Chapter 2 - 2
Chapter 2 - 2
ROOTS OF EQUATIONS
Bracketing Methods
This section on roots of equations deals with methods that exploit the fact that a function
typically changes sign in the vicinity of a root. These techniques are called bracketing
methods because two initial guesses for the root are required. As the name implies, these
guesses must “bracket,” or be on either side of, the root.
GRAPHICAL METHODS
A simple method for obtaining an estimate of the root of the equation f(x) = 0 is to make a
plot of the function and observe where it crosses the x axis. This point, which represents
the x value for which f(x) = 0, provides a rough approximation of the root.
Graphical techniques are of limited practical value because they are not precise. However,
graphical methods can be utilized to obtain rough estimates of roots. These estimates
can be employed as starting guesses for numerical methods discussed in this chapter.
THE BISECTION METHOD
The bisection method, which is alternatively called binary chopping, interval halving,
or Bolzano’s method, is one type of incremental search method in which the interval is always
divided in half. If a function change sign over an interval, the function value at the
midpoint is evaluated. The location of the root is then determined as lying at the midpoint
of the subinterval within which the sign change occurs. The process is repeated to obtain
refined estimates. A simple algorithm for the bisection calculation is shown below.
Step 1: Choose lower xl and upper xu guesses for the root such that the function changes
sign over the interval. This can be checked by ensuring that f(xl)f(xu) < 0.
Step 2: An estimate of the root xr is determined by
𝑥𝑙+𝑥𝑢
xr= 2
Step 3: Make the following evaluations to determine in which subinterval the root lies:
(a) If f(xl)f(xr) < 0, the root lies in the lower subinterval. Therefore, set xu = xr and
return to step 2.
(b) If f(xl)f(xr) > 0, the root lies in the upper subinterval. Therefore, set xl = xr and
return to step 2.
(c) If f(xl)f(xr) = 0, the root equals xr; terminate the computation.
1|Page
Example: Determine the root of 𝑥 3 + 4𝑥 2 - 10 using bisection method. Employ initial guesses
of xl = 1 and xu = 2 and iterate until the estimated error ℇa falls below a level of ℇs = 0.000001%.
Solution:
Iteration 1
xl =1, xu = 2
𝑥𝑙+𝑥𝑢 1+2
xr= = =1.5
2 2
1.25 − 1.5
εa = | |x100% =20%
1.25
Iteration 3
xl =1.25, xu = 1.5
𝑥𝑙+𝑥𝑢 1.25+1.5
xr= = = 1.3750
2 2
2|Page
1.3750 − 1.25
εa = | |x100% =9.09%
1.3750
Iteration 4
xl =1.25, xu = 1.3750
𝑥𝑙+𝑥𝑢 1.25+1.3750
xr= = = 1.3125
2 2
Iteration 5
xl =1.3125, xu = 1.3750
𝑥𝑙+𝑥𝑢 1.3125+1.3750
xr= = = 1.3438
2 2
3|Page
THE FALSE-POSITION METHOD
Although bisection is a perfectly valid technique for determining roots, its “brute-force” approach
is relatively inefficient. False position is an alternative based on a graphical insight.
A shortcoming of the bisection method is that, in dividing the interval from xl to xu into
equal halves, no account is taken of the magnitudes of f(xl) and f(xu). For example, if f(xl)
is much closer to zero than f(xu), it is likely that the root is closer to xl than to xu.
An alternative method that exploits this graphical insight is to join f(xl) and f(xu) by a
straight line. The intersection of this line with the x axis represents an improved estimate of
the root.
Using similar triangles in the above figure, the intersection of the straight line with the x axis
can be estimated as
𝑓(𝑥𝑙) 𝑓(𝑥𝑢)
=
𝑥𝑟−𝑥𝑙 𝑥𝑟−𝑥𝑢
4|Page
Cross-multiplying the above equation yields
This is one form of the method of false position. Note that it allows the computation of the root xr
as a function of the lower and upper guesses xl and xu. It can be put in an alternative form by
expanding it:
𝑥𝑢𝑓(𝑥𝑙) 𝑥𝑙𝑓(𝑥𝑢)
xr= -
𝑓(𝑥𝑙)−𝑓(𝑥𝑢) 𝑓(𝑥𝑙)−𝑓(𝑥𝑢)
𝑓(𝑥𝑢)(𝑥𝑙−𝑥𝑢)
xr= xu -
𝑓(𝑥𝑙)−𝑓(𝑥𝑢)
which is the same as the former equation we obtained for false position method. We use this form
because it involves one less function evaluation and one less multiplication. In addition, it is
directly comparable with the secant method which will be discussed in this chapter.
