Skip to main content

Section 10.1 Euclidean Domains (EDs)

“Find your domain and serve it to the world.”
―Myles Munroe

Definition 10.1. Euclidean Domain.

A Euclidean domain (ED) is a domain \(R\) together with a norm function \(N: R \to \Z_{\geq 0}\) such that \(N(0) = 0\) and the following property holds: for any two elements \(a, b \in R\) with \(b \ne 0\text{,}\) there are elements \(q\) and \(r\) of \(R\) such that
\begin{equation*} a = qb + r \;\text{and either}\; r = 0 \;\text{or}\; N(r) < N(b). \end{equation*}

Qual Watch.

Stating Definition 10.1 was Part (a) of [cross-reference to target(s) "jan-2020-4" missing or not unique] on the [cross-reference to target(s) "jan-2020" missing or not unique] qualifying exam.
In a Euclidean domain, division with remainder is possible, and the remainder is always smaller than the divisor. This property is similar to the Division Algorithm for integers, which is what makes Euclidean domains so useful in number theory and algebra.
Lets take a look at some examples to deepen our understanding.

Example 10.2. Trivial Norm.

A "degenerate example" is a field \(F\) equipped with the trivial norm \(N(r) = 0\) for all \(r\text{.}\) Given \(a,b \in F\) with \(b \ne 0\text{,}\) we have \(a =(ab^{-1})b + 0\text{.}\) Note that in this example there is no need to include \(r=0\) in the description of the Division Algorithm, since \(b \ne 0\) implies \(|0| < |b|\text{.}\) This is not the case in other examples. Also observe that as we’ve defined remainders they are not unique. For example, in dividing \(13\) by \(5\text{,}\) both
\begin{equation*} 13 = 2 \cdot 5 + 3 \text{ and } 13 = 3 \cdot 5 + (- 2) \end{equation*}
are considered valid. This calculation shows, more generally, that if \(b\) is a unit, then for all \(a\) there exists an equation \(a = bq + r\) with \(r = 0\text{,}\) not matter what norm \(N\) is used. One could make remainders (and hence quotients) unique for \(\Z\) by insisting that remainders always be non-negative, but this is not part of the abstract theory since it doesn’t generalize to all cases well.

Example 10.3. Polynomial Rings over Fields are Euclidean Domains.

The next classical example is \(F[x]\) with \(F\) a field and where we define the norm to be polynomial degree: \(N(f(x)) = \deg(f(x))\) if \(f\neq 0\) and \(N(0)=0\text{.}\) This ring is a Euclidean Domain because of the familiar long division for polynomials, as proved in Theorem 8.63.

Example 10.4. Guassian Integers are a Euclidean Domain.

The ring \(R = \Z[i]\) of Gaussian integers is a Euclidean domain with \(N\) being the usual complex (Euclidean) square norm \(N(a+bi)=a^2+b^2\text{.}\) Let \(\alpha\text{,}\) \(\beta\) \(\in \Z[i]\) and let \(\frac{\alpha}{\beta}=p+qi \in \Q(i)\) (here we use that the fraction field of \(\Z[i]\) is \(\Q(i)\)). Now pick \(s,t\in\Z\) so that \(|p-s|\leq 1/2\) and \(|q-t|\leq 1/2\text{.}\) We have
\begin{equation*} \a=\beta(s+ti)+\beta(p+qi)-\beta(s+ti). \end{equation*}
Set \(q=s+ti\) and set \(r=\beta(p+qi)-\beta(s+ti)=\beta(s+ti-(p+qi))\) and notice that \(q\in \Z[i]\) because \(s,t\in\Z\) and \(r\in \Z[i]\) by closure. If \(r=0\) we’re good, and if \(r\neq 0\) then, using that the complex squared norm is multiplicative as well as the Pythagorean Theorem and the choice for \(s,t\text{,}\) we have
\begin{equation*} N(r)=N(\beta(s+ti-(p+qi)))=N(\beta)N(s+ti-(p+qi))\leq N(\beta)\cdot(1/4+1/4)<N(\beta). \end{equation*}
Thus the norm function \(N\) makes \(\Z[i]\) into a Euclidean domain.

Definition 10.5. GCD.

Given elements \(a, b\text{,}\) not both \(0\text{,}\) of a Euclidean domain \(R\) with Euclidean norm \(N\text{,}\) a \(\gcd\) of \(a\) and \(b\) is an element \(g \in\R\) such that:
  1. \(g | a\) and \(g | b\text{;}\) and
  2. If \(d | a\) and \(d | b\text{,}\) then \(N (d) \leq N (g)\text{.}\)

Remark 10.6.

Note that \(b\mid a\) is equivalent to \(b\in\igen a\text{.}\)