2

Let $f(x)$ be a polynomial with complex coefficients such that $\exists n_0 \in \mathbb Z^+$ such that $f(n) \in \mathbb Z , \forall n \ge n_0$, then is it true that $f(n) \in \mathbb Z , \forall n \in \mathbb Z$ ?

user26857
  • 52,094

2 Answers2

1

No. The ring of integer-valued polynomials (polynomials in $\mathbf Q[x]$ with integer values in $\mathbf Z$) is a counter-example. A basis, as a $\mathbf Q$-vector space made up of the generalised binomial coefficients: $$\binom{x}{n}=\frac{x(x-1)\dots(x-n+1)}{n!}.$$

Bernard
  • 175,478
  • 2
    I asked whether $f$ is integer valued for all sufficiently large integers imply $f$ is integer valued for all integers , then how come "integer valued polynomials" is a counter example ?!? –  Apr 02 '15 at 08:41
  • 1
    He who can do more can do less. :o) – Bernard Apr 02 '15 at 08:44
1

Proof by induction on the degree of the polynomial. Trivially true for degree zero. Assume true for degree $n$. Let $f$ have degree $n+1$ and be integral for all sufficiently large integer arguments. Then same is true for $g(x)=f(x+1)-f(x)$, a polynomial of degree $n$. By the induction hypothesis, $g$ is integer-valued for all integer arguments. Then it follows that $f$ is.

Gerry Myerson
  • 179,216