5|Page
Example: Determine the root of 𝑥 3 + 4𝑥 2 - 10 using false position method. Employ initial guesses
of xl = 1 and xu = 2 and iterate until the estimated error ℇa falls below a level of ℇs = 0.000001%.
Solution;
Iteration 1
xl =1, xu = 2
f(xl) =13 + 4x12 – 10= -5
f(xu) =23 + 4x22 – 10= 14
𝑓(𝑥𝑢)(𝑥𝑙−𝑥𝑢)
xr= xu -
𝑓(𝑥𝑙)−𝑓(𝑥𝑢)
14(1−2)
xr= 2 - = 1.2632
−5−14
xl =1.2632, xu = 2
f(xl) =1.26323 + 4x1.26322 – 10= -1.6023
f(xu) =23 + 4x22 – 10= 14
𝑓(𝑥𝑢)(𝑥𝑙−𝑥𝑢)
xr= xu -
𝑓(𝑥𝑙)−𝑓(𝑥𝑢)
14(1.2632−2)
xr= 2 - = 1.3388
−1.6023−14
6|Page
Iteration 3
xl =1.3388, xu = 2
f(xl) =1.33883 + 4x1.33882 – 10= -0.4304
f(xu) =23 + 4x22 – 10= 14
𝑓(𝑥𝑢)(𝑥𝑙−𝑥𝑢)
xr= xu -
𝑓(𝑥𝑙)−𝑓(𝑥𝑢)
14(1.3388−2)
xr= 2 - = 1.3585
−0.4304−14
Iteration 4
xl =1.3585, xu = 2
f(xl) =1.35853 + 4x1.35852 – 10= -0.1100
f(xu) =23 + 4x22 – 10= 14
𝑓(𝑥𝑢)(𝑥𝑙−𝑥𝑢)
xr= xu -
𝑓(𝑥𝑙)−𝑓(𝑥𝑢)
14(1.3585−2)
xr= 2 - = 1.3635
−0.1100−14
Iteration 5
xl =1.3635, xu = 2
f(xl) =1.36353 + 4x1.36352 – 10= -0.0278
f(xu) =23 + 4x22 – 10= 14
𝑓(𝑥𝑢)(𝑥𝑙−𝑥𝑢)
xr= xu -
𝑓(𝑥𝑙)−𝑓(𝑥𝑢)
7|Page
14(1.3635−2)
xr= 2 - = 1.3648
−0.0278−14
Open Methods
For the bracketing methods we have seen the root is located within an interval prescribed by a
lower and an upper bound. Repeated application of these methods always results in closer
estimates of the true value of the root. Such methods are said to be convergent because they move
closer to the truth as the computation progresses in contrast, the open methods described in this
chapter are based on formulas that require only a single starting value of x or two starting values
that do not necessarily bracket he root. As such, they sometimes diverge or move away from the
true root as the computation progresses. However, when the open methods converge, they usually
do so much more quickly than the bracketing methods.
SIMPLE FIXED-POINT ITERATION
The formula for root estimation is obtained by rearranging the function f(x) = 0 so that x is on the
left-hand side of the equation:
x = g(x)
This transformation can be accomplished either by algebraic manipulation or by simply adding x
to both sides of the original equation. For example,𝑥 2 -2x +3 =0
𝑥 2 +3
can be simply manipulated to yield whereas sin x = 0 could be put into the form of Equation
2
above by adding x to both sides to yield x = sin x + x.
Example: Use simple fixed-point iteration to locate the root of f(x) = 𝑒 −𝑥 - x. Starting with an
initial guess of x0 = 0 and iterate up to fifth iteration, compute the approximate error for each
iteration.
