You are on page 1of 6

Solving Polynomial Equations

A polynomial is an expression of the following form:

p(x) = anxn + an-1xn-1 + an-2xn-2 + … a2x2 + a1x + a0

where an (called the coefficients) are real numbers, and n (called the degree) is a
non-negative integer. This is called a polynomial expression in x. The first
coefficient in the expression, an, is called the leading coefficient, and a0 is called
the trailing constant.

An especially important problem is solving the equation p(x) = 0. Often, it turns


out that there is more than one value for x that satisfies the equation. The values
for x that give p(x) = 0 are called the roots of the polynomial. Generalisations for
solutions exist for polynomials of degree 1, 2, 3 and 4, though the latter two have
computationally difficult solutions.

One of the most frequently used tools in finding the roots of a polynomial
equation is the factor theorem.

Factor Theorem
If p(x) is a polynomial and p(r) = 0, then r is a root and (x – r)
factors the polynomial. That is, gives a remainder of 0.

While the factor theorem provides a large clue when trying to find the roots of
the polynomial, we need more hints to make it easier to find which values of r
yield p(r) = 0.

Types of roots

Integer roots: r such that r is from the set of integers.

Rational roots: r = , where p is a factor of the trailing constant, and q is a factor


of the leading coefficient. (If q = ±1, then r is an integer root.)

Complex roots: z = a + bi, where i2 = -1 and is called the imaginary unit.


→ If z = a + bi is a root, then so is its complex conjugate, z = a – bi.

Irrational roots: Real roots that cannot be written as a fraction of integers.


→ To find irrational roots, one must use an approximation method such as
Newton’s method, Brent’s method, the bisection method, or Google.
Variations in sign

Make sure the terms of the polynomial are written from highest to lowest
degree. If the coefficient of a term has a different sign than the coefficient of the
next highest degree term, that is called a variation in sign and is counted as
one variation.

e.g. p(x) = 3x3 – 2x2 + x – 12

The coefficients are, in order: +, -, +, -; which counts as three variations in sign.

We can list the possible rational roots of this polynomial by taking the positive
and negative integer factors of the constant term. Factors of 12 include:
±1, ±2, ±3, ±4, ±6, and ±12. But, we also have to divide each of those by each of
the factors of 3: ±1, ±2, and ±3. So, our final set of possible rational roots is:

±, ±, ±, ±1, ±, ±, ±2, ±3, ±4, ±6, and ±12.

It’s important to remember that to take this set of rational roots, all the
coefficients in the polynomial must be integers. If they aren’t, you can factor out
the lowest common denominator, and then use the factors of the constant and
leading coefficient to find the potential rational roots.

Descartes’s Rule of Signs

Variations in sign can give us big clues as to how many positive or negative
roots a polynomial has. Knowing that a polynomial of degree n has n roots,
this can help us reduce the set of rational roots in which we attempt to use in the
factor theorem.

Descartes’s Rule of Signs


→ The number of positive roots of p(x) is given by the number
of variations in sign, or less that by an even number.
→ The number of negative roots of p(x) is given by the number of
variations in sign of p(–x), or less that by an even number.

For example, if there are three variations in sign, there are either three or one
positive root(s). If there are four variations in sign, there are four, two, or no
positive roots.

To find the number of possible negative roots, evaluate p(–x) and count the
variations in sign. This effectively means to switch the sign of the coefficients of
all odd-degreed terms in the expression.
(example continued on next page)

Descartes’s Rule of Signs (cont’d)


e.g. p(x) = 3x3 – 2x2 + x – 12 p(–x) = –3x3 – 2x2 – x – 12

There are three variations in sign in p(x), so there are either three or one
positive root(s). There no variations in sign in p(–x), meaning there are no
negative roots at all. So, there are no factors of the form (x + r), because r must
be positive, and we can effectively eliminate all the negative rational roots from
our earlier set. This potentially halves the amount of guesses we have to take
when using the factor theorem!

Our new set of possible rational roots omits any negative roots. So we have:

, , , 1, , , 2, 3, 4, 6, and 12.

p(x) is a cubic polynomial, meaning it has a third degree term – it has three
roots. We have two possible cases for the nature of its roots: one positive root, or
three positive roots. In the case that it has three positive roots, it would have no
other roots – so no complex or negative roots, and we’ll likely be able to find at
least one of its roots by the factor theorem using the rational roots from our list.

