algebra911.com

Chebyshev Polynomials

Q: What is the degree of the polynomial T_n(x)?

The degree of the Chebyshev polynomial $T_n(x)$ is exactly $n$. This means the highest power of $x$ in the polynomial is $x^n$. For example, $T_3(x) = 4x^3 - 3x$ is a degree-3 polynomial.

Download as PDF

Ever wondered if there's a special connection between polynomials and the cosine function from trigonometry? Discover Chebyshev polynomials, a fascinating sequence of functions that bridge algebra and geometry, with surprising applications in everything from computer science to engineering.

Chebyshev Polynomials — an original Algebra911 reference diagram defining chebyshev polynomials with its key formula and a worked example. — Chebyshev Polynomials: An Introduction

What Are Chebyshev Polynomials?

Chebyshev polynomials are a special sequence of polynomials, named after the Russian mathematician Pafnuty Chebyshev, that have powerful connections to trigonometry. Think of them as a family of functions where each member is related to the others in a predictable way. There are two main types: Chebyshev polynomials of the first kind, denoted as $T_{n} (x)$ , and of the second kind, denoted as $U_{n} (x)$ . In this lesson, we will focus on the more common first kind, $T_{n} (x)$ , which are fundamental in many areas of mathematics and engineering.

Unlike simple polynomials like $x^{2} + 2 x + 1$ , Chebyshev polynomials are defined by a deeper relationship with the cosine function. This trigonometric link gives them unique and extremely useful properties, especially on the interval from $- 1$ to $1$ . For each non-negative integer $n$ , there is one Chebyshev polynomial $T_{n} (x)$ . The first few are very simple: $T_{0} (x) = 1$ , $T_{1} (x) = x$ , and $T_{2} (x) = 2 x^{2} - 1$ . As you'll see, there's a beautiful pattern that allows us to generate any polynomial in the sequence from the previous two.

How Are They Defined? The Cosine Connection

The most elegant definition of Chebyshev polynomials of the first kind comes directly from trigonometry. It's a single, powerful formula that unlocks their entire structure:

T_n(\cos \theta) = \cos(n\theta)

Let's break this down. The formula says that if you take the cosine of an angle $θ$ , and plug that value into the $n$ -th Chebyshev polynomial, you get the exact same result as if you first multiplied the angle $θ$ by $n$ and then took the cosine. This seems like magic, but it allows us to generate the polynomials.

To turn this into a polynomial in $x$ , we use the substitution $x = \cos θ$ . Let's see how this works for the first few values of $n$ .

For $n = 0$ : $T_{0} (\cos θ) = \cos (0 \cdot θ) = \cos (0) = 1$ . So, $T_{0} (x) = 1$ .
For $n = 1$ : $T_{1} (\cos θ) = \cos (1 \cdot θ) = \cos (θ)$ . Since $x = \cos θ$ , we have $T_{1} (x) = x$ .

These are our starting points. To find the next one, $T_{2} (x)$ , we need a trigonometric identity.

Example 1

Find the Chebyshev polynomial $T_{2} (x)$ using the trigonometric definition.

Solution:

We start with the definition for $n = 2$ : $T_{2} (\cos θ) = \cos (2 θ)$ .

Now, we need to express $\cos (2 θ)$ in terms of $\cos θ$ . We use the double-angle identity from trigonometry: $\cos (2 θ) = 2 \cos^{2} θ - 1$ .

Substituting this back into our definition:

T_{2} (\cos θ) = 2 \cos^{2} θ - 1

Finally, we replace every $\cos θ$ with $x$ :

T_{2} (x) = 2 x^{2} - 1

And there it is! We've generated a standard quadratic polynomial from a purely trigonometric rule.

Can We Find Chebyshev Polynomials Without Trigonometry?

While the cosine definition is beautiful, it can be tedious to use trigonometric identities for higher values of $n$ . Fortunately, there is a much more direct way to generate the sequence of Chebyshev polynomials using a recurrence relation. A recurrence relation is a rule that defines each term in a sequence based on the preceding terms.