Solution:
Express the given function in the form of xi+1 = g(xi)
xi+1 = 𝑒 −𝑥𝑖
8|Page
Iteration 1
xi+1 = 𝑒 0 =1
Iteration 2
xi+1 = 𝑒 −1 = 0.367879
𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑎𝑝𝑝𝑟𝑜𝑥𝑖𝑚𝑎𝑡𝑖𝑜𝑛 − 𝑝𝑟𝑒𝑣𝑖𝑜𝑢𝑠 𝑎𝑝𝑝𝑟𝑜𝑥𝑖𝑚𝑎𝑡𝑖𝑜𝑛
εa = | |x100%
𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑎𝑝𝑝𝑟𝑜𝑥𝑖𝑚𝑎𝑡𝑖𝑜𝑛
0.367879 − 1
εa = | |x100% =71.8%
1
Iteration 3
xi+1 = 𝑒 −0.367879 = 0.692201
0.692201 − 0.367879
εa = | |x100% =46.9 %
0.692201
Iteration 4
xi+1 = 𝑒 −0.692201 = 0.500473
0.500473 − 0.692201
εa = | |x100% =38.3%
0.500473
Iteration 5
xi+1 = 𝑒 −0.500473 = 0.606244
0.606244 − 0.500473
εa = | |x100% =17.4 %
0.606244
9|Page
THE NEWTON-RAPHSON METHOD
The following figure shows Graphical depiction of the Newton-Raphson method. A tangent to the
function of xi [that is, f’(xi)] is extrapolated down to the x axis to provide an estimate of the root
at xi+1.
𝑓(𝑥𝑖 ) − 0
𝑓 ′ (𝑥𝑖 ) =
𝑥𝑖 − 𝑥𝑖+1
10 | P a g e
Example: Use the Newton-Raphson method to estimate the root of f(x) = 𝑒 −𝑥 - x, employing an
initial guess of x0 = 0. and iterate up to fourth iteration, compute the approximate error for each
iteration.
Solution:
The first derivative of the function can be evaluated as
𝑓 ′ (𝑥) = -𝑒 𝑥 – 1
Iteration 1
Xi = 0
𝑓 ′ (𝑥) = -𝑒 −𝑥 – 1
𝑓(𝑥𝑖 )
xi+1 = xi -
𝑓′ (𝑥𝑖 )
𝑒 −𝑥𝑖 − 𝑥𝑖
xi+1 = xi - =0.5000
−𝑒 −𝑥𝑖 – 1
Iteration 2
Xi = 0.5000
𝑒 −𝑥𝑖 − 𝑥𝑖 𝑒 −0.5000 − 0.5000
xi+1 = xi - = 0.5000 - = 0.5663
−𝑒 −𝑥𝑖 – 1 −𝑒 −0.5000 – 1
0.5663 − 0.5000
εa = | |x100% =11.7093%
0.5663
Iteration 3
Xi = 0.5663
𝑒 −𝑥𝑖 − 𝑥𝑖 𝑒 −0.5663 − 0.5663
xi+1 = xi - = 0.5663 - = 0.5671
−𝑒 −𝑥𝑖 – 1 −𝑒 −0.5663 – 1
0.5671 − 0.5663
εa = | |x100% = 0.1467%
0.5671
Iteration 4
11 | P a g e
Xi =0.5671
𝑒 −𝑥𝑖 − 𝑥𝑖 𝑒 −0.5671 − 0.5671
xi+1 = xi - = 0.5671 - = 0.5671
−𝑒 −𝑥𝑖 – 1 −𝑒 −0.5671 – 1
0.5671 − 0.5663
εa = | |x100% = 2.210610−5 %
0.5671
Thus, the approach rapidly converges on the true root. Notice that the true percent relative error at
each iteration decreases much faster than it does in simple fixed-point iteration
This approximation can be substituted into Newton-Raphson formula to yield the following
iterative equation:
𝑓(𝑥𝑖 )(𝑥𝑖−1 − 𝑥𝑖 )
𝑥𝑖+1 =𝑥𝑖 –
𝑓(𝑥𝑖−1 )−𝑓(𝑥𝑖 )
The above Equation is the formula for the secant method. Notice that the approach requires two
initial estimates of x. However, because f(x) is not required to change signs between the estimates,
it is not classified as a bracketing method.
Example: Use the Secant method to estimate the root of f(x) = 𝑒 −𝑥 - x, Start with initial estimates
of x−1 = 0 and x0 = 1.0. and iterate up to fourth iteration, compute the approximate error for each
iteration.
Solution: Recall that the true root is 0.56714329. . .
