You are on page 1of 10

Chapter 3.

1: Multiple
Roots

PEDRO FERNANDO QUIROGA


NOVOA

Escuela de Ingeniería de Petróleos

UNIVERSIDAD INDUSTRIAL DE
SANTANDER
Multiple Roots Multiple Roots
The multiple roots are determinate to polynomial
equations that have the general form:
 Basic operation of polynomials. f ( x )=a0 +a1 x+ a2 x 2 +…+ an x n

 Conventional methods f(x) = x 3 - 5x 2 + 7x -3


10

 Müller Method. 6

 Bairstow Method.

f(x)
2

 Complex roots -2

-4
0 0.5 1 1.5 2 2.5 3 3.5 4
x
Where n is the degree of the polynomial and a are 1. Basic operation of polynomials
the coefficients. The roots of polynomials can be
real and / or complex, and comply with three
rules:  Evaluation of polynomials:

 In an equation of degree n, there are n real Although this form of the equation:
or complex roots. It should be noted that f ( x )=a0 +a1 x+ a2 x 2 +…+ an x n
the roots are not necessarily different.
Is the most common, is not the best for
determining the value of a polynomial for a
 If n is odd there is at least one real root.
given value of x . Since for a polynomial of
order n requires n sums n(n+1) /2 multiplications.

 If there are complex roots, these are found So you can use the form nested to minimize
in conjugate pairs. the number of operations and reduce rounding
errors. For example for a polynomial of degree
3 in this way would be:
f ( x )=a0 + x ¿ ¿

 Deflation polynomial:
When determining the root of a polynomial of difficulty of finding an initial value and diverge in
degree n, if the procedure is repeated to find the the result.
root will yield the same result. Therefore it is
When there are complex roots, the closed
necessary to remove the root before continuing
methods are not applicable because the sign
so-called polynomial deflation. To perform this
change criterion is not applicable to complex
procedure should factor the polynomial, to
values. Therefore the Newton-Raphson method is
eliminate the root already found, or using
the most viable.
synthetic division, in either case the new
polynomial will have a degree less.

3. Müller Method:
2. Conventional methods

Conventional methods for finding the roots of a


polynomial are open and closed methods, the
efficiency of these methods depends if the
problem has complex roots. If there are only real
roots of the two methods are useful but have the The secant method obtained root of a function by
estimating a projection of a straight line on the x-
axis, through the values of the function. Müller's f (x 0)−f ( x 2 )=a( x 0 – x 2)2 +b ( x 0 – x2 )

method, works similarly but instead of projecting f (x 1)−f ( x 2)=a( x 1 – x 2)2 +b(x 1 – x 2)
a line using two points, requires three points to
(Equation 1 and 2)
calculate a parable.
Defining the differences:
For this we need three points
[ x0 , f ( x0 )],[ x1 , f ( x1 )] y [x 2 , f (x 2)]. The approximation we h0 =x 1 – x 0

can write as: h1 =x2 – x 1

f (x)=a( x – x 2)2 +b( x – x 2)+ c

The parabola's coefficients are calculated by f ( x¿¿ 1)– f (x 0 )


δ 0= ¿
solving the following system of equations. x1 – x 0

f ( x 0 ) =a ( x 0 – x 2) 2+ b(x 0 – x 2)+c

f (x 1)=a(x 1 – x 2 )2+ b( x 1 – x2 )+c f (x ¿¿ 1)


δ 1=f ( x¿ ¿2)– ¿¿
x 2 – x1
f (x 2)=a(x 2 – x2 )2+ b( x 2 – x 2)+c
These are replaced in equations 1 and 2 to obtain:
From the last equation we can see that the value
(h ¿ ¿ 0+ h1)b−(h ¿ ¿ 0+h1 )2 a=h 0 δ 0 +h1 δ 1 ¿ ¿
of C = f(x2). Substituting the values of C in the
other two equations we have: h1 b−h12 a=h 1 δ 1
The values of the coefficients are: With the initial values x 0 , x 1 y x 2 =4.5, 5.5 y 5 to
δ 1−δ 0 determine the roots of the equation.
a=
h1−h0
First the function is evaluated with the initial
values
b=a h1 +δ 1 f ( 4.5 )=20.625

f ( 5.5 )=82.875

c=f ( x 2 ) f ( 5 )=48

To find the root we applied of the alternative With these values are calculated:
quadratic formula, that is to say: h0 =5.5−4.5=1 h1=5−5.5=−0.5