The recurrence relation for Chebyshev polynomials of the first kind is:

T_{n+1}(x) = 2xT_n(x) - T_{n-1}(x)

To use this formula, we just need the first two polynomials to start, which we already found: $T_{0} (x) = 1$ and $T_{1} (x) = x$ . From there, we can generate all the others algebraically.

Example 2

Use the recurrence relation to find $T_{2} (x)$ , $T_{3} (x)$ , and $T_{4} (x)$ .

Solution:

We have our starting polynomials: $T_{0} (x) = 1$ and $T_{1} (x) = x$ .

To find $T_{2} (x)$ : Let $n = 1$ in the recurrence formula.

T_{1 + 1} (x) = 2 x T_{1} (x) - T_{1 - 1} (x)

T_{2} (x) = 2 x T_{1} (x) - T_{0} (x)

Now substitute the known polynomials:

T_{2} (x) = 2 x (x) - 1 = 2 x^{2} - 1

This matches the result we got from the cosine definition!

To find $T_{3} (x)$ : Let $n = 2$ in the recurrence formula.

T_{3} (x) = 2 x T_{2} (x) - T_{1} (x)

Substitute the polynomials we know now:

T_{3} (x) = 2 x (2 x^{2} - 1) - x = 4 x^{3} - 2 x - x = 4 x^{3} - 3 x

To find $T_{4} (x)$ : Let $n = 3$ in the recurrence formula.

T_{4} (x) = 2 x T_{3} (x) - T_{2} (x)

Substitute again:

T_{4} (x) = 2 x (4 x^{3} - 3 x) - (2 x^{2} - 1) = 8 x^{4} - 6 x^{2} - 2 x^{2} + 1 = 8 x^{4} - 8 x^{2} + 1

Here is a table of the first few Chebyshev polynomials of the first kind:

n	$T_{n} (x)$
0	$1$
1	$x$
2	$2 x^{2} - 1$
3	$4 x^{3} - 3 x$
4	$8 x^{4} - 8 x^{2} + 1$
5	$16 x^{5} - 20 x^{3} + 5 x$
6	$32 x^{6} - 48 x^{4} + 18 x^{2} - 1$

What Are the Key Properties of Chebyshev Polynomials?

Chebyshev polynomials aren't just a mathematical curiosity; they have several remarkable properties that make them incredibly useful, especially within the interval $x \in [- 1, 1]$ . This interval is special because it corresponds to the possible values of $\cos θ$ .

Boundedness: For any $x$ in the interval $[- 1, 1]$ , the value of $T_{n} (x)$ is always between $- 1$ and $1$ . That is, $| T_{n} (x) | \leq 1$ . This comes directly from the definition $T_{n} (\cos θ) = \cos (n θ)$ , since the cosine function itself never goes above 1 or below -1.
Equiripple Behavior: The graph of $T_{n} (x)$ wiggles up and down between $- 1$ and $1$ . It reaches its maximum value of $1$ or its minimum value of $- 1$ exactly $n + 1$ times on the interval $[- 1, 1]$ . This 'equal-ripple' nature is their most famous property and is key to their use in approximation.
Roots (Chebyshev Nodes): The polynomial $T_{n} (x)$ has exactly $n$ distinct roots, all of which are real and lie inside the interval $(- 1, 1)$ . These roots, called Chebyshev nodes, are not evenly spaced; they are clustered more densely toward the endpoints of the interval. The formula for the roots is $x_{k} = \cos (\frac{(2 k - 1) π}{2 n})$ for $k = 1, 2, . . ., n$ .
Leading Coefficient: For $n \geq 1$ , the coefficient of the highest power term ( $x^{n}$ ) in $T_{n} (x)$ is $2^{n - 1}$ . You can verify this in the table above. For $T_{4} (x)$ , the leading coefficient is $8$ , which is $2^{4 - 1}$ .
Even and Odd Functions: If $n$ is an even number, $T_{n} (x)$ is an even function (meaning $T_{n} (- x) = T_{n} (x)$ ), containing only even powers of $x$ . If $n$ is an odd number, $T_{n} (x)$ is an odd function (meaning $T_{n} (- x) = - T_{n} (x)$ ), containing only odd powers of $x$ .

