2

Let's say I'm trying to find a bijection from $\mathbb N^\mathbb N$, i.e., the set of all functions from $\mathbb N$ to $\mathbb N$, to some other set, say an open interval $(a,b)∈R$. What do I need to establish, to say there can exist a bijection? I do understand that I need to prove it's both injective and surjective, but how do you do it with a set of functions as the domain?

Let's say I'm trying to see if the injective part is true. So, I need to find a function that takes every possible sequence of natural numbers as input, and associates each of them with one real number in the interval $(a,b)$. But what does it mean to have a sequence of natural numbers as the input of a function? And even if that makes sense, what sort of functions can satisfy such a purpose? I hope I sound sane.

Train Heartnet
  • 2,534
  • 3
  • 25
  • 42

2 Answers2

6

Typically it's not so easy to explicitly write down a bijection. Instead you can use the Cantor-Bernstein-Schroeder theorem.

You can map an interval injectively into $\{0,1\}^{\mathbb N}$ using a base-$2$ representation.

You can map ${\mathbb N}^{\mathbb N}$ injectively into an interval using continued fractions.

Robert Israel
  • 448,999
  • Thank you. I have a related question, which I've posted as a comment under Henry's answer. Could you help me with that as well, please? – Train Heartnet Nov 12 '13 at 08:56
5

A function from $\mathbb N$ to $\mathbb N$ would be of the form

$$f(1)=n_1, f(2)=n_2, f(3)=n_3, f(4)=n_4, f(5)=n_5, \ldots$$ for particular values of $n_1,n_2,n_3,n_4,n_5,\ldots$, so producing a sequence of natural numbers. You can then regard the function and the sequence as equivalent.

A function from these to the real numbers is much as you might expect, so for example the continued fraction:

$$g(n_1,n_2,n_3,n_4,n_5,\ldots)= n_1 + \cfrac{1}{n_2 + \cfrac{1}{n_3 + \cfrac{1}{n_4 + \cfrac{1}{n_5+\cdots} } } }.$$

To show this is injective (which it is in this case) you need to show that each different sequence produces a different real number

To show this is surjective (which it is not quite in this case - try finding a rational) you need to show that each real number is produced by some sequence

Henry
  • 157,058
  • Thank you. I have a related question on this: I understand that if $A$ and $B$ are two sets, and if there exists a bijection from $A$ to $B$, then $|A|=|B|$. But is the converse true? I mean, can I say that if $|A|=|B|$, then there exists a bijection from $A$ to $B$? – Train Heartnet Nov 12 '13 at 08:54
  • @JobinIdiculla: that is almost the definition of $|A|$. At most, there is a third set $C$ with the same cardinality, so with a bijection from $A$ to $C$ and a bijection from $B$ to $C$, and since they are bijections you can combine them into a single bijection. – Henry Nov 12 '13 at 10:24
  • Oh, I understand now. Just a related question: Since I'm interested in mapping from $\mathbb N^\mathbb N$ to an interval $(a,b)$ in $\mathbb R$, in the example that you provided, I would have to prove that each sequence produces a different real number contained in that interval. Now, I know that for each sequence of natural numbers, a continued fraction gives a unique irrational number. But how do I tweak the continued fraction, to restrict its output values to a certain interval? – Train Heartnet Nov 13 '13 at 07:15
  • @JobinIdiculla: If you have numbers $x \in (0,\infty)$ then $\dfrac{ax+b}{x+1} \in (a,b)$. That does not deal with the issue of rational values. – Henry Nov 13 '13 at 08:30