4

Given a ring $R$ and ideals $A,C$ suppose we have $A + B' =A + B = C.$ I was wondering then what can we say about relation between $B$ and $B'$.

Clearly, $B$ may not equal $B'$, but can we say something? Does it follow that $B= B' + D$ where $D$ is an ideal contained in $A$? Thanks!

Tom Mosher
  • 1,349
  • 8
  • 12
  • Also, I was wondering, is there a way to make subtraction of ideals well defined in some sense? – Tom Mosher Jul 15 '13 at 19:33
  • In general no. But suppose they are ideals in a polynomial ring and you fix a monomial order. Then you can take the generators of $A$ followed by the generators of $C$. Apply Buchberger algorithm to get a Groebner basis of $C$. The generators added to the basis after the generators of $C$ begin to be used give you a $B$. It is unique after making all these choices. – OR. Jul 15 '13 at 19:46
  • @RGB Thank you for this information! Just from curiosity could you do something similar if you are dealing with ring of algebraic integers? – Tom Mosher Jul 16 '13 at 15:22
  • Yes, but the answer is kind of trivial. Suppose $(a)+J=(b)$. The above translates here to find the greatest common divisor of $a$ and $b$, but that is $b$. So you get $J=(b)$. Notice that $b$ divides $a$ for the problem to have a solution. – OR. Jul 16 '13 at 15:30
  • Right. I see. Thank you! – Tom Mosher Jul 16 '13 at 15:31

1 Answers1

2

In $\mathbb Z$, $(p)+(q)=(p)+(r)$ for all distinct primes $p,q,r$.

More generally, if $M$ is any maximal ideal in a ring $R$ and $I,J$ are any two ideals that are not contained in $M$, then $M+I=M+J$. So there is enormous freedom in constructing the situation you describe with no particular conditions on the ideals.

Ittay Weiss
  • 79,840
  • 7
  • 141
  • 236