Let $k$ be a commutative ring with unit and let $(A,\varepsilon)$ be a (not necessarily commutative) augmented, finitely generated $k$-algebra with augmentation ideal $I$.

If $\mu(A)$ denotes the minimal number of algebra generators of $A$ and $\mu(I)$ the minimal number of generators of $I$ as ideal, we have $\mu(I) \le \mu(A)$.

(For, if $A$ is generated as algebra by $x_1,...,x_n$, then it's also generated by $y_i := x_i - \varepsilon(x_i)\cdot 1\in I$. If $y$ is any element in $I$, write $y = a + f\langle y_1,...,y_n\rangle$ where $a \in k$ and $f$ is a polynomial in the (noncommuting) vaiables $y_1,..,y_n$ having no constant term. Hence $0=\varepsilon(y)=\varepsilon(a)=a$, i.e. $y = f\langle y_1,...,y_n\rangle$ is in the ideal generated by $(y_1,...,y_n)$. This shows $I=(y_1,...,y_n)$ and hence $\mu(I) \le n$)

I wonder what can be said in the case of group algebras:

Question: Does for finitely generated groups the equality $\mu(\mathbb{Z}[G]) = \mu(I)$ hold ?

Note: If $\mu(G)$ denotes the minimal number of generators of the group $G$, we have for finite $G$: $$\mu(I) \le \mu(\mathbb{Z}[G]) \le \mu(G).$$ So the equality would follow from $\mu(I)=\mu(G)$ but I don't know if the latter is true.


The paper The presentation rank of a direct product of finite groups by Cossey, Gruenberg and Kovacs shows that the difference between the minimal number of generators of $G$ and $I$ can be as big as you like for finite groups.

  • $\begingroup$ Great! Thank you. The paper also let's construct simple explicit examples: Let $H\neq 1$ be a finite solvable group and let $G = H^k$ be the direct product. Then $\mu(I_H)=\mu(H)$ (p. 597 and Prop. 1) and $\mu(I_G)=\mu(I_H)$ (Theorem 2). Hence $\mu(\mathbb{Z}[G]) = k\cdot \mu(\mathbb{Z}[H])=k\cdot \mu(I_H) = k\cdot \mu(I_G) > \mu(I_G)$ for $k> 1$. $\endgroup$ Sep 16 '15 at 22:51
  • $\begingroup$ BTW: The paper gives a cohomological formula for $\mu(I)$. Do you know if there is also some kind of formula/characterization for $\mu(\mathbb{Z}[G])$ ? $\endgroup$ Sep 16 '15 at 23:34
  • $\begingroup$ I don't think there is any such formula $\endgroup$ Sep 16 '15 at 23:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.