The representation dimension of quantum complete intersections. (Q947529)

Lower and upper bounds for the representation dimension (a concept due to M.~Auslander) for a certain class of finite dimensional algebras are given. For decades it was not clear whether there exist finite dimensional algebras of representation dimension greater than \(3\) until \textit{R.~Rouquier} has shown [Invent. Math. 165, No. 2, 357-367 (2006; Zbl 1101.18006)] that the exterior algebra of an \(n\)-dimensional \(k\)-vector space has representation dimension \(n+1\). Subsequently several other classes of examples with large representation dimension were constructed and methods for the computation of bounds developed. The paper under review belongs to these contributions. Let \(k\) be a field. A quantum complete intersection \(\Lambda\) (of codimension \(n\)) is a \(k\)-algebra given by generators \(x_1,\dots,x_n\) with relations \(x_i^{a_i}=0\) (\(i=1,\dots,n\)) and \(x_ix_j=q_{ij}x_jx_i\) (\(1\leq i<j\leq n\)), where the exponents \(a_i\geq 2\) and the coefficients \(q_{ij}\) are non-zero elements of \(k\). This is a finite dimensional Frobenius algebra and is in particular selfinjective. The first main result of the paper is that the representation dimension of \(\Lambda\) is at most \(2n\). The second main result treats a special case, where all exponents \(a_i\) are the same integer \(a\geq 2\) and all \(q_{ij}\) are the same element \(q\in k\), and which is an \(a\)-th root of unity. In this case it is shown that the representation dimension of \(\Lambda\) is at least \(n+1\). One essential method which is applied for proving the (more crucial) lower bound is a result of \textit{R.~Rouquier} [Lemma 4.11 in J. K-Theory 1, No. 2, 193-256 (2008; Zbl 1165.18008)] which is also known as the Ghost Lemma.