Iteration 1
x−1 = 0
x0 = 1.0
𝑓(𝑥−1 )=1
12 | P a g e
𝑓(𝑥0 )= -0.6321
−0.6321(0− 1)
𝑥1 =1 – = 0.6127
1−(−0.6321)
Iteration 2
x−1 = 1
x0 = 0.6127
𝑓(𝑥−1 )= -0.6321
𝑓(𝑥0 )= -0.0708
−0.0708(1−0.6127)
𝑥1 =-0.6321 – = 0.5638
−0.6321−(−0.0708)
0.5638− 0.6127
εa = | |x100%=8.6659%
0.5638
Iteration 3
x−1 = 0.6127
x0 = 0.5638
𝑓(𝑥−1 )= -0.0708
𝑓(𝑥0 )= 0.0052
0.0052(0.6127−0.5638)
𝑥1 =0.5638 – = 0.5672
−0.0708−0.0052
0.5672− 0.5638
εa = | |x100%= 0.5875%
0.5672
Iteration 4
x−1 = 0.5638
x0 = 0.5672
𝑓(𝑥−1 )= 0.0052
𝑓(𝑥0 )= -4.241910−5
−4.2429x10−5 (0.5638−0.5672)
𝑥1 =0.5672 – = 0.5671
0.0052−(−4.241910−5 )
0.5671− 0.5672
εa = | |x100%= 0.0048%
0.5671
13 | P a g e
Modified Secant Method
Rather than using two arbitrary values to estimate the derivative, an alternative approach involves
a fractional perturbation of the independent variable to estimate f (x)
𝑓(𝑥𝑖 +𝛿𝑥𝑖 ) − 𝑓(𝑥𝑖 )
𝑓 ′ (𝑥𝑖 ) ≈
𝛿𝑥𝑖
where δ = a small perturbation fraction. This approximation can be substituted into Newton-
Raphson method formula to yield the following iterative equation:
𝛿𝑥𝑖 𝑓(𝑥𝑖 )
𝑥𝑖+1 =𝑥𝑖 –
𝑓(𝑥𝑖 +𝛿𝑥𝑖 ) − 𝑓(𝑥𝑖 )
Example: Use the modified secant method to estimate the root of f(x)=𝑥 4 - 𝑥 3 - 3𝑥 2 - x. Use a value
of 0.01 for δ and start with x = 2 and iterate up to fourth iteration, compute the approximate error
for each iteration.
Solution:
Iteration 1
x0 =2
𝑓(𝑥0 )= -6
x0 + δx0 =2.0200
𝑓(𝑥0 + 𝛿𝑥0 )= -5.8539
𝛿𝑥0 𝑓(𝑥0 )
𝑥1 =𝑥0 – = 2.8391
𝑓(𝑥0 +𝛿𝑥0 ) − 𝑓(𝑥0 )
14 | P a g e
Iteration 2
x0 =2.8391
𝑓(𝑥0 )= 15.0653
x0 + δx0 =2.8675
𝑓(𝑥0 + 𝛿𝑥0 )= 16.4955
𝛿𝑥0 𝑓(𝑥0 )
𝑥1 =𝑥0 – = 2.5359
𝑓(𝑥0 +𝛿𝑥0 ) − 𝑓(𝑥0 )
2.5359− 2.8391
εa = | |x100%= 11.9558%
2.5359
Iteration 3
x0 =2.5359
𝑓(𝑥0 )= 3.2187
x0 + δx0 = 2.5613
𝑓(𝑥0 + 𝛿𝑥0 )= 3.9906
𝛿𝑥0 𝑓(𝑥0 )
𝑥1 =𝑥0 – = 2.4286
𝑓(𝑥0 +𝛿𝑥0 ) − 𝑓(𝑥0 )
2.4286− 2.5359
εa = | |x100%= 4.4172%
2.4286
Iteration 4
x0 =2.5359
𝑓(𝑥0 )= 3.2187
x0 + δx0 = 2.5613
𝑓(𝑥0 + 𝛿𝑥0 )= 3.9906
𝛿𝑥0 𝑓(𝑥0 )
𝑥1 =𝑥0 – = 2.4286
𝑓(𝑥0 +𝛿𝑥0 ) − 𝑓(𝑥0 )
2.4286− 2.5359
εa = | |x100%= 4.4172%
2.4286
15 | P a g e
Roots of Polynomials
In this section we will discuss methods to find the roots of polynomial equations of the general
form f𝑛(𝑥)= a0 + a1x + a2𝑥 2 + ··· + an𝑥 𝑛
where n = the order of the polynomial and the a’s = constant coefficients.
The roots of such polynomials follow these rules:
1. For an nth-order equation, there are n real or complex roots. It should be noted that these roots
will not necessarily be distinct.
2. If n is odd, there is at least one real root.
3. If complex roots exist, they exist in conjugate pairs (that is, λ + μi and λ − μi), where i = √−1
The efficacy of methods we have seen so far depends on whether the problem being solved
involves complex roots. If only real roots exist, any of the previously described methods could
have utility. However, the problem of finding good initial guesses complicates both the bracketing
and the open methods, whereas the open methods could be susceptible to divergence.
