Volume of a n-dimensional Ball: Part 2
1. Introduction
The first installment of this post investigated the question of computing the volume of a unit ball (and by extension any ball) of dimensions. The main result we obtained was that the volume can be computed recursively as , where
In this post, we will use this result to prove an explicit formula for .
2. The Formula
A quick search online yields the following explicit formula for :
where the function extends the factorial function to most of the real numbers. Note that actually equals , so to keep things simple, we will replace the denominator with , and it will be understood that this factorial is extended to the real numbers in place of . Note that it retains the important property that .
3. More About an
Before we begin the proof, we will investigate the properties of in greater detail. It seems quite difficult to obtain an explicit formula for at first glance (if any of you know of a simple way to evaluate that integral, please let me know in the comments!), but we can attempt to apply some tools from calculus to simplify that integral. Specifically, because we can write as times a smaller term, we think of applying integration by parts (inverse of product rule). Let
- (by the Chain Rule)
Then
A short computation shows that for , for . Thus, the first term on the right hand side equals 0. Furthermore, after cancelling the negative signs and factoring out from the integral, we may collect the and replace it with . Hence,
Thus, we have obtained a recursive formula for the sequence itself, in terms of .
4. The Proof of the Formula
Now, we have sufficient information to tackle the problem. We know that ; we can iterate this once more to get , and a third time, and so on. Thus, we have . Note that if we were to extend the definition of a unit sphere as “the set of all points a distance of 1 unit away from the center” to a -dimensional space, we see that a -D sphere is simply a line segment of length 2 around its center. Thus, is the “volume” (length for -D) of a -D ball, so . However, a quick computation shows that as well, so we may write
We will prove that this is equivalent to the above known formula for (from now on, assume is our derived formula and is the already-known formula. The following proof may be confusing, so here is an outline:
- First, we will show that and have the same values at .
- Then, we will show that the ratios and .
- Finally, we will prove that if the ratio , then that implies .
Observe how these steps are sufficient to prove that at all points. Essentially, we are proving that they start out the same way, and the scale factor between the first term and the term two steps ahead is the same in both sequences (and also for second and fourth terms), and finally that if this “two step ratio” is the same for both sequences at some term, then they are still the same after two more terms. The last two parts together prove that the “two step ratio” is always the same in both sequences, and that coupled with the first part shows that the terms themselves are always the same.
We will first show that for . This can be done as a straightforward computation (note that , , and ).
Now, we prove that and . Once more, these are simple computation problems and are left for the reader to verify.
Finally, we will prove that if , then . First of all, note that
Also,
Now suppose that . We want to prove that . By our recursion for , we have and . Plugging these in, we have
as claimed.
So we are finally done! We have fully proven the formula for the volume of a ball of any number of dimensions! Here are a few particular fruits of our labor: