Blocks of tame representation type and related algebras (Q1188690): Difference between revisions

A finite-dimensional algebra A over an (algebraically closed) field F is called tame if A has infinitely many isomorphism classes of indecomposable modules but, for any positive integer d, there are finitely many F[T]-A-bimodules \(M_ i\), free over F[T], such that almost all indecomposable A-modules of dimension d are of the form \(N\otimes_{F[T]}M_ i\) for some i and some simple F[T]-module N. Now let G be a finite group and B a block of the group algebra FG. It is known (although not proved in this book) that B is tame if and only if char F\(=2\) and the defect groups of B are dihedral, quaternion or semidihedral. Over the past few years the author has published a series of papers on such blocks culminating in a classification up to Morita equivalence. Each Morita equivalence class of finite-dimensional algebras contains a unique basic algebra (this means that all simple modules have dimension 1), and this basic algebra is isomorphic to the path algebra of a (uniquely determined) quiver (recording the possible extensions of simple modules) modulo an ideal of relations. Erdmann's classification theorem lists the possible quivers and generators for the possible ideals of relations. The present book gives a comprehensive account of these results, including a number of new facts and some new proofs of known facts. However, for some of the details, the reader is also referred to the original papers. The major tool used is the Auslander-Reiten quiver of an algebra which records the irreducible maps between indecomposable modules. Suppose now that B is a block of FG with dihedral, quaternion or semidihedral defect group D. In order to deal with the indecomposable FD- modules when D is dihedral the author considers, more generally, special biserial algebras; these can be defined as path algebras of a quiver Q modulo an ideal I of relations where (1) Any vertex of Q is the starting point of at most two arrows; \((1')\) Any vertex of Q is the end point of at most two arrows; (2) For any arrow b in Q, there is at most one arrow c in Q ending at the starting point of b such that bc\(\not\in I\); \((2')\) For any arrow c in Q, there is at most one arrow b starting at the end point of c such that bc\(\not\in I\). The author presents (without proof) a classification of the indecomposable modules for a special biserial algebra. It is known that in case D is (generalized) quaternion all FD- modules are periodic of period dividing 4, but there does not seem to be a good parametrization of the indecomposable modules so far. In case D is semidihedral there exists a recent classification of the indecomposable FD-modules due to Crawley-Boevey, and this is presented in this book (again without proof). The classification of the indecomposable FD-modules in these various cases makes it possible to determine the graph structure of the stable Auslander-Reiten quiver of FD. This is the quiver obtained from the Auslander-Reiten quiver by deleting the vertices corresponding to projective modules and the associated arrows. The next step is to consider Brauer correspondents b of B in \(N_ G(V)\) where V is a non-cyclic subgroup of D. In this case there is a general theory of nilpotent blocks and their extensions due to Puig and the reviewer which could be applied, but the author prefers to take a more elementary approach to compute the corresponding crossed products. After the structure of these Brauer correspondents is clear Green correspondence and new results by Kawata are applied to get the structure of the stable Auslander-Reiten quiver \(\Gamma_ s\) of B: If D is dihedral then \(\Gamma_ s\) consists of 1-tubes, at most two 3-tubes and components of tree class \(A^{\infty}_{\infty}\) or \(\tilde A_{1,2}\); if D is quaternion then \(\Gamma_ s\) consists of 1-tubes and 2-tubes; if D is semidihedral then \(\Gamma_ s\) consists of tubes of rank at most 3 and components of tree class \(A^{\infty}_{\infty}\) and \(D_{\infty}\), and at most one 3-tube occurs. This motivates the following definition: A finite-dimensional algebra A over F is said to be of dihedral (quaternion, semidihedral) type if A is tame, symmetric, connected, with non-singular Cartan matrix and stable Auslander-Reiten quiver as described above. The aim is a classification of these algebras and to recognize the tame blocks among these. In order to classify the algebras the author considers the hearts of indecomposable projective modules and certain modules at ends of components of \(\Gamma_ s\). It turns out that in each case there are at most 3 simple modules. Then the various possibilities for the quiver and the relations are determined. (It is shown that the algebras of dihedral type are special biserial, so their indecomposable modules are known.) In order to filter out the tame blocks the center and the elementary divisors of the Cartan matrix are used. The possible decomposition matrices are determined. As an application, the Brauer-Suzuki theorems on groups with quaternion Sylow 2-subgroups are proved. At the end the results are summarized in form of a table. In my opinion the author has done a good job in making these important developments accessible in form of a book. Some minor comments: The book could have benefitted from a subject index (I couldn't find the definition of a crossed product used on p. 126). A diagram is missing on p. 187. I also noticed some minor mathematical inaccuracies: The lattice of ideals (not right ideals) of an algebra is Morita invariant (p. 4). The treatment of Clifford theory for blocks on p. 133 requires crossed products. On pp. 135-136 one should ignore the statements concerning N. The definition of a subsection on p. 149 requires the Brauer homomorphism (similarly, the comment on top of p. 150 is not correct). I also feel that it should have been said more clearly that tame blocks with only one simple module are nilpotent in the sense of Alperin-Broué, so by a result of Puig they are Morita equivalent to the group algebra of their defect groups.