When complex roots are possible, the bracketing methods cannot be used because of
the obvious problem that the criterion for defining a bracket (that is, sign change) does not
translate to complex guesses. Of the open methods, the conventional Newton-Raphson method
would provide a viable approach.
However, as might be expected, it would be susceptible to convergence problems. For this reason,
special methods have been developed to find the real and complex roots of polynomials. In this
section we will discuss the Müller and Bairstow methods.
MÜLLER’S METHOD
Recall that the secant method obtains a root estimate by projecting a straight line to the x axis
through two function values. Müller’s method takes a similar approach, but projects a parabola
through three points.
16 | P a g e
The above figure shows comparison of two related approaches for locating roots.
Müller’s method consists of deriving the coefficients of the parabola that goes through the three
points. These coefficients can then be substituted into the quadratic formula to obtain the point
where the parabola intercepts the x axis—that is, the root estimate. The approach is facilitated by
writing the parabolic equation in a convenient form,
f2(x) = a(𝑥 − 𝑥2 )2 + b (x - 𝑥2 ) + c
We want this parabola to intersect the three points [x0, f(x0)], [x1, f(x1)], and [x2, f(x2)]. The
coefficients of the above equation can be evaluated by substituting each of the three points to give
f(𝑥0 )= a(𝑥0 − 𝑥2 )2 + b (x0 - 𝑥2 ) + c
f(𝑥1 )= a(𝑥1 − 𝑥2 )2 + b (x1 - 𝑥2 ) + c
f(𝑥2 )= a(𝑥2 − 𝑥2 )2 + b (x2 - 𝑥2 ) + c
c= f(𝑥2 )
Thus, the coefficient c is merely equal to the function value evaluated at the third guess, x2. This
result can then be substituted into above equations to yield two equations with two unknowns:
f(𝑥0 ) - f(𝑥2 )= a(𝑥0 − 𝑥2 )2 + b (x0 - 𝑥2 )
f(𝑥1 ) - f(𝑥2 )= a(𝑥1 − 𝑥2 )2 + b (x1 - 𝑥2 )
Algebraic manipulation can then be used to solve for the remaining coefficients, a and b.
One way to do this involves defining a number of differences,
h0 = x1 − x0
h1 = x2 − x1
𝑓(𝑥1 ) − 𝑓(𝑥0 )
𝛿0 = 𝑥1 − 𝑥0
𝑓(𝑥2 ) − 𝑓(𝑥1 )
𝛿1 = 𝑥2 − 𝑥1
ℎ1 𝑏 - ℎ1 2 a =ℎ1 𝛿1
which can be solved for a and b. The results can be summarized as
𝛿1 − 𝛿0
a=
ℎ1 + ℎ0
17 | P a g e
b=aℎ1 + 𝛿1
c= f(x2)
then the root x3 is evaluated using the following formula
−2𝑐
x3 =x2 +
𝑏±√𝑏2 −4𝑎𝑐
Once x3 is determined using the above formula, the process is repeated. The points are discarded
according to the following strategies.
✓ If only real roots are being located, we choose the two original points that are nearest the
new root estimate, x3.
✓ If both real and complex roots are being evaluated, a sequential approach is employed. That
is, just like the secant method, x1, x2, and x3 take the place of x0, x1, and x2
Example: Use Müller’s method with guesses of x0, x1, and x2 = 4.5, 5.5, and 5, respectively, to
determine a root of the equation f(x) = 𝑥 3 − 13x − 12
Note that the roots of this equation are −3, −1, and 4.
