I've come across an interesting connection between irreducible polynomials and polynomial constraints. For example, consider the basic quadratic:

$$af^2 + bf + c = 0$$

If we're working in a ring, and $f$ is assumed to be irreducible, then $c$ must be some multiple of $f$:

$$c = mf$$

Furthermore, substituting this back into the original equation, an additional constraint is discovered for $b$:

$$b + m = qf$$

Substituting this back into the quadratic equation, we find that $a=-q$.

So we now have a system of equations:

$$af^2 + bf +c = 0$$ $$c - mf = 0$$ $$b + m + af = 0$$

and I'm starting to wonder if I can translate my assumption of irreducibility into polynomial constraints that can be handled by a Groebner basis calculation.

For example, if the ring is $\mathbb{Z}$, and $f^2-6f-7=0$ is solved by 7 (irreducible), then $c$ is a multiple of 7, its multiple is -1, and $b$ (-6) plus that multiple is also a multiple of 7 (again -1), and $a$ is the negative of that multiple.

I'm working in a more complex system where (for example) $dx(nd)$ has to be a multiple of some irreducible $f$ with a zero $x$-derivative, so I'm led to conclude that $nd$ must have the form $fq+c$ where $c$ has a zero $x$-derivative. Now I'm thinking of adding $nd=fq+c$ and $c_x=0$ to my system and doing a differential Groebner basis calculation (maybe).

Can I find an exhaustive list of polynomial constraints triggered by the assumption of irreducibility? Exhaustive in the sense of a finite basis for an ideal.

Has anyone seen anything like this before? Any references?

  • $\begingroup$ What do you mean by a constraint? Can you give a definition, and then formulate a short self-contained question. As it is phrased, it is not clear what you are asking. In addition: 1) You appear to divide by f in the step starting with "Furthemore". The element f might be not invertible. 2) Then you start talk about Groebner bases --- so, what kind of ring are you in? $\endgroup$ – Boris Bukh Sep 4 '15 at 19:02
  • $\begingroup$ @Boris Bukh: A constraint is an equation like $c=mf$, an additional polynomial that wouldn't otherwise be implied (though $c=mf$ isn't a good example, since it's implied even without irreducibility). My question is: Can anyone provide a reference that expounds on this theory? 1) I don't divide by $f$, since it factors out, but I do assume that it isn't a zero-divisor, and 2) Let's say I'm in Q[x], for the first example, or Q[x,t] for the second one $\endgroup$ – Brent Baccala Sep 4 '15 at 20:06
  • $\begingroup$ I've posted a follow-up question, maybe a better one, here: mathoverflow.net/questions/217402/… $\endgroup$ – Brent Baccala Sep 4 '15 at 20:26