3

Consider the structure of a rooted tree independent of its underlying set, (i.e. in the sense of trees as combinatorial species). I know a number of ways which we can encode any such tree in natural numbers, but all of them fail to be a bijective, i.e. Some numbers can not be decoded into any sensible tree structure. I would be very pleased to see such concrete correspondence. Thank you.

Hooman
  • 3,647
  • If you can encode each structure as a natural number, and there are infinitely many different structures you can collapse the range of the encoding to have a bijection with $\mathbb N$. – Asaf Karagila Nov 11 '12 at 23:49

2 Answers2

6

See Y. Abe, Tree representation of positive integers, Applied Mathematics Letters Volume 7, Issue 1, January 1994, Page 57:

Abstract. We define a natural one-to-one correspondence between the set of positive integers and the set of equivalence classes of isomorphic rooted trees. By this correspondence, one can translate number-theoric concepts or quantities of a positive integer into graph-theoretic ones of a rooted tree. For example, a positive integer is prime if and only if the corresponding rooted tree is planted, that is, the valency of the root is equal to $1$.

  • 1
    Note that this result was first discovered by Matula (1968): “A natural rooted tree enumeration by prime factorization”. – Arya McCarthy Apr 01 '21 at 14:19
3

See also A new bijection between natural numbers and rooted trees (Cappello).

Guillermo Mosse
  • 1,457
  • 1
    Repeating my comment above: this result was first discovered by Matula (1968): “A natural rooted tree enumeration by prime factorization”, predating Cappello by 20 years. – Arya McCarthy Apr 01 '21 at 14:19