Skip to content

Reciprocal Equations

Reciprocal equations have a distinct characteristic that if a number is a root, its reciprocal is also a root. For a polynomial equation of the form

\[ a_0x^n + a_1x^{n-1} + \ldots + a_{n-1}x + a_n = 0, \quad ... [1]\]

the reciprocal roots property implies that if \( \alpha \) is a root, then \( \frac{1}{\alpha} \) must also satisfy the equation. This leads us to consider the polynomial

\[ a_nx^n + a_{n-1}x^{n-1} + \ldots + a_1x + a_0 = 0, \quad ... [2]\]

where the coefficients are in reverse order compared to the original polynomial. This new polynomial will have roots that are the reciprocals of the original polynomial's roots.

Since both equations [1] and [2] have the roots, their coefficients are proportional, that leads to:

\[ \frac{a_0}{a_n} = \frac{a_1}{a_{n-1}} = \ldots = \frac{a_{n-1}}{a_1} = \frac{a_n}{a_0}. \]

Let the above proportionality be equal to \( \lambda \).

Then observe that the first ratio \(\frac{a_0}{a_n}=\lambda\) and the last ratio \(\frac{a_n}{a_0}=\lambda\).

Multiply both ratios:\(\frac{a_0}{a_n}.\frac{a_n}{a_0}={\lambda}^2\) \(\implies \lambda = \pm 1\)

and hence, \( \lambda \) is either \( 1 \) or \( -1 \). The nature of \( \lambda \) dictates two types of reciprocal equations:

  1. Type I: When \( \lambda = 1 \), the coefficients are symmetric with respect to the midpoint of the sequence: \( a_r = a_{n-r} \) for \( r = 0, 1, 2, \ldots, n \). That is, \(a_0=a_n, a_1=a_{n-1}, a_2 = a_{n-2}, ...\)

    For example: \(2x^4-3x^3+40x^2-3x+2=0\) and \(3x^5-2x^4-7x^3-7x^2-2x+3=0\) are examples of reciprocal equations of type I.

  2. Type II: When \( \lambda = -1 \), the coefficients are symmetric but alternate in sign: \( a_r = -a_{n-r} \) for \( r = 0, 1, 2, \ldots, n \). That is, \(a_0=-a_n, a_1=-a_{n-1}, a_2 = - a_{n-2}, ...\)

    For example: \(2x^4-3x^3+3x-2=0\) and \(3x^5-2x^4+7x^3-7x^2+2x-3=0\) are examples of reciprocal equations of type II

Type I reciprocal equations of even degree are the most important. As other reciprocal equations Type I odd degree, Type II odd and even degree can be converted to Type I even degree. We will learn in the next section how we can solve these equations.

Important Property

I. A reciprocal polynomial equation \( p(x) = 0 \) of Type I has the property that its polynomial function \( p(x) \) satisfies the condition:

\[ x^n p\left(\frac{1}{x}\right) = p(x) \]

for every \( x \neq 0 \), where \( n \) is the degree of the polynomial. This means that the polynomial remains unchanged when each \( x \) term is replaced by its reciprocal and then multiplied by \( x^n \).

For example:

To verify whether the polynomial \( p(x) = x^6 + 7x^5 + 7x + 1 \) is a Type I reciprocal polynomial, we need to check if it satisfies the condition:

\[ x^n p\left(\frac{1}{x}\right) = p(x) \]

Since the degree \( n \) of the polynomial is 6, let's substitute \( x \) with \( \frac{1}{x} \) and multiply by \( x^6 \):

\[ x^6 p\left(\frac{1}{x}\right) = x^6 \left(\left(\frac{1}{x}\right)^6 + 7\left(\frac{1}{x}\right)^5 + 7\left(\frac{1}{x}\right) + 1\right) \]

Now, let's simplify the expression:

\[ x^6 p\left(\frac{1}{x}\right) = 1 + 7x + 7x^5 + x^6 \]

This is the original polynomial \( p(x) \) rewritten, which shows that \( p(x) \) satisfies the condition for a Type I reciprocal polynomial. Hence, the polynomial \( p(x) \) satisfies the equation \( x^6 p\left(\frac{1}{x}\right) = p(x) \).

