The Lie algebra structure of the first Hochschild cohomology group for monomial algebras (Q1600157)

Let \(k\) be a field of any characteristic, and fix a finite quiver \(Q\). A monomial algebra \(\Lambda\) is a finite dimensional \(k\)-algebra which is isomorphic to \(kQ/I\) where \(kQ\) is the \(k\)-algebra of paths in \(Q\) and \(I\) is some (minimal) ideal generated by paths of length \(\geq 2\). The subset of \(n\) edge paths in \(kQ\) is denoted \(Q_n\). Two paths are called parallel if they the have same starting and terminal vertices. Denote by \(B\) the set of classes in \(kQ/I\) of paths in \(kQ\) which do not contain a subpath in \(Z\) and whose classes therefore form a basis of \(\Lambda\). In the research announcement under review, the author studies the Lie algebra \(H^1(\Lambda,\Lambda)\) of outer derivations of a monomial algebra \(\Lambda\) which is given by the first Hochschild cohomology space. Using a minimal resolution of \(\Lambda\) obtained by \textit{M. J. Bardzell} [J. Algebra 188, No. 1, 69-89 (1997; Zbl 0885.16011)], \(H^1(\Lambda,\Lambda)\) can be shown to be isomorphic to a subquotient of \(k(Q_1//B)\), the \(k\)-vector space generated by a pair of paths \((a,\gamma)\) with \(a\in Q_1\), \(\gamma\in B\) and \(a\) parallel to \(\gamma\). The author then translates the Lie bracket of derivations into a bracket of classes of paths in \(k(Q_1//B)\). The explicit expression is \[ [(a,\gamma),(b,\varepsilon)]=(b,\varepsilon^{(a,\gamma)})-(a,\gamma^{(b,\varepsilon)}). \] Here, \((a,\gamma)\) and \((b,\varepsilon)\) are classes of pairs in \(k(Q_1//B)\), and for \((a,\gamma)\in k(Q_1//B)\) and \(\varepsilon\) a path in \(Q\), \(\varepsilon^{(a,\gamma)}\) denotes the sum of all non-zero classes in \(kQ/I\) of paths obtained by replacing one occurence of \(a\) in \(\varepsilon\) by \(\gamma\). The pairs \((c,\eta)\) are bilinear w.r.t. \(k\) in the two entries. This explicit description of the bracket allows the author to study the graded Lie algebra \(H^1(\Lambda,\Lambda)\) using the geometry of the quiver and to deduce criteria of solvability, semi-simplicity, commutativity and nilpotency. The interested reader is invited to consult the hopefully forthcoming full article.