Silting and tilting for weakly symmetric algebras (Q2688941)

Let \(A\) be a finite-dimensional \(k\)-algebra and \(\operatorname{mod}(A)\) be the class of finitely generated modules. Consider a complete system of primitive orthogonal idempotents \(e_1,\dots,e_n\) and define \(P_i:=e_iA\). Recall that \(A\) is weakly symmetric if \(P_i:=D\circ \operatorname{Hom}_A (P_i,A)\) \(\forall i\), where \(D:=\operatorname{Hom}_k (-,k)\) (see [\textit{T. Nakayama} and \textit{C. Nesbitt}, Ann. Math. (2) 39, 659--668 (1938; JFM 64.0081.03)] ). Tilting complexes are a class of objects that play a similar role in the derived category \(D^b (\operatorname{mod}(A))\) to that played by progenerators in Morita theory for module categories (see [\textit{J. Rickard}, J. Lond. Math. Soc., II. Ser. 39, No. 3, 436--456 (1989; Zbl 0642.16034)] ). Given their importance, these objects have been studied by different authors. It is known, for example, that the class of tilting complexes \(\operatorname{tilt}(A)\) has a partial order \(\geq\), defined as \(X\geq Y\) if and only if \(\operatorname{Hom}_{\mathcal{K}^b (\operatorname{proj}(A))} (X,Y[n])\) \(\forall n>0\) (see [\textit{D. Happel} and \textit{L. Unger}, Algebr. Represent. Theory 8, No. 2, 147--156 (2005; Zbl 1110.16011)]); and that there exists a mutation, which consists in the construction of a tilting complex from a given one by substituting a direct summand (see [\textit{C. Riedtmann} and \textit{A. Schofield}, Comment. Math. Helv. 66, No. 1, 70--78 (1991; Zbl 0790.16013)]). Such mutation has different applications, however, it has the disadvantage that not every summand can be replaced by another one to construct a new tilting complex. For this reason, the mutation and the partial order relation defined in \(\operatorname{tilt}(A)\) were extended to the class of silting complexes \(\operatorname{silt}(A)\) in [\textit{T. Aihara} and \textit{O. Iyama}, J. Lond. Math. Soc., II. Ser. 85, No. 3, 633--668 (2012; Zbl 1271.18011)]. In this context, \(A\) is said to be silting-discrete if the set \(\{ T \in \operatorname{silt}(A)\mid P\geq T\geq P[1] \}\) is finite for all \(P\in \operatorname{silt}(A)\). The notion of tilting-discrete is defined similarly (see [\textit{T. Aihara} and \textit{Y. Mizuno}, Algebra Number Theory 11, No. 6, 1287--1315 (2017; Zbl 1412.16008)]). The article under review is motivated by the question of which finite-dimensional algebras are silting-discrete. The main result tells us that if \(A\) is weakly symmetric and tilting-discrete, then it is also silting-discrete and in this case \(\operatorname{silt}(A)=\operatorname{tilt}(A)\). Moreover, the authors give a family of weakly symmetric finite-dimensional algebras with silting complexes that are not tilting.