II. For a Type II reciprocal polynomial equation \( p(x) = 0 \), the defining characteristic is that its polynomial function \( p(x) \) satisfies the condition:

\[ x^n p\left(\frac{1}{x}\right) = -p(x) \]

for all \( x \neq 0 \), where \( n \) is the degree of the polynomial. This means that the polynomial changes sign but otherwise remains unchanged when each \( x \) term is replaced by its reciprocal and then multiplied by \( x^n \).

For example:

Let \( p(x) = 2x^5 + 4x^4 - 5x^3 + 5x^2 - 4x - 2 = 0 \). It is clearly a type II polynomial.

Replace \( x \) with \( \frac{1}{x} \) and multiply by \( x^5 \):

\[ x^5 p\left(\frac{1}{x}\right) = x^5 \left(2\left(\frac{1}{x}\right)^5 + 4\left(\frac{1}{x}\right)^4 - 5\left(\frac{1}{x}\right)^3 + 5\left(\frac{1}{x}\right)^2 - 4\frac{1}{x} - 2\right) \]

Simplify the equation:

\[ x^5 p\left(\frac{1}{x}\right) = 2 + 4x - 5x^2 + 5x^3 - 4x^4 - 2x^5 \]

Comparing this with the original polynomial \( p(x) \), we see that:

\[ x^5 p\left(\frac{1}{x}\right) = -p(x). \]

Solving Type I even degree Reciprocal Equation

To solve a Type I reciprocal equation of even degree \(2m\), with the general form

\[ a_0x^{2m} + a_1x^{2m-1} + \ldots + a_1x + a_0 = 0, \]

we can simplify the solving process as follows:

  1. Divide the entire equation by \( x^m \) to normalize it:

    \[ a_0x^m + a_1x^{m-1} + \ldots + a_1x^{-m+1} + a_0x^{-m} = 0. \]
  2. Regroup the terms by pairing the \( x \) terms and their reciprocal power terms:

    \[ a_0(x^m + \frac{1}{x^m}) + a_1(x^{m-1} + \frac{1}{x^{m-1}}) + \ldots + a_{m-1}(x + \frac{1}{x}) = 0. \]
  3. Introduce a new variable \( z \) where \( z = x + \frac{1}{x} \). Notice that \( z \) will generate all the necessary powers when raised to higher powers itself:

    \[ z^2 = x^2 + 2 + \frac{1}{x^2} \implies x^2 + \frac{1}{x^2} = z^2 - 2,\]
    \[ z^3 = x^3 + 3x + \frac{3}{x} + \frac{1}{x^3} \implies x^3 + \frac{1}{x^3} = z^3-3z\]
    \[ \vdots \]
    \[ z^m = x^m + \ldots + \frac{1}{x^m}. \]
  4. The original equation can now be rewritten in terms of \( z \) to get a new polynomial equation of degree \( m \), which is half the original degree:

    \[ b_0z^m + b_1z^{m-1} + \ldots + b_{m-1}z + b_m = 0, \]

    with each \( b_i \) being a combination of the original coefficients \( a_i \).

  5. Solve this new polynomial equation for \( z \) using an appropriate method.

  6. Once the values for \( z \) are found, use them to construct quadratic equations to solve for \( x \) since \( z = x + \frac{1}{x} \) implies \( x^2 - zx + 1 = 0 \).

Example

Solve:

\[ 6x^4 - 35x^3 + 62x^2 - 35x + 6 = 0. \]

Solution:

We first notice the symmetry in the coefficients, indicating that it is indeed a Type I equation. The next steps involve transforming the equation and solving it:

  1. Divide the entire equation by \( x^2 \):

    \[ 6x^2 - 35x + 62 - \frac{35}{x} + \frac{6}{x^2} = 0. \]
  2. Regroup terms to form symmetric pairs around \( x^2 \) and \( x^{-2} \):

    \[ 6\left(x^2 + \frac{1}{x^2}\right) - 35\left(x + \frac{1}{x}\right) + 62 = 0. \]
  3. Let \( z = x + \frac{1}{x} \), hence \( z^2 = x^2 + 2 + \frac{1}{x^2} \), and \( x^2 + \frac{1}{x^2} = z^2 - 2 \).

