partial fractions in Euclidean domains


This entry states and proves the existence of partial fraction decompositions on an Euclidean domain.

In the following, we use ν to denote the Euclidean valuation function of an Euclidean domain E, with the convention that ν⁢(0)=-∞.

For a gentle introduction:

  1. 1.

    See partial fractions of fractional numbers (http://planetmath.org/PartialFractions) for the case when E consists of the integers and ν⁢(k)=|k| for k≠0.

  2. 2.

    See partial fractions of expressions for the case when E consists of polynomialsMathworldPlanetmathPlanetmathPlanetmath over the complex field, with ν⁢(p) being the degree of the polynomial p.

  3. 3.

    See partial fractions for polynomials for the case when E is the ring of polynomials over any field, and ν is the degree of polynomials.

Theorem 1.

Let p, q1≠0 and q2≠0 be elements of an Euclidean domain E, with q1 and q2 be relatively prime. Then there exist α1 and α2 in E such that

pq1⁢q2=α1q1+α2q2.
Proof.

By the Euclidean algorithmMathworldPlanetmath, we can obtain elements s1 and s2 in E such that

1=s1⁢q1+s2⁢q2.

Then

pq1⁢q2=p⁢s2q1+p⁢s1q2,

so we can take α1=p⁢s2 and α2=p⁢s1. ∎

Theorem 2.

Let p and q≠0 be elements of an Euclidean domain E, and n be any positive integer. Then there exist elements α1,…,αn,β in E such that

pqn=β+α1q+α2q2+⋯+αnqn,ν⁢(αj)<ν⁢(q).
Proof.

Let r0=p. Iterating through k=1,…,n in order, using the division algorithmPlanetmathPlanetmath, we can find elements rk and sk such that

rk-1=rk⁢q+sk,ν⁢(sk)<ν⁢(q).

Then

p=r0 =r1⁢q+s1
=(r2⁢q+s2)⁢q+s1
=…
=rn⁢qn+sn⁢qn-1+sn-1⁢qn-2+⋯+s2⁢q+s1
pqn =rn+snq+sn-1q2+⋯+s2qn-1+s1qn.

So set β=rn and αj=sn-j+1. ∎

Theorem 3.

Let p and q≠0 be elements of an Euclidean domain E. Let q=ϕ1n1⁢ϕ2n2⁢⋯⁢ϕknk be a factorization of q to prime factorsMathworldPlanetmathPlanetmath ϕi. Then there exist elements αi⁢j,β in E such that

pq=β+∑i=1k∑j=1niαi⁢jϕij,ν⁢(αi⁢j)<ν⁢(ϕi).
Proof.

Apply Theorem 1 inductively to obtain elements si in E such that

pq=∑i=1ksiϕini

(the factors ϕi are relatively prime). Then apply Theorem 2 to obtain elements αi⁢j and βi in E such that

siϕini=βi+∑j=1niαi⁢jϕij

with ν⁢(αi⁢j)<ν⁢(ϕi). Take β=β1+⋯+βk. ∎

Title partial fractions in Euclidean domains
Canonical name PartialFractionsInEuclideanDomains
Date of creation 2013-03-22 15:40:18
Last modified on 2013-03-22 15:40:18
Owner stevecheng (10074)
Last modified by stevecheng (10074)
Numerical id 4
Author stevecheng (10074)
Entry type Result
Classification msc 13F07
Synonym partial fraction decomposition in Euclidean domains