Tilting classes over wild hereditary algebras. (Q2386055)

Generalizing the notion of a tilting module in the representation theory of finite-dimensional algebras, one defines the notion of a tilting module over a ring as follows: Let \(T\) be a module over a ring \(A\). \(T\) is called a tilting module if the projective dimension of \(T\) is at most \(1\), \(\text{Ext}^1_A(T,T^{(I)})=0\) for any set \(I\), and there is a short exact sequence \(0\to T_0\to T_1\to 0\), where \(T_i\) is a direct summand in a direct sum of copies of \(T\). If \(T\) is a tilting module then the class \(\text{Gen}(T)\) of all homomorphic images of direct sums of copies of \(T\) is a torsion class in \(A\)-Mod, and \(\text{Gen}(T)\) is the kernel \(T^\perp\) of the functor \(\text{Ext}_A^1(T,-)\). This class of modules in \(A\)-Mod is called the tilting class generated by \(T\). An interesting result of Angeleri-Hügel, Tonolo and Trlifaj says that for any set \(\mathcal S\) of finitely presented modules of projective dimension at most \(1\) the class \({\mathcal S}^\perp\) is always a tilting class, that is, there is always a tilting module \(T\) (that may be infinitely generated) such that \({\mathcal S}^\perp=T^\perp\). In the paper under review the authors consider the tilting classes over a connected wild hereditary algebra. The main question considered here is: Given a set of finitely generated modules \(\mathcal S\), when is the tilting class \({\mathcal S}^\perp={\mathcal F}^\perp\) for a finitely generated tilting module \(F\)? If such a tilting module does not exist, how to construct explicitly the infinitely generated tilting module \(F\)? The authors answer the question completely for \(\mathcal S\) a class of preprojective modules or a class of preinjective modules. Also some results are given for classes consisting of regular modules.