On Heller lattices over ramified extended orders. (Q2570169)

Let \(G\) be a finite group, and let \(R\) be a complete discrete valuation domain with quotient field \(K\) and residue class field \(k\). Starting with the observation that the radical \(\text{Rad\,}P\) of an indecomposable projective \(RG\)-module \(P\) is again indecomposable unless the \(RG\)-block of \(P\) is representation-finite ([Arch. Math. 75, No. 2, 92--97 (2000; Zbl 1015.20005)]; see also \textit{A. Wiedemann} [Commun. Algebra 12, 2525--2578 (1984; Zbl 0566.16014), ibid. 2619--2629 (1984; Zbl 0542.16006), ibid. 3053--3069 (1984; Zbl 0558.16013)] for a more general result), the author studies indecomposable \(RG\)-lattices of the form \(\Omega E\), where \(\Omega\) denotes the Heller operator. In the given example, \(\text{Rad\,}P=\Omega S\), where \(S:=P/\text{Rad\,}P\) is a simple \(kG\)-module. Thus let \(R\) be an extension of a complete discrete valuation domain \(R'\) of ramification index at least 3. Assume that \(\Lambda=R\otimes_{R'}\Lambda'\) for some \(R'\)-order \(\Lambda'\). Let \(\pi\) be a uniformizer of \(R\) and \(\overline\Lambda:=\Lambda/\pi\Lambda\). Under some mild hypothesis, the author shows for a non-projective indecomposable \(\overline\Lambda\)-module \(M\), that \(\Omega M\) is indecomposable provided that \(\Omega_{\overline\Lambda}M\) is so and \(M\) is realizable over the residue class field of \(R'\). In particular, this applies when \(\overline\Lambda\) is a quasi-Frobenius algebra. Under this assumption, the lifting of short exact sequences of \(\overline\Lambda\)-modules to \(\Lambda\)-lattices is studied. For a \(\overline\Lambda\)-module \(M\) with the above properties, it is shown that modulo \(\pi\), the Auslander-Reiten sequence \(\tau(\Omega M)\rightarrowtail E\twoheadrightarrow\Omega M\) reduces to the direct sum of the almost split sequence \(\tau M\rightarrowtail\theta M\twoheadrightarrow M\) and the split short exact sequence \(\Omega M\rightarrowtail\Omega M\oplus\Omega M\twoheadrightarrow\Omega M\). The presentation is rather technical, due to the fact that bases are used instead of invariant concepts. A review of the essential arguments in a more abstract context would certainly clarify which of the results remain valid for ramification index 2.