The moduli space of complex 5-dimensional Lie algebras (Q290447)

The philosophy in the article is that any moduli space of algebras of a fixed dimension, can be naturally stratified by orbifolds. The meaning of this is of course not very stringent as we don't know if there exists a moduli space, and the stratification in orbifolds could be given in several ways. However, assuming that a moduli does exist, and letting the orbifolds be equivariant versal families, the deformation theory and cohomology in particular cases, gives possible constructions of, and knowledge about moduli. In the case of complex Lie algebras, the orbifolds consist of quasi-projective spaces. They are obtained by removing a divisor from \(\mathbb P^n_{\mathbb C}\) together with an action of a symmetric group. The main result of the article is that the strata that consists of more that one point can be given projective coordinates. The stratification is given in a unique form by applying deformation theory: The strata are given as smooth deformations, and the strata are connected by jump deformations. The authors also state that deformations are smooth along the families of points to which they jump, which is a very good property of the stratification. The authors find that each family (with more than one point) of solvable complex Lie algebras of dimension \(d\), \(d\leq 5\), contains one special nilpotent element. Then each family can be given by the action of a symmetric group on \(\mathbb P^n_{\mathbb C}\), and the generic point of \(\mathbb P^n_{\mathbb C}\) corresponds to the nilpotent. The meaning of this is that if \(\mathbb P^n_{\mathbb C}\) is given by projective coordinates \((p_0,\dots,p_n)\), then the generic point is \((0,\dots,0)\). The authors claim that this point is usually excluded by algebraic geometers, but it has to be present in this picture as there is an algebra corresponding to the generic point. Thus we can read out of the article that by geometric invariant theory there is a dense orbit, so that there does not exist an algebraic quotient. The generic element in one family may be isomorphic to the generic point in another family, and this represents the only overlap between the families. The authors point out that, contrary to earlier results by them and others, the goal of this article is not to simply determine a list of the algebras, but also to understand how the moduli space is glued together. Then the cohomology of the algebras has to be computed, and in addition the versal deformations of the algebras. The result is a more natural decomposition of the moduli space, and divides it into fewer strata than earlier results. The article starts by applying this new view to make a summary of the moduli spaces of \(3\) and \(4\)-dimensional algebras. The lists are given with the dimension of the cohomology of interest, and the versal deformations are included to some extent. Then a more thorough discussion of the \(5\)-dimensional complex Lie algebras is given: Given an exact sequence \(0\rightarrow M\rightarrow L\rightarrow W\rightarrow 0\) where \(L\) is a Lie-algebra which is an extension of \(W\) by \(M\). \(M\) is an ideal in \(L\), and \(W\) is the quotient algebra \(W=L/M\). The algebra structures of \(M\), \(L\), and \(W\) is denoted \(\mu\), \(d\), \(\delta\) respectively. One can write \(d=\delta+\mu+\lambda+\psi\) where \(\lambda\) and \(\mu\) are additional algebra structures. Then \(d\) is a algebra structure exactly when the \textit{compatibility condition}, the \textit{Maurer-Cartan condition}, and the \textit{cocycle condition} are satisfied. This is the basis of the classification of the algebras, and the computation of the versal deformation which is the main result of the article. The article is easy to read, it classifies the complex Lie algebras of dimension five, and gives the essential information about the moduli space. The article serves as a main step to the general classification of finite dimensional complex Lie algebras.