MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Unlike algebraic K-theory, equivariant K-theory of affine space (over a field $k$) can be quite nontrivial, depending on the action of the group in question. For example, if one takes the standard action of $\mathbb G_m$ on $\mathbb A^n$, then the $\mathbb G_m$-equivariant K-theory of $\mathbb A^n$ can be computed using a localization sequence corresponding to the inclusion of the origin inside $\mathbb A^n$. (In the localization sequence in equivariant K-theory, one uses that the quotient of $\mathbb A^n \backslash \{0\}$ is $\mathbb P^{n-1}$ and uses the projective bundle theorem, for example, to proceed.)

How does one compute $K_i^G(\mathbb A^n)$ in other cases? Is there a general procedure for doing so? For example, take $G = \mathbb G_m \times \cdots \times \mathbb G_m$ (product of $n$ copies) and consider the diagonal action on $\mathbb A^n$. How does one compute $K_i^G(\mathbb A^n)$ in this case? Is it easier to write down the description of $K_0^G(\mathbb A^n)$?

share|cite|improve this question
up vote 7 down vote accepted

In fact, just like in the case of non-equivariant algebraic $K$-theory, there is a homotopy-invariance for equivariant $K$-theory (of nonsingular varieties). Addressing your question specifically, for any action of an algebraic group $G$ on affine space, the answer is: $$ K^G_i({\Bbb A}^n) = R(G) \otimes K_i(k), $$ where $k$ is the ground field and $R(G) = K^G_0(k)$ is the representation ring. If $G$ is a (split) torus, the representation ring is just the group ring of the character group; choosing a basis $x_1,\ldots,x_n$ of characters identifies $R(G)$ with the Laurent polynomial ring ${\Bbb Z}[x_1^\pm,\ldots,x_n^\pm]$.

(The general statement is that when $f\colon X \to Y$ is a $G$-equivariant affine bundle, there is an isomorphism $$f^*\colon G^G_*(Y) \to G^G_*(X),$$ where $G^G_*$ denotes equivariant $K$-theory of coherent sheaves. For nonsingular varieties $X$ one has $G^G_*(X) = K^G_*(X)$. See Theorem 4.1 of Thomason's foundational paper "Algebraic K-theory of group scheme actions".)

share|cite|improve this answer
Is it standard to denote $K$-cohomology with the $*$ as subscript? Please say no. – Allen Knutson Feb 19 '14 at 23:26
@Allen, very regrettably: yes. Worse, the standard convention of writing K-homology as "G" leads to the atrocity I committed above. For better or worse, after spending a little time with the literature I am starting to be convinced there are at least some good reasons for all this... – Dave Anderson Feb 20 '14 at 1:30
@DaveAnderson: Thanks a lot for the answer. I didn't know about Thomason's this paper. – Carl Feb 20 '14 at 7:34
@DaveAnderson: I am now having the following doubt: Consider the standard $\mathbb G_m$ action on $\mathbb A^n$ and consider the localization exact sequence as I had mentioned in the question. The part of the sequence I am worried about goes as follows: $ \cdots \to K_1^G(\mathbb A^n) \to K_1(\mathbb P^{n-1}) \to K_0^G(\{0\}) \to K_0^G(\mathbb A^n) \to K_0(\mathbb P^{n-1}) \to 0$. Now, the groups $K_0^G(\{0\})$ and $K_0^G(\mathbb A^n)$ are the same, however, the map between them in the localization sequence is not an isomorphism. – Carl Feb 20 '14 at 14:19
So the map $K_1^G(\mathbb A^n) \to K_1(\mathbb P^{n-1})$ is not surjective. Is this right? – Carl Feb 20 '14 at 14:23

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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