The strong no loop conjecture for mild algebras. (Q657961)

Let \(\Lambda\) be a finite dimensional associative algebra over an algebraically closed field with a simple module \(S\) of finite dimension. The strong no loop conjecture says that this implies \(\text{Ext}_\Lambda^1(S,S)=0\), i.e. that the quiver of \(\Lambda\) has no loop at the point corresponding to \(S\). The conjecture is known for several families of algebras including: monomial algebras; truncated extensions of semi-simple rings; bound quiver algebras \(\mathbf k\mathcal Q/I\) such that for each loop \(\alpha\in\mathcal Q\) there exists \(n\in\mathbb N\) with \(\alpha^n\in I/(IJ+JI)\) where \(J\) denotes the ideal generated by the arrows; special biserial algebras; and algebras having only two points in their quiver and radical cube zero. In the paper under review the conjecture is proved in case that \(\Lambda\) is mild, which means that \(\Lambda\) has a distributive lattice of two-sided ideals and each proper factor algebra \(\Lambda/J\) is representation-finite. In fact, it is sufficient that a ``small neighborhood'' of the support of the projective cover of \(S\) is mild. The paper is well-written, the proofs are of a constructive nature often involving several cases developed in detail, and many visual examples and graphs are provided and used throughout the paper. All this makes the paper more readable and definitely helps the reader in better grasping the arguments.