Parametrizations of infinite biconvex sets in affine root systems. (Q2493356): Difference between revisions

Let \(\Delta\) be the root system of a Kac-Moody Lie algebra \(\mathfrak g\), \(\Delta_+\) the set of all positive roots relative to the root basis \(\Pi=\{\alpha_i\mid i\in{\mathbf I}\}\), and \(W=\langle s_i\mid i\in{\mathbf I}\rangle\) the Weyl group of \(\mathfrak g\), where \(s_i\) is the reflection associated with \(\alpha_i\). We call a subset \(B\subset\Delta_+\) a biconvex set if it satisfies the following conditions: (i) \(\beta,\gamma\in B\), \(\beta+\gamma\in\Delta_+\Rightarrow\beta+\gamma\in B\); (ii) \(\beta,\gamma\in\Delta_+\setminus B\), \(\beta+\gamma\in\Delta_+\Rightarrow\beta+\gamma\in\Delta_+\setminus B\). In this paper, the author investigates in detail relationships between infinite reduced words and infinite real biconvex sets in the case where \(\mathfrak g\) is an arbitrary untwisted affine Lie algebra. The study is applied to the classification of ``convex orders'' on \(\Delta_+\), which is indispensable to construct ``convex bases'' of Poincaré-Birkhoff-Witt type of the strictly upper-triangular subalgebra \(U^+_q\) of the quantized universal enveloping algebra \(U_q(\mathfrak g)\).