  4. Substitute \( z \) into the equation:

    \[ 6(z^2 - 2) - 35z + 62 = 0, \]
    \[ 6z^2 - 35z + 50 = 0. \]
  5. This is now a quadratic equation in \( z \). Solve for \( z \):

    Let's solve \( 6z^2 - 35z + 50 = 0 \) using the quadratic formula:

    \[ z = \frac{-(-35) \pm \sqrt{(-35)^2 - 4 \cdot 6 \cdot 50}}{2 \cdot 6}. \]

    This yields two possible solutions for \( z \):

    \[ z_1 = \frac{35 + 5}{12} = \frac{40}{12} = \frac{10}{3} \]
    \[ z_2 = \frac{35 - 5}{12} = \frac{30}{12} = \frac{5}{2} \]

Thus, the solutions for \( z \) are \( z_1 = \frac{10}{3} \) and \( z_2 = \frac{5}{2} \).

Having found \( z \) values \( \frac{10}{3} \) and \( \frac{5}{2} \), we can now find \( x \) by solving the quadratic equations derived from the substitution \( z = x + \frac{1}{x} \). This gives us two quadratic equations to solve for \( x \):

For \( z_1 = \frac{10}{3} \):

\[ \begin{align*} x + \frac{1}{x} &= \frac{10}{3} \\ x &= \frac{10 \pm \sqrt{(-10)^2 - 4 \cdot 3 \cdot 3}}{2 \cdot 3} \\ x &= \frac{10 \pm \sqrt{100 - 36}}{6} \\ x &= \frac{10 \pm \sqrt{64}}{6} \\ x &= \frac{10 \pm 8}{6} \end{align*} \]
\[ \begin{align*} x_1 &= \frac{10 + 8}{6} = \frac{18}{6} = 3 \\ x_2 &= \frac{10 - 8}{6} = \frac{2}{6} = \frac{1}{3} \end{align*} \]
\[ 3x^2 - 10x + 3 = 0 \]

For \( z_2 = \frac{5}{2} \):

\[ x + \frac{1}{x} = \frac{5}{2} \]
\[ 2x^2 - 5x + 2 = 0 \]
\[ \begin{align*} x &= \frac{5 \pm \sqrt{(-5)^2 - 4 \cdot 2 \cdot 2}}{2 \cdot 2} \\ x &= \frac{5 \pm \sqrt{25 - 16}}{4} \\ x &= \frac{5 \pm \sqrt{9}}{4} \\ x &= \frac{5 \pm 3}{4} \end{align*} \]

This results in two more solutions for \( x \):

\[ x_3 = \frac{5 + 3}{4} = \frac{8}{4} = 2 \]
\[ x_4 = \frac{5 - 3}{4} = \frac{2}{4} = \frac{1}{2} \]

In summary, the solutions for the original quartic equation \( 6x^4 - 35x^3 + 62x^2 - 35x + 6 = 0 \) are \( x_1 = 3 \), \( x_2 = \frac{1}{3} \), \( x_3 = 2 \), and \( x_4 = \frac{1}{2} \).

Solving other Reciprocal Equations

Type I odd degree Reciprocal Equations

Theorem: Conversion of Type I Odd-Degree Reciprocal Equations to Type I Even-Degree

Let \( p(x) \) be a Type I reciprocal polynomial of odd degree \( 2m+1 \), given by

\[ p(x) = a_0x^{2m+1} + a_1x^{2m} + \ldots + a_{2m}x + a_0 = 0, \]

where the coefficients are symmetric with respect to the center, fulfilling the condition \( a_i = a_{2m+1-i} \) for all \( i \). Then \( p(x) \) can be factored as \( p(x) = (x + 1)q(x) \), where \( q(x) \) is a Type I reciprocal polynomial of even degree \( 2m \). In other words, a Type I odd-degree reciprocal equation can be converted to a Type I even-degree equation by dividing it with \(x + 1\).

Proof:

To prove that a Type I odd-degree reciprocal equation can be converted to a Type I even-degree equation by dividing it with \(x + 1\), let's begin with the general form of a Type I odd-degree reciprocal equation:

\[ p(x) = a_0x^{2m+1} + a_1x^{2m} + \ldots + a_{2m}x + a_0 = 0. \]

Notice that the coefficients are symmetrical and the equation is of odd degree \(2m+1\).

