17
$\begingroup$

With the exception of finite fields and "trivial examples", which rings $R$ are such that Quillen's algebraic $K$ groups $K_i(R)$ are completely known for all $i\geq 0$?

Here, by "trivial examples" I mean for instance examples like $R = \mathbb{F}_q[t]$, or other examples that can easily be deduced from Quillen's computation of $K_i(\mathbb{F}_q)$, the zero ring, etc.

I believe there are no others but I'm not an expert (just curious) and I am having trouble deducing this from the literature.

For instance the $K$-groups of an algebraically closed field are divisible but (correct me if I'm wrong) the uniquely divisible part has not been determined.

$\endgroup$
11
$\begingroup$

For commutative rings, all the K-theory computations I am aware of are discussed in Chapter VI of Weibel's $K$-book (actually with the exception of finite fields).

Function rings: I would say that for smooth affine curves over finite fields, all the $K$-groups have been computed "essentially": Theorems VI.6.4 and VI.6.7 of the $K$-book completely describe the $K$-theory of smooth projective curves over finite fields. If $\overline{C}$ is a smooth projective curve over $\mathbb{F}_q$ which has an $\mathbb{F}_q$-rational point $P$, then you can split off one summand $K_\bullet(\mathbb{F}_q)$ to get a complete description of $K_\bullet(\mathbb{F}_q[\overline{C}\setminus\{P\}])$. For more general smooth affine curves, I believe one can work out the localization sequence and compute $K$-theory of those as well.

I do not think that there are many complete results for higher-dimensional varieties over finite fields, the basic obstruction would be Bass's conjecture or Parshin's conjecture (i.e., finite generation of K-theory of finitely generated commutative $\mathbb{F}_q$-algebras is expected but not yet proved, so all torsion computations leave a uniquely divisible indeterminacy).

Number rings: Sections VI.8 and VI.9 of the $K$-book have fairly complete calculations for $K$-theory of rings of $S$-integers in number fields (if you accept computations in terms of étale cohomology). These computations all follow from the solution of the Bloch-Kato conjecture which implied the Quillen-Lichtenbaum conjecture which enabled $K$-theory computations in terms of étale cohomology (roughly). The complete calculation of the $K_\bullet(\mathbb{Z})$ still depends on the Kummer-Vandiver conjecture in degrees $0$ mod $4$.

Further results: Everything else that I know is for special coefficients. With torsion coefficients, we know $K$-theory of algebraically closed fields, as you mentioned (Chapter VI.1). With torsion coefficients prime to the residue characteristic the K-theory of local fields and their dvrs can be computed using rigidity methods. For torsion coefficients at the residue characteristic, the $K$-theory of local fields is computed using trace methods (Chapter VI.7). Check out the $K$-book for more information. With a grain of salt, I would say that all other known computations of $K$-theory are deduced from the above via projective bundle formula, localization sequence or some such method.


Non-commutative rings: I do not know so much about those. The Farrell-Jones conjecture has been proved for a couple of classes of groups (hyperbolic groups, CAT(0)-groups), see papers of Bartels, Lück, Reich. This is a major step toward computing $K$-theory of group rings of such groups. However, even with the Farrell-Jones conjecture, it still remains to compute the source of the assembly map, i.e., the equivariant homology. I do not know if the solution to the Farrell-Jones conjecture has been used in completely computing $K$-theory of group rings, you'd have to ask an expert in the field.

Of course, in the non-commutative setting there are also the flasque rings whose $K$-theory is trivial but I guess these qualify as "trivial examples".

| cite | improve this answer | |
$\endgroup$

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.