Discrete derived categories. I: Homomorphisms, autoequivalences and t-structures (Q509983): Difference between revisions

In this article, the authors considered connected derived discrete algebras (in the sense of Vossiek) of finite global dimension defined over an algebraically closed field $k$ that is not derived equivalent to a finite dimensional hereditary algebra of Dynkin type. It's known that up to derived Morita equivalence, such an algebra $\Lambda$ can be realized as the path algebra $\Lambda(r,n,m)$ of the quiver with relations given in Fig. 1 for some values of $r,n,m$ with $n>r$. The authors investigated several key aspects of the bounded derived categry $\mathcal{D}^b(\Lambda)$ of finite dimensional right modules over $\Lambda$: the structure of homomomorphism spaces, the autoequivalence group of $\mathcal{D}^b(\Lambda)$, and bounded t-structures and bounded co-t-structures on $\mathcal{D}^b(\Lambda)$. \par To get the structure of homomomorphism spaces in $\mathcal{D}^b(\Lambda)$, a basic knowledge is the AR quiver of $\mathcal{D}^b(\Lambda)$ which has $3r$ components and can be grouped into three groups $\{\mathcal{X}^0, \dots, \mathcal{X}^{r-1}\},\{\mathcal{Y}^0,\dots,\mathcal{Y}^{r-1}\}, \{\mathcal{Z}^0,\dots,\mathcal{Z}^{r-1}\}$ (see Theorem B of [Classification of discrete derived algebras]). Denote $\mathcal{X}=\mathrm{add}\bigcup_i \mathcal{X}^i,\mathcal{Y}=\mathrm{add}\bigcup_i\mathcal{Y}^i, \mathcal{Z}=\mathrm{add}\bigcup_i\mathcal{Z}^i$. In \S3, the authors give a combinatorial description via the AR quiver of which indecomposable objects admit nontrivial homomomorphisms between them, which they called ``Hom-hammocks''. In \S6, the authors proved a strong result that the dimensions of homomomorphism spaces between indecomposables in $\mathcal{D}^b(\Lambda)$ have a common bound. This captured in a way the ``smallness'' of discrete derived categories. \par To calculate the autoequivalence group of $\mathcal{D}^b(\Lambda)$, the authors established an autoequivelence of $\mathcal{D}^b(\Lambda)$ given by an exceptional cycle (see Definition 4.2), in a way similar to estabilishing the spherical functor given by a spherical object. Such an equivalence actually exists for any exceptional cycle in a $k$-linear indecomposable algebraic triangulated category of finite type with Serre duality (see Theorem 4.5). Then in Theorem 5.7, they calculated the group of autoequivalences of $\mathcal{D}^b(\Lambda$, which in fact is generated by $T_\mathcal{X}, T_{\mathcal{Y}},$ the translation functor $\Sigma$ and the outer automorphisms of $\Lambda$ (actually just nonzero scalars), where $T_{\mathcal{X}}$ and $T_{\mathcal{Y}}$ are respectively the twist functors associated to the exceptional cycles $E_*$ and $F_*$ defined before Lemma 5.1. \par To classify bounded t-structures and bounded co-t-structures on $\mathcal{D}^b(\Lambda)$, the authors first observed that the heart of any bounded t-structure on $\mathcal{D}^b(\Lambda)$ is a length category with finitely many isoclasses of indecomposables (Proposition 7.1). Then by virtue of the bijections established by \textit{S. Koenig} and \textit{D. Yang} in their paper [Doc. Math. 19, 403--438 (2014; Zbl 1350.16010)], it sufficies to classify equivalence classes of silting objects in $\mathcal{D}^b(\Lambda)$. The authors showed that any silting subcategory of $\mathcal{D}^b(\Lambda)$ contain an indecomposable object $Z$ from some $\mathcal{Z}$ component (Corollary 7.11) and that $\mathrm{thick}_{\mathcal{D}^b(\Lambda)}(Z)^{\perp}\simeq \mathcal{D}^b(kA_{n+m-1})$ (Proposition 7.6). In particular, any silting subcategory coincides with $\mathrm{add}(M)$ for some silting object $M$ in $\mathcal{D}^b(\Lambda)$. Using Aihara and Iyama's silting reduction, this allows the authors to establish the bijection between silting sucategories of $\mathcal{D}^b(\Lambda)$ and pairs $(Z, \mathbf{N})$, where $Z$ is an indecomposable object in $\mathcal{Z}$ and $\mathbf{N}$ is a silting subcategory of $\mathcal{D}^b(kA_{n+m-1})$ satisfying certain property (to get rid of redundancy). \par A detailed example for $\Lambda(2,3.1)$ is given in \S8.