Solution:
Iteration 1
X0=4.5
X1=5.5
X2=5
f(𝑥0 )= 4.53 – 13*4.5 – 12= 20.6250
f(𝑥1 )= 5.53 – 13*5.5 – 12=82.8750
f(𝑥2 )= 53 – 13*5 – 12= 48
h0 = x1 − x0= 5.5 − 4.5 = 1
h1 = x2 − x1= 5 − 5.5 = −0.5
𝑓(𝑥1 ) − 𝑓(𝑥0 ) 82.875 − 20.625
𝛿0 = = = 62.2500
𝑥1 − 𝑥0 5.5 − 4.5
𝛿 −𝛿 69.75 − 62.26
a= ℎ1 + ℎ 0 = =15
1 0 1−0.5
18 | P a g e
|𝑏 + √𝑏 2 − 4𝑎𝑐|=|62.25 + √62.252 − 4 ∗ 15 ∗ 48 |= 93.7946
Iteration 2
X0=5.5
X1=5
X2=3.9765
f(𝑥0 )= 82.8750
f(𝑥1 )= 48
f(𝑥2 )= -0.8163
h0 = -0.5
h1 = -1.0235
𝑓(𝑥1 ) − 𝑓(𝑥0 )
𝛿0 = = 69.75
𝑥1 − 𝑥0
𝑓(𝑥2 )− 𝑓(𝑥1 )
𝛿1 = = 47.6949
𝑥2 − 𝑥1
𝛿 −𝛿
a= ℎ1 + ℎ 0 = 14.4765
1 0
b=aℎ1 + 𝛿1 = 32.8780
c= f(x2)= -0.8163
19 | P a g e
4.0011− 3.9765
εa = | |x100%= 0.6139 %
4.0011
Iteration 3
X0=5
X1=3.9765
X2=4.0011
f(𝑥0 )= 48
f(𝑥1 )= -0.8163
f(𝑥2 )= 0.0368
h0 = -1.0235
h1 = 0.0246
𝑓(𝑥1 ) − 𝑓(𝑥0 )
𝛿0 = = 47.6949
𝑥1 − 𝑥0
𝑓(𝑥2 )− 𝑓(𝑥1 )
𝛿1 = = 34.7310
𝑥2 − 𝑥1
𝛿 −𝛿
a= ℎ1 + ℎ 0 = 12.9775
1 0
b=aℎ1 + 𝛿1 = 35.0498
c= f(x2)= 0.0368
4− 2.5359
εa = | |x100%= 0.0262%
4
20 | P a g e
Iteration 4
X0=3.9765
X1=4.0011
X2=4
f(𝑥0 )= -0.8163
f(𝑥1 )= 0.0368
f(𝑥2 )= 2.4684∗ 10−5
h0 = 0.0246
h1 = -0.0010
𝑓(𝑥1 ) − 𝑓(𝑥0 )
𝛿0 = = 34.7310
𝑥1 − 𝑥0
𝑓(𝑥2 )− 𝑓(𝑥1 )
𝛿1 = = 35.0126
𝑥2 − 𝑥1
𝛿 −𝛿
a= ℎ1 + ℎ 0 = 11.9775
1 0
b=aℎ1 + 𝛿1 = 35.0000
c= f(x2) = 2.4684∗ 10−5
BAIRSTOW’S METHOD
fn(x)= a0 + a1x + a2𝑥 2 +…+an𝑥 𝑛 can be divided by the factor x − t to yield a second polynomial
that is one order lower fn-1(x)= b1 + b2x + b3𝑥 2 +…+bn𝑥 𝑛−1
with a remainder R = b0, where the coefficients can be calculated by the recurrence relationship
bn= an
bi=ai +bi+1t for i=n-1 to 0
21 | P a g e
Note that if t were a root of the original polynomial, the remainder b0 would equal zero.
To permit the evaluation of complex roots, Bairstow’s method divides the polynomial
by a quadratic factor x2 - rx - s. If this is done to fn(x), the result is a new polynomial
fn-2(x)= b2 + b3x +… bn-1𝑥 𝑛−3 +bn𝑥 𝑛−2
with a remainder
R = b1(x - r) + b0
As with normal synthetic division, a simple recurrence relationship can be used to perform the
division by the quadratic factor:
bn = an
bn−1 = an−1 + rbn
bi = ai + rbi+1 + sbi+2 for i = n − 2 to 0
The quadratic factor is introduced to allow the determination of complex roots. This relates to the
fact that, if the coefficients of the original polynomial are real, the complex roots occur in conjugate
pairs. If x2 − rx − s is an exact divisor of the polynomial, complex roots can be determined by the
quadratic formula. Thus, the method reduces to determining the values of r and s that make the
quadratic factor an exact divisor. In other words, we seek the values that make the remainder term
equal to zero.
Inspection of R = b1(x - r) + b0 leads us to conclude that for the remainder to be zero, b0 and b1
must be zero. Because it is unlikely that our initial guesses at the values of r and s will lead to this
result, we must determine a systematic way to modify our guesses so that b0 and b1 approach zero.
