Lie algebras admitting elliptic semialgebras (Q1085262)

Let L be a real Lie algebra with bracket [, ] and dim L\(<\infty\). In some neighbourhood of the point \(0\in L\) one can define a group operation by the Campbell-Hausdorff formula. A convex cone \(K\subset L\) with a vertex 0 is called a semialgebra if it is a local semigroup. It is natural to call a semialgebra which is an elliptic cone an elliptic semialgebra. A metric \(<, >\) in L is said to be invariant if for all a,b,c\(\in L\) holds \(<[a,b],c>=<b,[c,a]>.\) Theorem. If a Lie algebra has an elliptic semialgebra K, then either K is given by an invariant Lorentz metric or L is a basic affine algebra (i.e. is the Lie algebra of the group of affine transformations of \({\mathbb{R}}^{n-1}\) generated by translations and homotheties). Corollary. If an exponential Lie algebra L has an elliptic semialgebra K, then L is either commutative or a basic affine algebra. The order in a basic affine algebra was studied by \textit{K. H. Hofmann} and \textit{J. Lawson} [Semigroup Forum 23, 343-358 (1981; Zbl 0531.22002)], and the reviewer [Sov. Math., Dokl. 15, 416-419 (1975); translation from Dokl. Akad. Nauk SSSR 215, 35-37 (1974; Zbl 0319.50015); Sib. Mat. Zh. 27, No.3(157), 51-67 (1986; Zbl 0599.51016)].