On the double-affine Bruhat order: the $\varepsilon =1$ conjecture and classification of covers in ADE type (Q1740603): Difference between revisions

For any Kac-Moody group $G$, the authors prove that the Bruhat order on the semidirect product $W_{\mathcal{T}} = \mathcal{T} \rtimes W$ of the Weyl group $W$ and the Tits cone $\mathcal{T}$ for $G$ is strictly compatible with a $\mathbb{Z}$-valued length function. In the proof, the inversion sets play an important role. For a positive root $\beta \in G$ and for an integer $n\in \mathbb{Z}$, there is an associated reflection element $s_{\beta[n]}$. In general $s_{\beta[n]} \notin W_{\mathcal{T}}$. For suitable $x\in W_{\mathcal{T}}$ with $xs_{\beta[n]}\in W_{\mathcal{T}}$, they construct an inversion set $\mathrm{Inv}_{x}^{++}(s_{\beta[n]})$ and prove that $xs_{\beta[n]}>x$ if and only if $\mathrm{Inv}_{x}^{++}(s_{\beta[n]})$. In particular, they prove that if $xs_{\beta[n]}>x$ then $l(xs_{\beta[n]})-l(x)=\#(\mathrm{Inv}_{x}^{++}(s_{\beta[n]}))$, where $l:W_{\mathcal{T}}\rightarrow \mathbb{Z} \oplus \mathbb{Z}\varepsilon\rightarrow \mathbb{Z}$ is the composed length function obtained by setting $\varepsilon=1$. \par Then, they consider the question that for $x,y\in W_{\mathcal{T}}$ whether $y$ covers $x$ if and only if $x<y$ and $l(x)=l(y)-1$. They answer this question positively when $G$ is untwisted affine ADE type. For this type, using explicit control of the pairings between roots and coroots, they reduce the problem to a calculation that is essentially the case of affine $\mathrm{SL}_2$, where they verify this situation by explicit computations.