Boundary solutions of the classical Yang-Baxter equation (Q1359682)

Let \({\mathfrak g}\) be a finite-dimensional simple Lie algebra over the complex numbers (or split over the real numbers) with Cartan subalgebra \({\mathfrak h}\). Then \textit{A. A. Belavin} and \textit{V. G. Drinfel'd} [Funct. Anal. Appl. 16, No. 3, 159-180 (1982; Zbl 0504.22016)] have given a constructive classification of the unitary solutions of the modified classical Yang-Baxter equation (MCYBE) for \({\mathfrak g}\) by using an admissible triple as a `discrete' parameter and an affine subvariety of \({\mathfrak h}\wedge{\mathfrak h}\) as a `continuous' parameter. In the paper under review the authors discuss certain solutions of the MCYBE with uniquely determined continuous parameter which generalize the Cremmer-Gervais solution for \({\mathfrak sl}(3)\). In the case of \({\mathfrak sl}(n)\) the simple roots can be identified with \(\{1,2,\dots,n-1\}\) in their natural order, and it is shown that a generalized Cremmer-Gervais triple exists only if the omitted root is relatively prime to \(n\). Moreover, there is a complete description of the corresponding solution of the MCYBE. In the above mentioned paper \textit{A. A. Belavin} and \textit{V. G. Drinfel'd} also give a homological interpretation of the unitary solutions of the classical Yang-Baxter equation (CYBE) for \({\mathfrak g}\) which was used by \textit{A. Stolin} [Math. Scand. 69, No. 1, 57-80 (1991; Zbl 0727.17005)] to obtain for \({\mathfrak sl}(n)\) a nonconstructive description of the unitary solutions of the CYBE in terms of quasi-Frobenius Lie subalgebras of \({\mathfrak g}\). Since every even-dimensional abelian subalgebra of \({\mathfrak g}\) is quasi-Frobenius, it seems impossible to obtain in this way a complete classification. Therefore the authors propose to study (and possibly to classify) those unitary solutions of the CYBE which are contained in the Zariski closure of the set of unitary solutions of the MCYBE. They also show for \({\mathfrak sl} (n)\) that these boundary solutions do not exhaust all unitary solutions of the CYBE. The most interesting boundary solutions come from maximal parabolic subalgebras. For \({\mathfrak sl} (n)\) it follows from a result due to \textit{A. G. Ehlashvili} [Funct. Anal. Appl. 16, No. 4, 326-328 (1982; Zbl 0519.17005)] that a maximal parabolic subalgebra is quasi-Frobenius if and only if the omitted root is relatively prime to \(n\). Moreover, then the maximal parabolic subalgebra is Frobenius and the corresponding solution of the CYBE is uniquely determined. The authors conjecture (and prove in general for \(i=1\) and in the case \(n=5\) for \(i=2)\) that these solutions of the CYBE are contained in the closure of a suitable \(SL(n)\)-orbit of a generalized Cremmer-Gervais solution of the MCYBE. It is also shown that for \(i=1\) the generalized Cremmer-Gervais solution of the MCYBE and the boundary solution of the CYBE are part of a natural three-dimensional simple submodule of \({\mathfrak sl}(n)\wedge{\mathfrak sl}(n)\) under the adjoint action of the principal three-dimensional subalgebra of \({\mathfrak sl}(n)\). The authors close the paper by indicating how their results might be extended to establish the above conjecture and by discussing briefly how to obtain corresponding solutions of the quantum Yang-Baxter equation.