Normal nilpotent varieties in \(F_4\) (Q1270040)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Normal nilpotent varieties in \(F_4\) |
scientific article |
Statements
Normal nilpotent varieties in \(F_4\) (English)
0 references
9 June 1999
0 references
This paper deals with the question whether a given nilpotent variety \(N\) in a complex simple Lie algebra is normal or not. Normality is a very desirable property for \(N\) since it implies rational singularities and the Gorenstein property. Kostant showed first that the full nilpotent variety is always normal. Then Kraft and Procesi showed that all nilpotent varieties in a Lie algebra of type \(A_\ell\) are normal. Afterwards the case of the other classical simple Lie algebras was settled. Then non-normal varieties do appear. For the exceptional types many cases are still open. The author settles the normality problem for all nilpotent varieties in a simple Lie algebra of type \(F_4\). The method of proof is based on cohomological principles and comparison of multiplicities leading to three normality criteria. They can be programmed on a computer. Here in particular the first criterion turns out to be useful for computer calculations. Let \(G\) denote the adjoint group of \(\text{lie }g\) with \(\mathbb Z\)-grading \(\text{lie } g=\bigoplus _i \text{lie } g_i\), \(P\) the parabolic subgroup of \(G\) corresponding to \(\text{lie } g=\bigoplus_{i\geq 0}\text{lie } g_i\) and \(\text{lie } m\) the nilpotent subalgebra \(\bigoplus_{i\geq d}\text{lie } g_i\). A characteristic \((D,d)\) is called good, if \(\dim N(D,d)=\dim \text{lie } g+\dim\text{lie } m-\dim \text{lie } p\). The criterion is: for a good characteristic \((D,d)\) the variety \(N(D,d)\) equals the corresponding normal affine variety \(\widetilde{N}(D,d)\) if the following condition is satisfied: suppose that the nonzero dominant weight \(\lambda\) occurs in \(H^i(G/P,\text{lie } L _{G/P}(\Lambda ^i(\text{lie } g/\text{lie } m)^{\widetilde{\;}})^*)\) for \(i >0\). Then this dominant weight does not occur in the homogeneous part of \(\mathbb C [N]\) of degree \(i\). In the last part of the paper the results of the computations are presented. According to the author the hardest case for type \(F_4\) is to show the normality of the nilpotent variety with characteristics \((0,0,1,0)\).
0 references
normal variety
0 references
nilpotent variety
0 references
complex simple Lie algebra
0 references
computer calculations
0 references
0 references