According to Bairstow The changes, Δr and Δs, needed to improve our guesses can be estimated
using the following
c2 Δr + c3 Δs =-b1
c1 Δr + c2 Δs =-b0
cn = bn
cn−1 = bn−1 + rcn
ci = bi + rci+1 + sci+2 for i = n − 2 to 1
At each step, an approximate error in r and s can be estimated as follows
Δr
|𝜀𝑎,𝑟 |=| 𝑟 |100%
Δs
|𝜀𝑎,𝑠 |=| 𝑠 |100%
22 | P a g e
When both of these error estimates fall below a prespecified stopping criterion εs, the values of the
roots can be determined by
r ± √𝑟 2 +4𝑠
x= 2
23 | P a g e
Solving equations using Cramer's rule method
D=c2xc2-c1xc3= (-2.6) x (-2.6) -(-1.1) x0.7=7.53
D1=b0xc3-b1xc2=3.5x0.7-2.2x(-2.6) =8.17
D2=b1xc1-b0xc2=2.2x(-1.1)-3.5x (-2.6) =6.68
𝐷1
Δr = =1.085
𝐷
𝐷2
Δs = = 0.8871
𝐷
Next, the computation is repeated using the revised values for r and s
Iteration 2
b3=a3=0.7
b2=a2+rb3 = -4+2.085x0.7=-2.5405
b1 = a1 + rb2 + sb3 = 6.2+2.085x(-2.5405)-0.1129x0.7=0.824
b0 = a0 + rb1 + sb2 = -2+2.085x0.824-0.1129x(-2.5405)=0.0049
c3=b3=0.7
c2=b2+rc3 = -2.5405 +2.085x0.7=-1.081
c1 = b1 + rc2 + sc3 = 0. .824+2.085x(-1.081)-0.1129x0.7=-1.5089
Thus, the simultaneous equations to solve for Δr and Δs are
c2 Δr + c3 Δs =-b1
c1 Δr + c2 Δs =-b0
-1.081 Δr + 0.7 Δs =-0.824
-1.5089Δr – 1.081Δs =-0.0049
Solving equations using Cramer's rule method
24 | P a g e
D=c2xc2-c1xc3= (-1.081)x(-1.081)-(-1.5089)x0.7=2.2248
D1=b0xc3-b1xc2= 0.0049x0.7-0.824x(-1.081)=0.8942
D2=b1xc1-b0xc2= 0.824x(-1.5089)-0.0049x(-1.081)=-1.2381
𝐷1
Δr = =0.4019
𝐷
𝐷2
Δs = = -0.5565
𝐷
Next, the computation is repeated using the revised values for r and s
Iteration 3
b3=a3=0.7
b2=a2+rb3 =-4+2.4869x0.7=-2.2591
b1 = a1 + rb2 + sb3 = 6.2+2.4869x(-2.2591)-0.6694x0.7=0.1131
b0 = a0 + rb1 + sb2 = -2+2.4869x0.1131-0.6694x(-2.2591)=-0.2065
c3=b3=0.7
c2=b2+rc3 = -2.2591+2.4869x0.7=-0.5183
c1 = b1 + rc2 + sc3 = 0.1131+2.4869x(-0.5183)-0.6694x0.7=-1.6444
Thus, the simultaneous equations to solve for Δr and Δs are
c2 Δr + c3 Δs =-b1
c1 Δr + c2 Δs =-b0
-0.5183 Δr + 0.7 Δs =-0.1131
-1.6444Δr – 0.5183Δs = 0.2065
Solving equations using Cramer's rule method
D=c2xc2-c1xc3= (-0.5183)x(-0.5183)-(-1.6444)x0.7=1.4197
25 | P a g e
D1=b0xc3-b1xc2= (-0.2065)x0.7-0.1131x(-0.5183)=-0.086
D2=b1xc1-b0xc2= 0.1131x(-1.6444)-(-0.2065)x(-0.5183)=-0.293
𝐷1
Δr = =-0.0606
𝐷
𝐷2
Δs = = -0.2064
𝐷
Next, the computation is repeated using the revised values for r and s
The computation can be continued (see Table below), with the result that after six iterations the
method converges on values of r = 2.4357 (|𝜀𝑎,𝑟 |= 0%) and s = -0.8715 (|𝜀𝑎,𝑟 | = 0%)
26 | P a g e
Iteration Δr Δs r s
1 a3=0.7 b3=0.7 c3=0.7 1.085 0.8871 2.085 -0.1129
a2=-4 b2=-3.3 c2=-2.6
a1=6.2 b1=2.2 c1=-1.1
a0=-2 b0=3.5