But if p(x) has only one positive root, then its other two roots must be either
irrational or complex. Now, suppose the other two roots were both irrational –
they can’t be positive, because we determined by Descartes’s rule that p(x) has
only one positive root. The irrational roots can’t be negative either, because p(x)
has no negative roots. Therefore, the other two non-positive roots of p(x) must be
complex roots. Because we know that complex roots come in pairs of
conjugates, the remaining two roots are x = a + bi and x = a – bi.

We can figure that what will likely happen is that the cubic polynomial has one
positive rational root r from our list, and dividing will yield a quadratic with
imaginary roots. Then we can find those with the quadratic formula, if we have to,
or we can leave the quadratic in expanded form.

The magical part is that we figured out the nature of the roots of this polynomial
without doing any factoring – we eliminated half our set of possible rational roots,
and discovered that carrying out the division will yield a quadratic that can’t be
factored over the real numbers. This method might take some time to get used
to, but if you can start to see these patterns quickly, then it will save you time
when trying to solve equations like these.

Intermediate Value Theorem


The intermediate value theorem has a few ingredients.

→ An open interval (a, b) on the domain.


→ A function f(x) continous for x∈(a, b)
For our purposes, we have a function
p(x), which is continuous for x∈R.
→ f(a) and f(b) have different signs.

The intermediate value theorem then


states that there exists at least one
c∈(a, b) such that f(c) = 0. This means
that c is a real root (integer, rational, or
irrational) of p(x).

We can use this when plugging in values to find p(x) = 0. For example:
If we evaluate p(1) and get a negative value, and then p(5) is a positive value, we
know that somewhere on the interval x∈(1, 5) lies a root of p(x). Whether this
root is rational or irrational is question we can answer later.

From this, we can suggest a more counter-intuitive method that can make the
brute-force method of the factor theorem a little less tedious: when choosing
values from your set of possible rational roots, choose values that are
reasonably far apart from each other. That way, if two “far apart” values yield
opposite signs for p(x), you know to try values that occur between those two
values.
Boundaries on roots

If we use synthetic division to divide the polynomial p(x) by a binomial (x – r) –


that is, to carry out the division statement – and we end up finding that r is not a
factor (because it leaves a remainder after the division), we can still use
information from the resulting quotient that tells us about boundaries on the set of
possible rational roots.

Suppose r is a rational number from our set of possible rational roots, which we
listed by looking at the polynomial p(x). We carry out using synthetic division,
and find that there is a remainder, meaning r is not a root of p(x).

→ If the bottom row (the row under the bar, where the quotient is written)
contains only non-negative numbers (zero and positive numbers), then r is an
upper bound on the roots – meaning that if real roots exist for p(x), then there
are no real roots greater than r.
→ If, after carrying out , we find that the bottom row contains only non-negative
numbers, then r is a lower bound on the roots – no real roots exist for p(x) that
are lower than r.

Notice that the lower bound rule tells us to evaluate p(–x) and then divide by
(x – r). This can be time consuming and annoying, so we can use a different,
equivalent rule for finding the lower bound on the roots:

→ If after carrying out the division , the bottom row of synthetic division contains
numbers alternating in non-positivity and non-negativity, then r is a lower
bound on the roots.

Let’s look at the content of this rule. Suppose synthetic division of a polynomial
p(x) by (x – r) gives a “bottom” row” of numbers like these: 1, 0, 4, –2, 0, –3.

The numbers take the form: +, 0, +, –, 0, –. These alternate in their sign, as it is


put, because zero can function as non-positive or non-negative. The first
number is positive, the second is non-positive (because zero is non-positive), the
third is positive, non-positive, zero (positive or non-negative), and non-positive
again. Because these numbers successfully have alternating signs, r must be a
lower bound on the roots, and no other real roots are lower than r.

Finding the boundaries on the roots can seriously reduce the amount of
possible rational roots, making our brute-force factor theorem work even easier.
Combine these root boundary rules with Descartes’s rule of signs and the
intermediate value theorem to solve these polynomial equations supa-fast.
Other important facts

Perfect square (a + b)2 = a2 + 2ab + b2

Perfect cube (a + b)3 = a3 + 3a2b + 3ab2 + b3

Difference of squares a2 – b2 = (a + b)(a – b)

Sum of cubes a3 + b3 = (a + b)(a2 – ab + b2)

Difference of cubes a3 – b3 = (a – b)(a2 + ab + b2)

You might also like