08 April 2016

prerequisites: The post An introduction to Horner’s method

1. Introduction

The goal of this post is to derive Taylor polynomials using Horner’s method for polynomial division.

The post is structured as follows. In Section 2, we introduce the concept of Taylor polynomials and Taylor’s theorem. In Section 3, we derive a procedure for obtaining Taylor polynomials using Horner’s method for polynomial division. The post is concluded in Section 4.

2. Taylor polynomials and Taylor’s theorem

In this section, we first state the polynomial remainder theorem followed by the definitions of Taylor series and Taylor polynomials, which we use to finally state Taylor’s theorem.

As pointed out in the previous blog post, if we divide a polynomial, , with a binomial, , the remainder of the division is equal to , which is captured by the polynomial remainder theorem.

Theorem 1 (Polynomial remainder theorem). Given a polynomial,

where , and a binomial,

where , the remainder of dividing with , denoted , is equal to . Furthermore, divides if and only if .

Next, we define Taylor series and Taylor polynomials in order to state Taylor’s theorem.

A Taylor series is the representation of a function as an infinite sum of terms, calculated from the values of the function’s derivatives at a specific point.

Definition 1 (Taylor series). Given a function and a natural number , the Taylor series of is,

which can be written as,

A Taylor series with a finite number of terms, , is called a Taylor polynomial and written,

Since we are working solely with polynomials, and not other types of functions, we are able to restate any polynomial as a Taylor polynomial, calculating the exact same values. This brings us to the following simplified version of Taylor’s theorem - without the error function - defined over polynomials and natural numbers.

Theorem 2 (Taylor’s theorem) Given a polynomial, , and two natural numbers, and , the -th order Taylor polynomial of , , at the point is,

Having stated the above definitions and theorems, we now show how to obtain Taylor polynomials using Horner’s method.

3. Generating Taylor polynomials

From Theorem 2, we know that given a polynomial,

where , and a , the Taylor polynomial of at the point is,

where every occurrence of the variable has been substituted with and every coefficient has been substituted with . Thus, we need a way to compute these new values using Horner’s method.

If we let and , we can calculate the coefficients of – without the use of Horner’s method – by evaluating and its first three derivatives for ,

which yields the -rd order Taylor polynomial of at point ,

Looking at the calculations above, we do not only have to evaluate four polynomials and divide each of them with a factorial, but we also have to take the repeated derivative of . It would be useful if we could calculate these values using our existing definitions. From Theorem 1, we know that dividing with a binomial, ,1

yields the quotient and remainder , which is also equal to , since . This corresponds to the result of Formula \ref{eq:taylor-poly-ex-p-2}, which is also why we have subscripted the remainder with a , since it is the value of the coefficient of with index ,

Furthermore, it turns out that if we keep dividing the obtained quotient, a pattern emerges that connects the remainders of the subsequent divisions with the remaining coefficients of . If we divide the quotient of the first division, , with the same binomial as before, ,

we get the quotient and remainder . In line with the previous result, we notice that the remainder, , is equal to the result of Formula \ref{eq:taylor-poly-ex-pp-2}, i.e., the value of the coefficient of with index ,

If we repeat this procedure once more with the quotient ,

we get the remainder , which matches the coefficient with index in Formula \ref{eq:taylor-poly-ex-ppp-2},

and the quotient , which is also equal to the last remainder, , since is constant, and therefore it is also equal to the coefficient with index in Formula \ref{eq:taylor_poly_ex_pppp-2},

Now, with the following coefficients in hand,

the -rd order Taylor polynomial of at point becomes,

which is equal to the last Taylor polynomial in Formula \ref{eq:taylor-poly-ex-p-2-result}. Thus, we have demonstrated how to obtain the Taylor polynomial of a polynomial at a point , by repeatedly dividing the resulting quotient polynomials with a binomial, , using Horner’s method, where is the initial polynomial to be divided.2

4. Conclusion

In this post, we have shown how to obtain Taylor polynomials with Horner’s method for polynomial division.

In our next post, we use what we have learned from this – and the previous – blog post to derive Moessner’s sieve3 from Horner’s method.

  1. Note that we use the tabular representation for polynomial division, which we introduced in Formula 7 of the previous blog post

  2. See “The wonder of Horner’s method” (2003) by Alex Pathan and Tony Collyer. 

  3. Moessner’s sieve is the procedure described in “Eine Bemerkung über die Potenzen der natürlichen Zahlen” (1951) by Alfred Moessner, and the term Moessner’s sieve was first coined by Olivier Danvy in the paper “A Characterization of Moessner’s sieve” (2014).