The number of the Gabriel-Roiter measures admitting no direct predecessors over a wild quiver. (Q635447)

Let \(\Lambda\) be a connected Artin algebra and \(\mathrm{mod\,}\Lambda\) be the category of finitely generated left \(\Lambda\)-modules. Given a chain \(M_\bullet\colon M_1\subset M_2\subset\cdots\subset M_t\) of indecomposable submodules of \(M\in\mathrm{mod\,}\Lambda\), the lengths of these submodules \(|M_\bullet|=\{|M_1|,\dots,|M_t|\}\) form a subset of \(\mathbb N\). Define a total order on the set of all subsets of \(\mathbb N\) as follows. For different subsets \(I,J\) of \(\mathbb N\) write \(I<J\), if the smallest element of \((I\setminus J)\cup(J\setminus I)\) belongs to \(J\). Then \(\mu(M)=\max\{|M_\bullet|\}\) is the Gabriel-Roiter (GR, for short) measure of \(M\), where the maximum is taken over all chains \(M_\bullet\) of indecomposable submodules of \(M\). A subset \(I\) of \(\mathbb N\) is called a GR measure for \(\Lambda\) if there is an indecomposable \(M\in\mathrm{mod\,}\Lambda\) with \(\mu(M)=I\). A sequence of GR measures is called a GR segment if it is closed with respect to direct successors and direct predecessors. Assume that \(k\) is an algebraically closed field and \(\Lambda\) is a finite-dimensional \(k\)-algebra of infinite representation type. Then in the paper under review it is conjectured that the following conditions are equivalent: (1) \(\Lambda\) has wild type; (2) \(\Lambda\) admits infinitely many GR segments; (3) there are infinitely many GR measures having no direct predecessors. This conjecture is proven for the wild quiver \(Q\) with three vertices \(1\), \(2\), \(3\) and three arrows \(1\to 2\), \(2\to 3\), \(2\to 3\).