2 a3=0.7 b3=0.7 c3=0.7 0.4019 -0.5565 2.4869 -0.669
a2=-4 b2=2.5405 c2=1.081
a1=6.2 b1=0.824 c1=1.5089
a0=-2 b0=0.0049
3 a3=0.7 b3=0.7 c3=0.7 -0.0606 -0.2064 2.4264 -0.8758
a2=-4 b2=2.2591 c2=0.5183
a1=6.2 b1=0.1131 c1=-1.6444
a0=-2 b0=0.2065
4 a3=0.7 b3=0.7 c3=0.7 0.0093 0.0043 2.4357 -0.8714
a2=-4 b2=-2.3015 c2=-0.6031
a1=6.2 b1=0.0026 c1=-2.0737
a0=-2 b0=0.0218
5 a3=0.7 b3=0.7 c3=0.7 0.0001 0 2.4357 -0.8715
a2=-4 b2=-2.295 c2=-0.5901
a1=6.2 b1=0.0001 c1=-2.0472
a0=-2 b0=0.0002
6 a3=0.7 b3=0.7 c3=0.7 0 0 2.4357 -0.8715
a2=-4 b2=-2.295 c2=-0.59
a1=6.2 b1=0 c1=-2.047
a0=-2 b0=0
27 | P a g e
c4=b4+rc5 = -4.5 +(-1) (1) =-5.5
c3 = b3 + rc4 + sc5 =6.25 +(-1) (-5.5) +(-1) (1) =10.75
c2 = b2 + rc3 + sc4 =0.375 +(-1) (10.75) +(-1) (-5.5) = -4.8750
c1 = b1 + rc2 + sc3 = -10.5 +(-1) (-4.875) +(-1) (10.75) = -16.3750
Next, the computation is repeated using the revised values for r and s
Iteration 2
b5=a5=1
b4=a4+rb5 = -3.5+( -0.6442) (1) = -4.1442
b3 = a3 + rb4 + sb5 =2.75 + (-0.6442) ( -4.1442) + (0.1381) (1) = 5.5578
b2 = a2 + rb3 + sb4 =2.125 +(-0.6442) (5.5578) + (0.1381) ( -4.1442) = -2.0276
b1 = a1 + rb2 + sb3 = -3.875 +(−0.6442) ( -2.0276) + (0.1381) (5.5578) = -1.8013
b0 = a0 + rb1 + sb2 =1.25 +(−0.6442) ( -1.8013) + (0.1381) ( -2.0276) = 2.1304
c5=b5=1
c4=b4+rc5 = -4.1442 +(-0.6442) (1) = -4.7884
c3 = b3 + rc4 + sc5 =5.5578 +(-0.6442) ( -4.7884) +( 0.1381) (1) = 8.7806
28 | P a g e
c2 = b2 + rc3 + sc4 =-2.0276 +(-0.6442) (8.7806) +(0.1381) ( -4.7884) = -8.3453
c1 = b1 + rc2 + sc3 = -1.8013 +(-0.6442) ( -8.3453) +(0.1381) (8.7806) = 4.7873
the simultaneous equations to solve for Δr and Δs are
c2 Δr + c3 Δs =-b1
c1 Δr + c2 Δs =-b0
-8.3453 Δr + 8.7806 Δs =1.8013
4.7873 Δr – 8.3453Δs =-2.1304
which can be solved for Δr = 0.1331 and Δs = 0.3316. Therefore, our original guesses can be
corrected to
r = −0.6442 + 0.1331 = −0.5111
s = 0.1381 + 0.3316 = 0.4697
Δr 0.1331
|𝜀𝑎,𝑟 |=| 𝑟 |100% =|−0.5111|100% =26.0%
Δs 0.3316
|𝜀𝑎,𝑠 |=| 𝑠 |100% =|0.4697|100% =70.6%
The computation can be continued, with the result that after four iterations the method
converges on values of r = -0.5 (|𝜀𝑎,𝑟 |= 0.063%) and s = 0.5 (|𝜀𝑎,𝑠 |= 0.040%). Then the roots can
r ± √𝑟 2 +4𝑠 −0.5 ± √−0.52 +4∗0.5
be determined by x = = = 0.5, -1.0
2 2
At this point, the quotient is the cubic equation f(x) = x3 − 4x2 + 5.25x − 2.5
Bairstow’s method can be applied to this polynomial using the results of the previous step, r = −0.5
and s = 0.5, as starting guesses. Five iterations yield estimates of r = 2 and s = −1.249, which can
be used to compute
r ± √𝑟 2 +4𝑠 2± √22 +4∗−1.249
x= = = 1 ± 0.499i
2 2
𝑠
At this point, the quotient is a first-order polynomial that can be directly evaluated by x= - 𝑟 =
29 | P a g e