We group terms symmetrically from the highest and lowest powers towards the center:

\[ p(x) = a_0(x^{2m+1} + 1) + a_1(x^{2m} + x) + \ldots + a_m(x^{m+1} + x^m) = 0. \]

By inspection, we can see that each grouped term contains a factor of \(x + 1\). So we can express \(p(x)\) as:

\[ p(x) = (x + 1)q(x), \]

where \(q(x)\) is a polynomial of degree \(2m\), which makes it an even-degree polynomial.

To show that \(q(x)\) is also a Type I equation, we can consider the transformation:

\[ x \rightarrow \frac{1}{x}. \]

Since \(p(x)\) is a Type I equation, we have \(p(\frac{1}{x}) = p(x)/x^{2m+1}\). Now, let's apply this to the equation \(q(x)\):

\[ x^{2m}q\left(\frac{1}{x}\right) = \frac{p(\frac{1}{x})}{\frac{1}{x} + 1} = \frac{p(x)/x^{2m+1}}{\frac{x+1}{x}} = \frac{p(x)}{x+1}. \]

We know that \(p(x) = (x+1)q(x)\), so when we substitute \(p(x)\) back into the equation, we get:

\[ x^{2m}q\left(\frac{1}{x}\right) = q(x). \]

This demonstrates that \(q(x)\) satisfies the property of a Type I equation since it remains unchanged under the transformation \( x \rightarrow \frac{1}{x} \) after being multiplied by \(x^{2m}\). Hence, \(q(x)\) is a Type I even-degree equation.

Example

Problem:

Solve \(4x^3 - 13x^2 - 13x + 4 = 0\).

Solution:

Since the equation is a Type I odd degree reciprocal equation, we factor \(x + 1\) from the polynomial:

\[ (x + 1)(4x^2 - 17x + 4) = 0. \]

Now, we find the roots of the quadratic equation \(4x^2 - 17x + 4 = 0\) by applying the quadratic formula or factoring. The roots are \( x = 1/4 \) and \( x = 4 \).

Therefore, the roots of the original cubic equation are \( x = -1, x = 1/4, \) and \( x = 4 \).

Type II odd degree Reciprocal Equation

Theorem: Factorization of Type II Reciprocal Polynomials of Odd Degree

Let \( p(x) \) be a Type II reciprocal polynomial of odd degree \( 2m+1 \), defined by

\[ p(x) = a_0x^{2m+1} + a_1x^{2m} + \ldots + a_{2m}x + a_{2m+1} = 0, \]

where the coefficients satisfy \( a_i = -a_{2m+1-i} \) for all \( i \). Then \( p(x) \) can be factorized as \( p(x) = (x - 1)q(x) \), where \( q(x) \) is a Type I reciprocal polynomial of even degree \( 2m \).

Proof:

To prove that Type II reciprocal equations of odd degree can be factorized into a product of a Type I even degree equation and \( x-1 \), let's start with a Type II reciprocal polynomial of degree \( 2m+1 \):

\[ p(x) = a_0x^{2m+1} + a_1x^{2m} + \ldots + a_{2m}x + a_{2m+1} = 0, \]

where the coefficients satisfy the relationship \( a_i = -a_{2m+1-i} \) for \( i = 0, \ldots, 2m \).

First, we show that \( p(1) = 0 \). Substituting \( x = 1 \) into \( p(x) \), we have:

\[ p(1) = a_0 + a_1 + \ldots + a_{2m} + a_{2m+1}. \]

Since \(a_0=-a_{2m+1}, a_1=-a_{2m}, ..., a_m=-a_{m+1}\), we have:

\[ p(1) = (a_0+a_{2m+1})+ (a_1+a_{2m})+ ...+ (a_m+a_{m+1}) = 0. \]

We can use also prove that \(p(1)=0\) by using fact that,

\[ x^{2m+1} p\left(\frac{1}{x}\right) = -p(x). \]

Substitute \( x = 1 \):

\[ 1^{2m+1} p(1) = -p(1), \]
\[ p(1) = -p(1). \]

This implies \( p(1) = 0 \), which means, by remainder theorem, \( x - 1 \) is a factor of \( p(x) \). So we can write \( p(x) \) as:

\[ p(x) = (x-1)q(x), \]

where \( q(x) \) is a polynomial of degree \( 2m \).

