How Do Primary Ideals in a PID Relate to Their Product and Intersection?

  • Thread starter Thread starter Bashyboy
  • Start date Start date
  • Tags Tags
    Pid
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
1 reply · 2K views
Bashyboy
Messages
1,419
Reaction score
5

Homework Statement



Let ##R## be a principal ideal domain. (a) Every proper ideal is a product of ##P_1...P_n## of maximal ideals which are uniquely determined up to an order. (b) ##P## is a primary (not prime) ideal if and only ##P=(p^n)## for some prime ##p \in R## and ##n \in \Bbb{N}##. (c) If ##P_1,...,P_k## are primary ideals such that ##P_i = (p_i^{n_i})## and the ##p_i## are distinct primes, then ##P_1...P_k = P_1 \cap ... \cap P_k##. (d) Every proper ideal in ##R## can be expressed (uniquely up to order) as the intersection of a finite number of primary ideals.

Homework Equations

The Attempt at a Solution



I am currently working on part (c) and could use some help. Here is what I have so far. Since we are taking the ##p_i## to be distinct, I believe we can also take them not be associates of each other, but I'm not entirely certain. First note that ##(p_1^{n_1})...(p_k^{n_k}) = (p_1^{n_1}..p_k^{n_k})##, since we are working in a commutative ring. Let ##x \in (p_1^{n_1}..p_k^{n_k})##. Then ##p_1^{n_1}...p_k^{n_k} \mid x## and therefore ##p_i^{n_i} \mid x## for every ##i## and finally ##x \in (p_i^{n_i})## for every ##i##, proving that ##(p_1^{n_1})...(p_k^{n_k}) \subseteq \bigcap_{i=1}^k (p_i^{n_i})##.

Now we try to prove the other set inclusion. First, since ##\bigcap_{i=1}^k (p_i^{n_i})## is the intersection of ideals, it must also be an ideal, from which we can conclude ##\bigcap_{i=1}^k (p_i^{n_i}) = (y)## for some ##y \in R##. This ##y## must be nonzero and not a unit, which means that there exist primes/irreducibles ##c_i## and integers ##m_i \in \Bbb{N}## such that ##y = c_i^{m_1} ... c_s^{m_s}##, because ##R## is also a UFD. Since ##(y) = \bigcap_{i=1}^k (p_i^{n_i})##, ##c_i^{m_1} ... c_s^{m_s}## must be in the intersection, which implies that, for every ##i##, ##p_i^{n_i} \mid c_i^{m_1} ... c_s^{m_s}## and therefore ##p_i \mid c_i^{m_1} ... c_s^{m_s}##. Because ##p_i## is prime, it follows that ##p_i \mid c_j^{m_j}## for some ##j##, and through induction we get ##p_i \mid c_j##. This means that ##c_j = \ell_{ij} p_i## for some ##\ell_{ij} \in \Bbb{N}##, and since ##c_j## is also irreducible, ##\ell_{ij}## must be a unit, from which it follows ##c_j## and ##p_i## are associates. So, for every ##i##, there exists a ##j## such that ##p_i## and ##c_j## are associates, so ##k \le s## (can we take, WLOG, ##k=s##?) For convenience, relabel the elements so that ##p_i## and ##c_i## are associates...

-----------------------

Okay. At this point, I am unsure of how to proceed. I am trying to show that ##c_i^{m_1} ... c_s^{m_s} \in (p_1^{n_1})...(p_k^{n_k})##, but the path is obscure for some reason, although I believe I am close. I could use some hints.
 
Physics news on Phys.org
I think it is a bit too complicated what you wrote. Let me try to sort it a little bit by only using what you already have. We have to show:
Bashyboy said:
##P_1...P_k = P_1 \cap ... \cap P_k##
I don't think you need the six lines lines for the inclusion ##P_1...P_k \subseteq P_1 \cap ... \cap P_k##
All ##P_i## are ideals, so any product ##q_1\ldots q_k = q_1 \ldots q_{i-1}\cdot q_i \cdot q_{i+1}\ldots q_k \in P_i##, i.e. the inclusion holds.
For the other direction, say ##Q = P_1 \cap ... \cap P_k## which ...
Bashyboy said:
since ##\bigcap_{i=1}^k (p_i^{n_i})## is the intersection of ideals, it must also be an ideal
... and is proper by definition. Now we have
Bashyboy said:
(a) Every proper ideal is a product of ##M_1...M_n## of maximal ideals which are uniquely determined up to an order.
So ##P_1 \ldots P_k \subseteq Q = (y) = M_1 \ldots M_n= (m_1) \ldots (m_n)=(m_1 \ldots m_n)## with primes ##m_i## because maximal ideals are prime and ##R## is a PID. (This follows without any excursions to UFD and irreducibility.)
Bashyboy said:
Since ##(y) = \bigcap_{i=1}^k (p_i^{n_i})##
##p_i\,\vert \,p_i^{n_i}\,\vert \,m_1 \ldots m_n=y\,## so w.l.o.g. ##\,p_i=m_i\,## for all ##1 \leq i \leq k##. Thus ##y=p_1\ldots p_k \cdot m_{k+1}\ldots m_n \cdot r##. Since all ##p_i^{n_i}## still divide ##y##, we can cancel them one by one by repetition of the argument, so ##p_1^{n_1}\ldots p_k^{n_k} \,\vert \, y##.
 
Last edited: