6
$\begingroup$

Let $G$ be your favorite complex reductive algebraic group, and consider its (algebraic) loop group $G((t))$. A role very similar to the Borel $B\subset G$ is played in the loop group by the Iwahori $$I=\{g\in G[[t]] : g\in B \pmod{t}\}.$$

In particular, $$G((t))=\bigsqcup_{w\in \widehat{W}} IwI$$ for $\widehat W$ the affine Weyl group of $G$. Now, $G((t))$ is an affine ind-scheme, so for any $w\in \widehat W$, it makes sense to talk about the ideal defined by $\overline{IwI}$.

In the comparable case of $\overline{BwB}$, this ideal is relatively easily described: the generalized minors of the fundamental representations which vanish on $\overline{BwB}$ generate the ideal.

Is there any comparable set of representations for $G((t))$?

You would make me a very happy man if you were to say the answer was just the fundamental representations of $G$ with $((t))$ adjoined. I should emphasize that I know this is true as sets; I am asking about ideals, which is notably stronger.

I should note that I would be OK with replacing $I$ with the parahoric $G[[t]]$, which is strictly weaker.

$\endgroup$
2
  • $\begingroup$ You probably already know this, but For G=GL_n the closures of the double Iwahori cells can be described by a set of inequalities of the form valuation of minor greater than or equal to something. This will follow from an argument using the proof idea of Smith normal form + Gaussian elimination. For general G, I expect a similar description with generalised minors. Tempting is to embed G into some GL_n and compare the IWI decompositions for G and GL_n. Maybe this will give a possibly classification dependent proof (but there should be a better way). $\endgroup$ Commented Oct 29, 2011 at 20:03
  • $\begingroup$ It's very easy to see that the valuations thing describes the subsets correctly in all cases (OK, maybe not easy, but not hard). I'm interested in the scheme structure (or more concretely, the ideal one gets) and whether it is reduced (radical). That's much harder to work out. $\endgroup$
    – Ben Webster
    Commented Oct 29, 2011 at 21:36

0

You must log in to answer this question.