Why Are These Polynomials So Important?

The special properties of Chebyshev polynomials make them essential tools in many advanced fields. While the details can be complex, the core ideas are quite intuitive.

The most significant application is in polynomial approximation. Imagine you have a very complicated function, like a messy sine wave or an exponential curve, and you want to find a simpler polynomial that is a 'best fit' for it. If you use a standard polynomial, it might be very close to the function in the middle but fly far away at the ends. A Chebyshev approximation, however, minimizes the maximum error across the entire interval. Because of its equiripple property, the error is spread out evenly in a wavelike pattern. This is called minimizing the 'minimax' error, and it's the most efficient way to approximate a function with a polynomial of a given degree.

This has practical consequences in:

Computer Science: When a computer or calculator finds $\sin (x)$ or $\log (x)$ , it isn't looking up the value in a giant table. It's using a very accurate, pre-calculated polynomial approximation. Often, these are based on Chebyshev polynomials because they provide the most accuracy for the least amount of computational work.
Numerical Analysis: When solving complex equations on a computer, small rounding errors can build up and lead to huge inaccuracies. Using methods based on Chebyshev polynomials (especially evaluating functions at the Chebyshev nodes) can make calculations much more stable and reliable.
Engineering: In electrical engineering, Chebyshev polynomials are used to design electronic filters (like those in a radio or a speaker system). A 'Chebyshev filter' allows certain frequencies to pass through while blocking others, and it has a very sharp transition between the pass and stop bands, which is a direct result of the polynomial's properties.

Key formulas for chebyshev polynomials by Algebra911.

How Can We Use Chebyshev Polynomials to Rewrite Expressions?

One interesting application of Chebyshev polynomials is that they can be used as a basis for other polynomials. This means we can write any polynomial as a sum of Chebyshev polynomials. This is often done to make the expression more stable for computer calculations. Let's see how to express a simple power of $x$ in terms of $T_{n} (x)$ polynomials.

Example 3

Express $x^{4}$ as a linear combination of Chebyshev polynomials.

Solution:

Our goal is to write $x^{4} = c_{4} T_{4} (x) + c_{3} T_{3} (x) + c_{2} T_{2} (x) + c_{1} T_{1} (x) + c_{0} T_{0} (x)$ for some coefficients $c_{i}$ .

Step 1: Start with the highest-degree Chebyshev polynomial, $T_{4} (x)$ , and solve for the highest power of $x$ .

From our table, we know $T_{4} (x) = 8 x^{4} - 8 x^{2} + 1$ .

Let's isolate $x^{4}$ :

8 x^{4} = T_{4} (x) + 8 x^{2} - 1

x^{4} = \frac{1}{8} T_{4} (x) + x^{2} - \frac{1}{8}

Step 2: We now have a term we need to eliminate: $x^{2}$ . Repeat the process with $T_{2} (x)$ .

We know $T_{2} (x) = 2 x^{2} - 1$ .

Isolate $x^{2}$ :

2 x^{2} = T_{2} (x) + 1

x^{2} = \frac{1}{2} T_{2} (x) + \frac{1}{2}

Step 3: Substitute the expression for $x^{2}$ back into our equation for $x^{4}$ .

x^{4} = \frac{1}{8} T_{4} (x) + (\frac{1}{2} T_{2} (x) + \frac{1}{2}) - \frac{1}{8}

Step 4: Simplify the expression.

x^{4} = \frac{1}{8} T_{4} (x) + \frac{1}{2} T_{2} (x) + \frac{1}{2} - \frac{1}{8}

x^{4} = \frac{1}{8} T_{4} (x) + \frac{1}{2} T_{2} (x) + \frac{4}{8} - \frac{1}{8}

x^{4} = \frac{1}{8} T_{4} (x) + \frac{1}{2} T_{2} (x) + \frac{3}{8}

Step 5: Express any remaining constant in terms of $T_{0} (x)$ .

Since $T_{0} (x) = 1$ , the constant $\frac{3}{8}$ is simply $\frac{3}{8} T_{0} (x)$ .

