Radius of convergence

Some math rules work in a circle.

This circle has a size. It is called a radius. Inside the circle, the math works well. This is a safe zone for numbers.

Sometimes, the math stops working. This happens if you go too far from the center. The math might break at a certain point. We call these points singularities.

The radius is the distance to the nearest break. It tells us how big our safe circle can be.

Some math rules work for all numbers. These have an infinite radius. This means the safe zone never ends.

Some math rules work in a circle.

This circle is a safe zone. Inside the circle, a power series works well. We call this the disk of convergence. The size of this circle is the radius of convergence.

Why does the circle have a limit? It stops when the math breaks. These break points are called singularities. A singularity is a spot where a function is not defined. To find the radius, we look for the nearest break. The radius is the shortest distance from the center to a singularity.

Sometimes, the math works for every number. This means the radius is infinite. In other cases, the math might be tricky on the very edge of the circle. It might work at some spots on the edge but not others.

Scientists use special plots to guess the radius. One way is a Domb–Sykes plot. This helps when we only know a few parts of the series. It lets us see where the math might break in the future.

In mathematics, a power series is a way to represent a function using a long list of terms. This series works perfectly within a specific area, which looks like a circular disk. The size of this disk is called the radius of convergence. Inside this circle, the series is reliable and gives the correct answer. We call this safe area the disk of convergence.

If the series works for every possible number, we say the radius is infinite. This means the function is an entire function.

The radius has a limit because of points where the math stops working. These tricky points are called singularities. A singularity is a value where a function is not defined. To find the radius, you measure the distance from the center to the nearest singularity. The radius is always the shortest distance to one of these points.

Even if a function seems fine on a straight line, a singularity might hide nearby in the complex plane. For example, the function 1/(1+z^2) has singularities at ±i. Even though these are not on the real number line, they set the radius of convergence to 1.

Mathematicians use different tests to find this radius. One way is the root test, which uses a special limit called the limit superior. There is also the ratio test, which is often easier to calculate. These tests help determine if the series converges or diverges. The Cauchy–Hadamard theorem explains how the radius relates to these tests. If the distance from the center is less than the radius, the series converges. If the distance is greater, the series diverges.

Sometimes, we do not know every part of a series. In science, we might only know a few terms, like ten or a hundred. In these cases, we must estimate the radius using a plot. One method is the Domb–Sykes plot.

You can plot the coefficients to see their behavior. By drawing a straight line through these points, you can guess where the series will break. The point where the line hits the axis helps you find the radius. This is a very practical way to handle difficult problems.

Understanding the radius also helps us know how fast a series works. This is called the rate of convergence. Near the center, the series reaches a precise answer very quickly. As you move toward the edge of the circle, it gets slower. For example, calculating a value for the series e^x might take only two terms near zero. However, if you move further away, you might need 141 terms to be accurate.

The boundary of the circle can also be very complicated. At the very edge, the series might work at some spots but fail at others.

In mathematics, a power series is a way to represent a function using an infinite sum of terms. A power series is defined by a center, which we call $a$, and a sequence of coefficients, $c_n$. While these series can be used to describe many functions, they do not always work for every number. There is a specific region where the series is reliable and gives a stable answer. This region is a circular area known as the disk of convergence. The size of this disk is defined by the radius of convergence, $r$.

The radius of convergence tells us exactly how far from the center we can go while the series still converges. If the distance from the center $a$ to a variable $z$ is less than $r$, the series converges. If that distance is greater than $r$, the series diverges, meaning it fails to produce a stable value. When the radius is a positive number, the series converges absolutely and uniformly on compact sets inside the disk. If the series converges for every possible complex number, we say the radius is infinite. In this case, the function is called an entire function.

To understand why a radius exists, we must look at singularities. A singularity is a point where a function is not defined or ceases to be holomorphic. In complex analysis, the radius of convergence is equal to the distance from the center to the nearest singularity. This distance is the shortest possible path to a point where the function fails. Even if a function looks smooth on a real number line, a singularity might exist in the complex plane. For example, the function $1/(1+z^2)$ has no singularities on the real line. However, it has singularities at $i$ and $-i$. Since the distance from the center $0$ to these points is $1$, the radius of convergence is $1$.

Mathematicians use several methods to find the precise radius of convergence. One theoretical method is the root test, which utilizes the limit superior of the terms. The Cauchy–Hadamard theorem uses this test to establish the relationship between the coefficients and the radius. Another common method is the ratio test, which is often easier to compute when the limit exists. The ratio test shows that if the limit of the ratio of successive terms is a certain value, the radius of convergence is finite. These tests allow mathematicians to move from a list of coefficients to a concrete measurement of the disk.

In practical scientific applications, we rarely know every coefficient in a series. We might only know a few terms, ranging from a couple to perhaps a hundred. Because we cannot use the root test on a finite list, we must use estimation techniques. One such method is the Domb–Sykes plot.

By plotting the coefficients and performing a linear fit, we can extrapolate the behavior of the series. The intercept of this line on the vertical axis provides an estimate for the reciprocal of the radius. This allows researchers to predict where a series will fail even with limited data.

When the coefficients have a complex pattern of signs, more advanced procedures are required. Mercer and Roberts proposed a method involving an associated sequence to handle these cases. By plotting these values and extrapolating to infinity, one can estimate the radius of convergence. This method also helps identify the degree and the angle of the nearest singularity. This level of detail is vital when the series is being used to model physical systems or complex mathematical structures.

It is also important to consider the rate of convergence. This describes how many terms are needed to reach a specific level of precision. Near the center of the disk, a series converges very quickly. For instance, calculating $e^x$ near zero might only require two terms for high accuracy. However, as you move toward the boundary, the rate of convergence slows down significantly. To achieve the same precision at a distance, you might need 141 terms instead of two.

The boundary of the disk of convergence, where the distance equals $r$, can exhibit very complex behavior. At this edge, the series might converge at some points and diverge at others. Some series converge at every point on the boundary, while others diverge everywhere on it. Even if a series converges everywhere on the boundary, it may not converge absolutely. This distinction is a key part of understanding the limits of power series and how they behave as they approach their breaking point.