Crystal bases and two-sided cells of quantum affine algebras (Q701895)

The authors construct bases of quantum enveloping algebras of affine Kac-Moody algebras and relate them to the global crystal basis of Kashiwara and Lusztig. It is noteworthy that their primary results hold for all affine algebras, both twisted and untwisted, and that their methods have very little of the case-by-case analysis that can bedevil Kac-Moody theory. The authors first construct a particularly nice basis \(\mathcal{B}\) of the positive part \(\mathbb{U}^+\) in the usual triangular decomposition of the quantum enveloping algebra \(\mathbb{U}\) associated to an affine Kac-Moody algebra \(\mathfrak{g}\). This PBW-type basis \(\mathcal{B}\) has a number of interesting properties, including the important one that the transition matrix between \(\mathcal{B}\) and the global crystal basis is upper triangular. One effect of this is that the integrality of the basis is not checked by tedious explicit computations of commutation relations, but via the global crystal basis and the transition matrix. In the simply-laced case, the basis \(\mathcal{B}\) coincides with that of \textit{J. Beck, V. Chari} and \textit{A. Pressley} [''An algebraic characterization of the affine canonical basis''. Duke Math. J. 99, No. 3, 455--487 (1999; Zbl 0964.17013)], but the construction techniques are different. In particular, the authors make powerful use of the extremal weight modules \(V(\lambda)\), \(\lambda \in P\), introduced by Kashiwara and prove a number of conjectures of Kashiwara's on the structure of the global crystal basis \(\mathcal{B}(\lambda)\) of \(V(\lambda)\). The authors give a Peter-Weyl-type decomposition of the global crystal basis \(\mathcal{B}(\tilde \mathbb{U})\) of the modified quantum enveloping algebra. The authors also prove a conjecture of Lusztig's on the two-sided cells of the level-zero part of \(\mathcal{B}(\tilde\mathbb{U})\). The paper is very clearly written and contains a wealth of interesting ideas and results.