So, our final answer is:

x^{4} = \frac{1}{8} T_{4} (x) + \frac{1}{2} T_{2} (x) + \frac{3}{8} T_{0} (x)

We have successfully rewritten $x^{4}$ using only Chebyshev polynomials.

What Are Some Common Mistakes to Avoid?

Mixing up the Recurrence Formula: A common error is getting the signs or coefficients wrong in the recurrence relation. Remember it's $T_{n + 1} (x) = 2 x T_{n} (x) - T_{n - 1} (x)$ . A plus sign instead of a minus will lead to completely different polynomials.
Forgetting the Starting Points: The recurrence relation is useless without the correct starting polynomials: $T_{0} (x) = 1$ and $T_{1} (x) = x$ . Don't accidentally start with $T_{0} (x) = 0$ or $T_{1} (x) = 1$ .
Applying Properties Outside $[- 1, 1]$ : The amazing property that $| T_{n} (x) | \leq 1$ is only guaranteed for values of $x$ between $- 1$ and $1$ . Outside this interval, the values of the polynomials grow extremely rapidly. For example, $T_{4} (2) = 8 (2)^{4} - 8 (2)^{2} + 1 = 128 - 32 + 1 = 97$ , which is much larger than 1.
Confusing $n$ and $x$ : Remember that $n$ is the index, or order, of the polynomial and must be a non-negative integer. The variable $x$ is the input to the function and can be any real number.

Quick Summary and Reference

Here are the most important takeaways about Chebyshev polynomials of the first kind, $T_{n} (x)$ .

Core Concept: A sequence of polynomials with a deep connection to trigonometry.
Trigonometric Definition: The defining property that connects them to the cosine function.

T_n(\cos \theta) = \cos(n\theta)

Recurrence Relation: The easiest way to generate the polynomials algebraically.

T_{n+1}(x) = 2xT_n(x) - T_{n-1}(x)

Starting Points: The two polynomials needed to kick off the recurrence relation.

T_0(x) = 1 \quad \text{and} \quad T_1(x) = x

Key Property: On the interval $[- 1, 1]$ , the polynomials are bounded between -1 and 1.

|T_n(x)| \le 1 \quad \text{for} \quad x \in [-1, 1]

Main Application: Finding the 'best fit' polynomial approximation for other functions by spreading the error out as evenly as possible.

Frequently Asked Questions

How do you pronounce 'Chebyshev'?

The name is Russian, and a common English pronunciation is 'sheb-ee-shef'. Pafnuty Chebyshev was a prominent 19th-century mathematician who made contributions to many areas of mathematics.

Are there other types of Chebyshev polynomials?

Yes, there are also Chebyshev polynomials of the second kind, denoted $U_{n} (x)$ . They are related to the sine function by the formula $U_{n} (\cos θ) = \frac{\sin ((n + 1) θ)}{\sin θ}$ and follow a similar recurrence relation.

Why is the interval from -1 to 1 so important for these polynomials?

This interval is crucial because the trigonometric definition relies on the substitution $x = \cos θ$ . The cosine function only produces values between -1 and 1, so this is the natural domain where their special properties, like being bounded by 1, hold true.

Do Chebyshev polynomials work for numbers bigger than 1?

Yes, you can plug any real number into a Chebyshev polynomial. However, outside the interval $[- 1, 1]$ , their values grow very quickly and they lose the special 'equiripple' and 'bounded' properties that make them so useful for approximation.

What is the degree of the polynomial T_n(x)?

The degree of the Chebyshev polynomial $T_{n} (x)$ is exactly $n$ . This means the highest power of $x$ in the polynomial is $x^{n}$ . For example, $T_{3} (x) = 4 x^{3} - 3 x$ is a degree-3 polynomial.

Is this related to Chebyshev's inequality from statistics?

They are named after the same mathematician, Pafnuty Chebyshev, but they are completely different mathematical concepts. Chebyshev's inequality is a rule in probability and statistics, while Chebyshev polynomials are a sequence of functions in algebra and analysis.

Where might I see these polynomials in a high school class?

You would typically encounter Chebyshev polynomials in an advanced pre-calculus course when studying trigonometric identities, or perhaps in a computer science or programming class that deals with numerical approximation methods.