×

# A Combinatorical Proof

Prove the following combinatorial:

The number of ways of placing $$n$$ labeled balls into $$n$$ indistinguishable boxes is equal to

$\dfrac{\displaystyle \sum_{k=0}^{\infty} \dfrac {(k+1)^{n-1}}{k!}}{e}$

Note by Sharky Kesa
1 month ago

## Comments

Sort by:

Top Newest

What you are counting is the Bell numbers, and your result is known as Dobinski's formula.

(By the way, this is not a bijection. A bijection is a 1-1 correspondence between two sets. What you want to prove is that a certain combinatorial formula is true.) · 1 month ago

Log in to reply

OK, thanks! However, I thought that a bijection could be constructed between these two by changing the summation. · 1 month ago

Log in to reply

Okay so I have a solution for putting $$n$$ balls in exactly $$k$$ boxes:

Define a function that takes an ordered pair of naturals to a nonnegative integer:

$$f(n,k)=\begin{cases} 1 & k=1 \\ 1 & n=k \\0 & k>n\\kf(n-1,k)+f(n-1,k-1) & \text{for other cases}\end{cases}$$

This function should satisfy it if my LaTeX is correct.

Now for putting $$n$$ balls in $$n$$ boxes, just do

$$\displaystyle \sum_{x=1}^{n}f(n,x)$$ · 1 month ago

Log in to reply

Now you just need to biject this summation to the other summation. · 1 month ago

Log in to reply

I looked it up and it I haven't seen any solutions which use a combinatorial bijection. · 1 month ago

Log in to reply

Can you think of any other solution tghen? It might help to create a bijection. · 1 month ago

Log in to reply

Are you sure that's correct? For example when $$n=1$$ we have $$1$$ way of putting a ball in a box, but the sum indicates that there are a few more than 1 way; for $$k=0$$ that term alone is already $$1$$. · 1 month ago

Log in to reply

I forgot to put a denominator in the sum but it has been rectified. · 1 month ago

Log in to reply

×

Problem Loading...

Note Loading...

Set Loading...