# Thread: valuation ring, discrete valuation

1. ## valuation ring, discrete valuation

Background: Let $\displaystyle K$ be a field. A discrete valuation on $\displaystyle K$ is a function $\displaystyle \nu$ on $\displaystyle K$ is a function $\displaystyle \nu$ $\displaystyle : K^* \rightarrow \mathbb{Z}$ such that
(i) $\displaystyle \nu(ab)=\nu(a)+\nu(b)$ for all $\displaystyle a, b \in K^*.$
(ii) $\displaystyle \nu$ is surjective
(iii) For all $\displaystyle x, y$ $\displaystyle \in K^*$ such that $\displaystyle y \neq -x$, $\displaystyle \nu(x+y)\geq$ min$\displaystyle \{ \nu(x), \nu(y) \}$. The set $\displaystyle R=$$\displaystyle \{ x \in K^* | \nu(x) \geq 0 \} \cup \{ 0 \} is called the valuation ring of \displaystyle \nu. Prove: (a) Prove that \displaystyle R is a subring of \displaystyle K. (b) Prove that for each nonzero element \displaystyle x \in K, \displaystyle x \in R, or \displaystyle x^{-1} \in R. (c) Prove that \displaystyle x \in R is a unit if and only if \displaystyle \nu(x)=0. (d) Now let \displaystyle K = \mathbb{Q} and let \displaystyle p be a prime. Given a nonzero element \displaystyle x \in \mathbb{Q}, write \displaystyle x=p^e \cdot \frac{c}{d} where \displaystyle e is an integer and \displaystyle p divides neither \displaystyle c nor \displaystyle d; then define \displaystyle \nu_p=e. Prove \displaystyle e is uniquely defined and that \displaystyle \nu_p is a discrete valuation on \displaystyle \mathbb{Q}. 2. Originally Posted by Erdos32212 Background: Let \displaystyle K be a field. A discrete valuation on \displaystyle K is a function \displaystyle \nu on \displaystyle K is a function \displaystyle \nu \displaystyle : K^* \rightarrow \mathbb{Z} such that (i) \displaystyle \nu(ab)=\nu(a)+\nu(b) for all \displaystyle a, b \in K^*. (ii) \displaystyle \nu is surjective (iii) For all \displaystyle x, y \displaystyle \in K^* such that \displaystyle y \neq -x, \displaystyle \nu(x+y)\geq min\displaystyle \{ \nu(x), \nu(y) \}. The set \displaystyle R=$$\displaystyle \{ x \in K^* | \nu(x) \geq 0 \} \cup \{ 0 \}$ is called the valuation ring of $\displaystyle \nu$.

Prove:
(a) Prove that $\displaystyle R$ is a subring of $\displaystyle K$.
(b) Prove that for each nonzero element $\displaystyle x \in K$, $\displaystyle x \in R$, or $\displaystyle x^{-1} \in R$.
(c) Prove that $\displaystyle x \in R$ is a unit if and only if $\displaystyle \nu(x)=0$.
(d) Now let $\displaystyle K = \mathbb{Q}$ and let $\displaystyle p$ be a prime. Given a nonzero element $\displaystyle x \in \mathbb{Q}$, write $\displaystyle x=p^e \cdot \frac{c}{d}$ where $\displaystyle e$ is an integer and $\displaystyle p$ divides neither $\displaystyle c$ nor $\displaystyle d$; then define $\displaystyle \nu_p=e$. Prove $\displaystyle e$ is uniquely defined and that $\displaystyle \nu_p$ is a discrete valuation on $\displaystyle \mathbb{Q}.$
(a) you only need to show that R is closed under addition and multiplication: let $\displaystyle x,y \in R,$ and put $\displaystyle u=x+y.$ if u = 0, then u obviously is in R. if $\displaystyle u \neq 0,$ then by (iii): $\displaystyle \nu(u) \geq \min \{\nu(x), \nu(y) \} \geq 0,$

because $\displaystyle \nu(x) \geq 0, \ \nu(y) \geq 0.$ hence $\displaystyle u \in R.$ to prove that $\displaystyle xy \in R,$ by (i): $\displaystyle \nu(xy) = \nu(x) + \nu(y) \geq 0.$ so $\displaystyle xy \in R.$

(b) first note that by (i): $\displaystyle \nu(1)=\nu(1) + \nu(1).$ thus: $\displaystyle \nu(1)=0.$ now suppose $\displaystyle x \in K.$ if x = 0, then x clearly is in R. so let $\displaystyle x \neq 0.$ suppose $\displaystyle x \notin R, \ x^{-1} \notin R.$ then $\displaystyle \nu(x) < 0, \ \nu(x^{-1}) < 0.$

thus: $\displaystyle 0=\nu(1)=\nu(x) + \nu(x^{-1}) < 0.$ contradiction!

(c) if $\displaystyle x \in R$ is a unit, then $\displaystyle x^{-1} \in R.$ we showed in (b) that $\displaystyle \nu(1)=0.$ hence: $\displaystyle \nu(x) + \nu(x^{-1})=0,$ which is possible only if $\displaystyle \nu(x)=\nu(x^{-1})=0,$ because $\displaystyle \nu(x) \geq 0, \ \nu(x^{-1}) \geq 0.$ now suppose that

$\displaystyle x \in R$ and $\displaystyle \nu(x)=0.$ we want to show that $\displaystyle x^{-1} \in R,$ that is $\displaystyle \nu(x^{-1}) \geq 0.$ well, as we've seen: $\displaystyle \nu(x^{-1}) = -\nu(x)=0.$

(d) it's obvious that $\displaystyle e$ is uniquely defined because of the unique factorization in the ring of integers. it's also trivial that $\displaystyle v_p$ satisfies the properties (i), (ii). the property (iii) is less obvious: suppose

that $\displaystyle x=p^r \frac{a}{b}, \ y=p^s \frac{c}{d},$ with $\displaystyle r \leq s$ and $\displaystyle a,b,c,d$ not divisible by p. so $\displaystyle \nu_p(x)=r, \ \nu_p(y)=s.$ then $\displaystyle x+y=p^r \frac{ad + p^{s-r}bc}{bd}.$ so since $\displaystyle p$ does not divide $\displaystyle bd,$ we have: $\displaystyle \nu_p(x+y) \geq r = \min \{r,s \}.$