−2 c 82.875 – 20.625
x 3−x 2= δ 0= =62.25
2
b ± √b −4 ac 5.5−4.5

48 – 82.875
The approximation error is determined by the δ 1= =69.75
5−5.5
formula:
With these values are calculated:
x 3−x 2
ε= | | x3
100 %
a=
δ 1−δ 0
h1−h0
=15

 Example:
ε =25.74 %

b=a h1 +δ 1=62.25 As the error is big, the new values are x 0 , x 1 y x 2 =


5.5, 5 y 3 . 976487.
The procedure is repeated and the final outcome
c=f ( x 2 )=48 of the result after four iterations is 4.
The square root of the discriminant is:
√ b2−4 ac=31.54461
Then as | 62.25 +31.54461 |> | 62.25-31.54461 |.
4. Bairstow Method:
It uses a positive sign in the denominator of the
quadratic formula and the new root will be The method is related to that of Müller and
estimated: Newton-Raphson. The method is based on
−2(48)
dividing a polynomial by a factor ( x−t ). For
x 3=5+ =3.976487
62.25+31.54461 example, the general polynomial
f n ( x )=a0 +a1 x+ a2 x 2 +…+ an x n

3.976487−5 Is divided by the factor x−t to have a lower


ε= | 3.976487|∗100 %
degree polynomial
f n−1 ( x )=b 1+ b2 x +b 3 x 2 +…+b n x n−1 para i=n−2 a 0

With a residual R=b , the coefficients are


0
One way to determine the values of r and s that
make zero waste is to use the Newton-Raphson
calculated as follows:
method. For this we need a linear approximation
b n=an
of b1 and b0 respect to r and s which we calculate
b i=ai +b i+1 t para i=n−1 a 0 using the Taylor series.
∂ b1 ∂ b1
b 1 ( r +∆ r , s+ ∆ s )=b1 + ∆ r+ ∆s
∂r ∂s
5. Complex Roots ∂ b0 ∂ b0
b 0 ( r + ∆ r , s+ ∆ s )=b 0+ ∆r+ ∆s
∂r ∂s
To assess the complex roots, Bairstow's
method divides the polynomial between factor Δr and Δs increases needed to improve the initial
x 2−rx−s. So the new polynomial thus: values are estimated by equating to zero the
above equations to give:
f n−2 ( x )=b 2+ b3 x+ …+ bn−1 xn−3 +b n x n−2
∂ b1 ∂ b1
R=bi ( x −r ) +b0 ∆r+ ∆ s=−b1
∂r ∂s
b n=an ∂ b0 ∂ b0
∆ r+ ∆ s=−b0
∂r ∂s
b n−1=an−1 +r bn

b i=ai +r∗bi +1+ s∗b i+2


Bairstow shows that the partial derivatives can be At each step an approximate error is calculated at
obtained by a procedure similar to synthetic r and s
division, thus:
c n=bn
|∆rr|100 %
|ε ar|=

∆s
c n−1=b n−1+ r cn |ε |=| s |100 %
as

c i=b i+ r ci +1+ s ci +2 para i=n−2 a 1


When two errors are below specified termination
Where: criteria, the roots are determined by...
∂ b0
=c 1 r ± √ r 2 +4 s
∂r x=
2
∂ b1 ∂b 0
= =c 2
∂r ∂s

∂ b1
=c 3
∂s

It is then
c 2 ∆ r +c 3 ∆ s=−b1

c 1 ∆ r +c 2 ∆ s=−b0
Bibliography

 DENNIS ZILL, (2009) “Differential Equations


with Boundary-Value Problems”. Chapter 1

 http://www2.fiu.edu/~pricer/Darcy
%20law.pdf

 http://www.ees.nmt.edu/Hydro/courses/ert
h441/lectures/L2_Darcys_Law.pdf

 http://www.material_simulacion.ucv.cl/elem
entos_de_los_modelos_matemat.htm

You might also like