scientific article; zbMATH DE number 4053870

From MaRDI portal
Publication:3789747





[For the entire collection see Zbl 0631.00003.] Consider an n-dimensional complex Lie algebra L with a fixed basis \(e_ 1,...,e_ n\) and with structural constants \(c^ k_{ij}\in {\mathbb{C}}\), \(i,j,k=1,...,n\), where \([e_ i,e_ j]=\sum_{k}c^ k_{ij}e_ k\). A collection \(c^ k_{ij}\) defines a point in \({\mathbb{C}}^{n^ 3}\). The Jacobi identity and skew-symmetry of multiplication in L shows that the point \((c^ k_{ij})\) belongs to an algebraic affine variety which is defined by equations corresponding to the Jacobi identity and skew symmetry. There is a natural action of the group GL(n,\({\mathbb{C}})\) in this variety \({\mathfrak A}_ n\). The variety \({\mathfrak A}_ 2\) is isomorphic to an affine plane \({\mathbb{C}}^ 2\). The variety \({\mathfrak A}_ 3\) is the union of two irreducible components \({\mathfrak A}_ 3^{(1)}\), \({\mathfrak A}_ 3^{(2)}\) of dimension 6 and degree 1 and 7. \({\mathfrak A}_ 3^{(1)}\simeq {\mathbb{C}}^ 6\) and it splits into four orbits of the group GL(3,\({\mathbb{C}})\) which have dimensions 0,3,5,6. Representatives of these orbits are the abelian algebra \({\mathbb{C}}^ 3\), the Heisenberg algebra, the algebra m(3) of motions of the Euclidean plane and sl(2). The component \(A_ 3^{(2)}\) is not a complete intersection and \(A_ 3^{(1)}\cap A_ 3^{(2)}\) is the union of the first three orbits in \(A_ 3^{(1)}.\) The main result of the paper determines the structure of \({\mathfrak A}_ 4\), \({\mathfrak A}_ 5\) and \({\mathfrak A}_ 6\). It is shown, for example, that there are 4 (respectively, 7 and 17) irreducible components in \({\mathfrak A}_ 4\) (respectively, in \({\mathfrak A}_ 5\) and \({\mathfrak A}_ 6)\). In \({\mathfrak A}_ 5\) there are two components of dimensions 19 and 21, other components have dimension 20. In \({\mathfrak A}_ 6\) two components have dimensions 31 and 32, the other components have dimension 30. In \({\mathfrak A}_ 4\) all components have dimension 12. In each component there are found representatives (representatives of generic GL(n,\({\mathbb{C}})\)-orbits) and their nilpotent radicals. This description is based on the following observations. (1) There exist only finitely many non-isomorphic nilpotent Lie algebras of dimension at most 6; any such algebra admits a \(non\)-\({}_ s(X)\) be the maximal \({\mathbb{Z}}_ 2\)-Eilenberg-MacLane spectrum splitting from bo\(\wedge \overline{bo}^ s\wedge X\). One may think of \(V_ s(X)\) as the \({\mathbb{Z}}_ 2\)-vector space spanned by an \({\mathcal A}^*\)-basis of a maximal \({\mathcal A}^*\)-free submodule of \(H^*(bo\wedge \overline{bo}^ s\wedge X)\). The authors show that \[ \pi_ t(KV_ s(X))\subset E_ 1^{s,t}(X;bo)=\pi_ t(bo\wedge \overline{bo}^ s\wedge X) \] is a subcomplex with respect to \(d_ 1\). Hence one can define the quotient complex (\({\mathcal C}^{*,*}(X),d)\) with \[ {\mathcal C}^{s,*}(X)=E_ 1^{s,*}(X;bo)/V_ s(X), \] the so-called ``bo-essential complex. (\({\mathcal C}^{*,*}(X),d)\) and its homology is the main theme of the paper. If the spectrum X is (bo,H)-prime [for this useful definition see \textit{H. R. Miller}, J. Pure Appl. Algebra 20, 287-312 (1981; Zbl 0459.55012)] then the ``bo-essential homology H(\({\mathcal C}^{*,*}(X),d_ 1)\) can be interpreted as an Ext-functor on a suitable abelian category. The second part of the paper deals with techniques for an effective computation of these Ext-groups. The authors introduce a new ``weight- filtration spectral sequence and restrict their attention to \(S^ 0\), the projective space \(P_ 1^{\infty}\) and the integral Brown-Gitler spectrum B(1) (with cohomology \(H^*(B(1))={\mathcal A}^*/{\mathcal A}^*(Sq^ 1,\chi (Sq^ n);\) \(n>2))\). These computations are then used to give a (complete) proof of the bounded torsion theorem [see the second author, Pac. J. Math. 92, 365-383 (1981; Zbl 0476.55021)] which has known applications to the geometric dimension of vector bundles. Finally the authors present a table of the \(E_ 2\)-term of the bo-ASS for \(\pi^ s_*\) in dimensions t-s\(\leq 20\). This table shows that no higher differentials appear in this range of dimensions.




Cited in
(33)








This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3789747)