Introduction
Infinite series are one of the most powerful tools in calculus and analysis, allowing us to represent complex functions as infinite sums of simpler polynomials. On the flip side, not every Taylor series converges for every value of $x$; they often converge only within a specific range. Among these, the Taylor series stands out as a fundamental method for approximating functions locally. Understanding how to find the radius of convergence is essential for determining exactly where these series are valid.
The radius of convergence is the distance from the center of the Taylor series to the boundary of the interval where the series converges. It acts as a "sphere of influence" for the series—inside this radius, the infinite sum behaves nicely and approximates the function, while outside it, the series diverges or becomes meaningless. Mastering the techniques to calculate this radius is a critical skill for anyone studying advanced calculus, differential equations, or complex analysis.
Detailed Explanation
To understand the radius of convergence, we must first understand the nature of a power series. A Taylor series centered at a point $a$ is a specific type of power series, written in the form:
$ \sum_{n=0}^{\infty} c_n (x - a)^n $
Here, $c_n$ represents the coefficients of the series (often derived from the derivatives of the function at point $a$), and $(x-a)$ is the variable term. The series is essentially a polynomial with infinitely many terms.
The concept of radius of convergence (denoted as $R$) describes the set of all $x$ values for which this infinite sum approaches a finite limit. If we plot the real line, the series will converge for all $x$ such that the distance between $x$ and the center $a$ is less than $R$. Mathematically, this is written as $|x - a| < R$.
Counterintuitive, but true.
It is crucial to distinguish between the radius and the interval of convergence. The radius $R$ is a non-negative number (or infinity), while the interval is the actual set of numbers, which might include or exclude the endpoints $a \pm R$ depending on whether the series converges at those specific points. To give you an idea, a series might have a radius of convergence of $5$, meaning it converges for $
$|x - a| < 5$, but we still need to check whether it converges when $x = a + 5$ or $x = a - 5$ separately. These endpoint checks are often the most delicate part of the analysis.
The Ratio Test and the Root Test
The two most common tools for determining the radius of convergence are the Ratio Test and the Root Test. Both are built on the idea that for a series $\sum c_n (x-a)^n$ to converge, its terms must eventually become small enough that the partial sums settle down to a finite value.
Ratio Test. Given the series $\sum c_n (x-a)^n$, consider the limit
$ L = \lim_{n \to \infty} \left| \frac{c_{n+1}}{c_n} \right|. $
If this limit exists (or is $+\infty$), then the radius of convergence is
$ R = \frac{1}{L}. $
When $L = 0$, we interpret $1/L$ as $\infty$, meaning the series converges for every real (or complex) value of $x$. When $L = \infty$, we set $R = 0$, so the series converges only at the center $x = a$.
Some disagree here. Fair enough.
Root Test. Alternatively, apply the Root Test by computing
$ \rho = \limsup_{n \to \infty} \sqrt[n]{|c_n|}. $
The radius of convergence is then
$ R = \frac{1}{\rho}, $
with the same conventions for $\rho = 0$ and $\rho = \infty$ as above. The Root Test is sometimes preferred when the coefficients $c_n$ are given by a formula that involves factorials or powers in a way that makes the $n$th root easier to evaluate than the ratio Nothing fancy..
Both tests are equivalent in the sense that they yield the same value of $R$ whenever the relevant limits exist. In practice, one chooses whichever computation is simpler for the series at hand.
Worked Examples
Example 1. Consider the exponential function $f(x) = e^x$, whose Taylor series about $a = 0$ is
$ \sum_{n=0}^{\infty} \frac{x^n}{n!}. $
Here $c_n = 1/n!$. Applying the Ratio Test:
$ L = \lim_{n \to \infty} \frac{1/(n+1)!}{1/n!Consider this: } = \lim_{n \to \infty} \frac{n! }{(n+1)!} = \lim_{n \to \infty} \frac{1}{n+1} = 0 And that's really what it comes down to..
Thus $R = 1/0 = \infty$. The exponential series converges for every real (and complex) number $x$. In fact, the series converges absolutely and uniformly on every bounded interval, which is why $e^x$ is so well behaved analytically.
Example 2. Take the geometric series $\sum_{n=0}^{\infty} x^n$. Here $c_n = 1$ for all $n$, so
$ L = \lim_{n \to \infty} \left| \frac{1}{1} \right| = 1, \qquad R = \frac{1}{1} = 1. $
The series converges when $|x| < 1$ and diverges when $|x| > 1$. And the first diverges (the partial sums grow without bound), while the second diverges as well because its terms do not approach zero. Even so, at the endpoints $x = 1$ and $x = -1$, the series becomes $\sum 1$ and $\sum (-1)^n$ respectively. Hence the interval of convergence is $(-1, 1)$, and the radius is $R = 1$ Turns out it matters..
Example 3. For the series $\sum_{n=0}^{\infty} \frac{n!}{n^n} x^n$, we again use the Ratio Test:
$ L = \lim_{n \to \infty} \frac{(n+1)!/(n+1)^{n+1}}{n!/n^n} = \lim_{n \to \infty} \frac{(n+1) \cdot n^n}{(n+1)^{n+1}} = \lim_{n \to \infty} \left( \frac{n}{n+1} \right)^n = \frac{1}{e} Most people skip this — try not to..
Therefore $R = e$. That said, the series converges absolutely for $|x| < e$ and diverges for $|x| > e$. Endpoint analysis would require a separate investigation, but the radius itself is cleanly determined Not complicated — just consistent..
Endpoint Behavior
As mentioned earlier, the radius $R$ tells us only about the open interval $|x-a| < R$. The behavior at the boundary points $x = a \pm R$ must be examined individually, typically by substituting the endpoint into the
