f ′ (x) = 3x2 − 0.33x

Step 2: Initial Guess

A reasonable initial guess is

x0 = 0.05 m
because the depth must lie between 0 and the diameter of the ball, 0.11 m.

Iteration 1

f (x0 )
x1 = x0 −
f ′ (x0 )

(0.05)3 − 0.165(0.05)2 + 3.993 × 10−4

x1 = 0.05 −
3(0.05)2 − 0.33(0.05)

1.118 × 10−4
x1 = 0.05 − = 0.06242
−9 × 10−3

0.06242 − 0.05
εa = × 100 = 19.90%
0.06242

Iteration 2

f (x1 )
x2 = x1 −
f ′ (x1 )

(0.06242)3 − 0.165(0.06242)2 + 3.993 × 10−4

x2 = 0.06242 −
3(0.06242)2 − 0.33(0.06242)

−3.97781 × 10−7
x2 = 0.06242 − = 0.06238
−8.90973 × 10−3

4
 0.06238 − 0.06242
εa = × 100 = 0.0716%
0.06238

Iteration 3

f (x2 )
x3 = x2 −
f ′ (x2 )

(0.06238)3 − 0.165(0.06238)2 + 3.993 × 10−4

x3 = 0.06238 −
3(0.06238)2 − 0.33(0.06238)

x3 ≈ 0.06238

0.06238 − 0.06238
εa = × 100 = 0%
0.06238

Final Result
The submerged depth is approximately

x ≈ 0.06238 m

Iteration Summary

Iteration Estimate Approximate Significant Digits Correct

Error
1 x1 = 0.06242 19.90% 0
2 x2 = 0.06238 0.0716% 2
3 x3 = 0.06238 0% 4

Advantages of Newton–Raphson Method

Advantages

1. It converges very fast near the root.

2. It usually requires fewer iterations than bisection or false position.

3. It has very high accuracy when the initial guess is good.

4. It has quadratic convergence.

Drawbacks of Newton–Raphson Method

1. Divergence Near an Inflection Point

Problem
If the initial guess or an iterated value is close to an inflection point where the derivative becomes
very small, the method may diverge away from the root before eventually coming back.

5
 Example

For an equation like

f (x) = (x − 1)3 + 0.512 = 0,
starting from x0 = 5.0, the iterates may first move away from the root because the estimate gets
close to the inflection point x = 1, where f ′ (x) = 0.

2. Division by Zero or Near Zero

Problem
If
f ′ (xi ) = 0,
the Newton–Raphson formula becomes undefined.
Even if f ′ (xi ) is very close to zero, the correction term

f (xi )
f ′ (xi )

becomes extremely large, causing unstable jumps and possible divergence.

3. Oscillation Near Local Maximum or Minimum

Problem
The method may oscillate around a local maximum or minimum rather than converging to a root.

Example

For
f (x) = x2 + 2,
there is no real root because
x2 + 2 = 0 =⇒ x2 = −2.
Also,
f ′ (x) = 2x
which becomes zero at x = 0, the local minimum. As a result, the method oscillates and may
eventually diverge.

4. Root Jumping

Problem
Sometimes the method may jump from one intended root region to another and converge to a
completely different root.

6
 Example

For
f (x) = sin x = 0,
if the initial guess is chosen near 2.4π, the iterates may unexpectedly jump and converge to the
root x = 0 instead of the intended nearby root around 2π.

Convergence Property

Quadratic Convergence
A theoretical analysis shows that
εt,i+1 ∝ (εt,i )2 .

Meaning

The true error at the next iteration is roughly proportional to the square of the previous error.
So the number of correct significant digits approximately doubles in each iteration. This is called
quadratic convergence.

Comparison with Bracketing Methods

Comparison

• Newton–Raphson is faster than bracketing methods like Bisection and False Position.

• However, convergence is not guaranteed.

• It requires derivative calculation.

• It is sensitive to the choice of initial guess.

Important Exam Points

Remember
1. Newton–Raphson is an open method.

2. It uses only one initial guess.

3. Main formula:
f (xi )
xi+1 = xi −
f ′ (xi )

4. It requires derivative f ′ (x).

5. It has quadratic convergence.

6. It is fast but not always guaranteed to converge.

7. Problems occur when f ′ (xi ) = 0 or near zero.

7
Final Summary

One-box summary

Newton–Raphson is an open method for solving f (x) = 0.

It starts from one initial guess and uses the tangent slope to improve the estimate.
f (xi )
xi+1 = xi −
f ′ (xi )
It converges very fast when the initial guess is good.
Its convergence is quadratic, so the accuracy improves rapidly.
However, it may diverge due to poor initial guesses, inflection points,
division by zero, oscillation near extrema, or root jumping.