1

In the case of the quotient in the polynomial ring of one variable, there are very good answers here How to deal with polynomial quotient rings

I am asking how we can compute the quotient ring of $K[x,y]$ ( $K$ is a field ) In the following cases :

  1. $K[x,y]/(f) $ where $f=f_1^{k_1} \cdot \dotsc \cdot f_n^{k_n}$ with irreducible $f_i$,
    do we still have this property $K[x,y]/(f) \cong \prod_{i=1}^{n} K[x,y]/(f_i^{k_i})$?

  2. $K[x,y]/(f) $ where $f$ is irreducible in $K[x,y]$

  3. For example what is the quotient ring of $\mathbb{C}[x,y]/(g) $ where $g=x^{p-1}+x^{p-2}+....+x+1-y$ i think that it may be isomoprphic to $\mathbb{C}[x,x^{p-1}+x^{p-2}+....+x+1]=\mathbb{C}[x]$ am i right?

1 Answers1

2

The answer to question $1$ is no. Consider for example the ring $F[x,y]\big/(xy)$, for any field $F$. Then the natural map to $\left[F[x,y]\big/(x)\right]\times\left[F[x,y]\big/(y)\right]$ is not an isomorphism, as it does not take any element of $F[x,y]\big/(xy)$ to $(\overline{0},\overline{1})$ (can you see why?). However, the result does hold in the one-variable case, since $F[x]$ is a PID and hence a Bézout domain.

(Edit: To answer your question below, one can completely characterize the polynomials $f=f_1^{k_1}\dots f_n^{k_n}$ such that $F[x,y]\big/(f)\cong\prod_{i=1}^n F[x,y]\big/(f_i^{k_i})$, where the $f_i$ are non-associates and irreducible, as follows: the isomorphism holds if and only if $(f_i)+(f_j)=R$ for every $i\neq j$. For one direction, we may use the Chinese Remainder Theorem; since $F[x,y]$ is a UFD, $(f)=\bigcap_{i=1}^n(f_i^{k_i})$, and, if $(f_i)+(f_j)=R$ for every $i\neq j$, we have that the $(f_i^{k_i})$ are pairwise coprime, as needed. For the other direction, suppose $(f_i)+(f_j)\neq R$. Then no element of $F[x,y]\big/(f)$ maps to the element of $\prod_{i=1}^n F[x,y]\big/(f_i^{k_i})$ with a $\overline{1}$ in the $j$-th position and $\overline{0}$ everywhere else. Indeed, if $p$ were such an element, $p$ would be divisible by $f_i$ and congruent to $1$ mod $f_j^{k_j}$, and hence we would have $1\in (f_i)+(f_j)$, a contradiction. Thus $f$ induces the desired isomorphism if and only if $(f_i)+(f_j)=R$ for every $i\neq j$, and this gives a complete characterization.)

The answer to question $2$ depends very much on the polynomial $f$, and it's hard to say anything specific without more details. What kind of results are you looking for? For instance, we know that the ring $F[x,y]\big/(f)$ will be an integral domain of Krull dimension $1$, but I'm not sure if this is the kind of thing you're looking for.

The answer to question $3$ is yes. Indeed, we can say something more general; let $F$ be any field, and let $p\in F[x]$ be any polynomial, and consider the map $f:F[x,y]\to F[x]$ defined by $x\mapsto x$ and $y\mapsto p$. Clearly $f$ is surjective, and we claim $\ker f=(p-y)$. The $\supseteq$ inclusion is clear, and for the $\subseteq$ inclusion suppose $f(q)=0$ for some $q\in F[x,y]$. Then in particular $q(x,p)=0\in F[x]\subseteq F[x,y]$. On the other hand, replacing every instance of $\overline{y}$ in $\overline{q(x,y)}$ by $\overline{p}=\overline{y}$ gives $$\overline{q(x,y)}=\overline{q(x,p)}=\overline{0}\in F[x,y]\big/(p-y),$$ and so $q(x,y)\in (p-y)$, as desired. Thus, by the first isomorphism theorem, $f$ induces the desired isomorphism $F[x,y]\big/(p-y)\cong F[x]$.

  • i ment in question 2 any kind of quotient of general polynomial exactly as the example you wrote ,can you give any other example like you wrote before (any other type of polynomial – Aster Phoenix Jan 05 '21 at 18:31
  • about question1 you said that the answer is no ,but do you mean that it never happen or there are some case when it holds .but it is an application of the CRT – Aster Phoenix Jan 05 '21 at 18:38
  • 1
    hi @Abdallahchaibeddrraa; I'm in a meeting right now, but will write up an answer to your questions later today. in particular, there is a very clean characterization of the polynomials $f$ for which question $1$ does hold, which I'll write up – Atticus Stonestrom Jan 05 '21 at 19:23
  • take your time ,please can you just explaine more argement of the inculsion you proved it ? – Aster Phoenix Jan 05 '21 at 20:02
  • 1
    hi @Abdallahchaibeddrraa; I've written up a complete answer to question $1$ of your post, hopefully it helps. for your question here, which inclusion do you mean? the one $\ker f\supseteq (p-y)$? feel free to let me know any questions you have about the argument :) – Atticus Stonestrom Jan 06 '21 at 06:46
  • i mean this inclusion $kerf \subset (p-y) $ – Aster Phoenix Jan 06 '21 at 08:20
  • please just this , in the case of $K[x]/(f)$ do we still need the condition that $(f_{i})+(f_{j})=R$ , $i \neq j$ or it does't matter it always hold(i mean the isomorphim) ? – Aster Phoenix Jan 06 '21 at 08:36
  • This is my attempt if it is right it is ok. $ker f=$ $ q(x, y) \in F[x, y] $s, t $f(q(x, y))=0 $i, e $q(x, p)=0 $thus all polynomial whose have $p-y$ as a factor in their decomposition so $ ker f \subset (p-y) $ – Aster Phoenix Jan 06 '21 at 08:57