To prove that \( q(x) \) is a Type I reciprocal polynomial, we show that it satisfies the identity \( x^{2m} q\left(\frac{1}{x}\right) = q(x) \). Using the fact that \( p(x) \) is Type II and the relationship between \( p(x) \) and \( q(x) \), we get:

\[ x^{2m+1} p\left(\frac{1}{x}\right) = -p(x), \]
\[ (x-1)x^{2m} q\left(\frac{1}{x}\right) = -(x-1)q(x). \]

Dividing both sides by \( x-1 \) (assuming \( x \neq 1 \)) gives us:

\[ x^{2m} q\left(\frac{1}{x}\right) = -q(x), \]

which after considering the Type II property \( x^{2m} q\left(\frac{1}{x}\right) = q(x) \) implies \( q(x) \) must be of Type I since the sign change is no longer present.

Thus, \( q(x) \) is a Type I even degree polynomial, and the original Type II odd degree polynomial \( p(x) \) can indeed be factorized into \( (x-1)q(x) \).

Type II Even degree Reciprocal equation

Theorem: Factorization of Type II Reciprocal Polynomials of Even Degree

Let \( p(x) \) be a Type II reciprocal polynomial of even degree \( 2m \), given by

\[ p(x) = a_0x^{2m} + a_1x^{2m-1} + \ldots + a_{2m-1}x + a_{2m} = 0, \]

where the coefficients satisfy \( a_i = -a_{2m-i} \) for \( i = 0, \ldots, m-1 \) and \( a_i = a_{2m-i} \) for \( i = m, \ldots, 2m \). Then \( p(x) \) can be factorized as \( p(x) = (x^2 - 1)q(x) \), where \( q(x) \) is a Type I reciprocal polynomial of degree \( 2m-2 \).

Proof:

  1. Since, \( p(x) \) is a Type II reciprocal polynomial of even degree \( 2m \) and it satisfies the identity

    \[ x^{2m}p\left(\frac{1}{x}\right) = -p(x), \]

    then setting \( x = 1 \) gives us

    \[ 1^{2m}p\left(\frac{1}{1}\right) = -p(1), \]
    \[ p(1) = -p(1), \]

    which implies \( p(1) = 0 \), indicating that \( x = 1 \) is a root of \( p(x) \).

    Similarly, setting \( x = -1 \) gives us

    \[ (-1)^{2m}p\left(\frac{1}{-1}\right) = -p(-1), \]
    \[ p(-1) = -p(-1), \]

    which implies \( p(-1) = 0 \), indicating that \( x = -1 \) is also a root of \( p(x) \).

    Since both \( x = 1 \) and \( x = -1 \) are roots, \( x - 1 \) and \( x + 1 \) are factors of the polynomial \( p(x) \). The product \( (x - 1)(x + 1) = x^2 - 1 \) is therefore a factor of \( p(x) \), and we can express \( p(x) \) as

    \[ p(x) = (x^2 - 1)q(x), \]

    where \( q(x) \) is the remaining polynomial of degree \( 2m - 2 \), which must be a Type I reciprocal polynomial because the Type II property is no longer in effect once the \( x^2 - 1 \) factor is removed.

  2. Now, let's demonstrate that \( q(x) \) is a Type I reciprocal polynomial. By the Type II property, we have:

    \[ x^{2m} p\left(\frac{1}{x}\right) = -p(x), \]

    which we apply after factoring out \( x^2 - 1 \):

    \[ (x^2 - 1)x^{2m-2} q\left(\frac{1}{x}\right) = -(x^2 - 1)q(x). \]
  3. Dividing both sides by \( x^2 - 1 \) (and noting that \( x \neq \pm1 \)), we obtain:

    \[ x^{2m-2} q\left(\frac{1}{x}\right) = q(x), \]

confirming that \( q(x) \) satisfies the Type I reciprocal property.

Therefore, the polynomial \( p(x) \) can be factorized into \( (x^2 - 1)q(x) \), with \( q(x) \) being a Type I reciprocal polynomial of degree \( 2m-2 \).`