Diagonalization in Compact Lie Algebras and a New Proof of a Theorem of Kostant

From MaRDI portal





The author utilizes the Lie algebra \(\mathfrak g\) of a compact connected semisimple Lie group \(G\) and the adjoint action of \(G\) on \(\mathfrak g\) in order to give an algorithmic solution for the problem of diagonalizing a matrix. The described procedure is rather simple; essentially, it reduces the problem to the study of actions of SU(2). The idea is to decompose \(Z\in{\mathfrak g}\) as \(Z= Z_ 0+\sum_{\alpha\in\Sigma^ +} Z_ \alpha\), with respect to the decomposition of \(\mathfrak g\) into root spaces under the adjoint action of a maximal torus. Then the author shows that there exists a sequence \(Z=Z^ 1,Z^ 2,\dots\) of elements of \(\mathfrak g\) in such that (1) \(Z^{n+1}\) is the image of \(Z^ n\) under an orthogonal mapping (with respect to a \(G\)-invariant positive definite form on \(\mathfrak g\)) determined by the component \(Z^ n_ \alpha\) of maximum norm, (2) \(Z^{n+1}\) has no \(\mathfrak g_ \alpha\) component, (3) the distance from \(Z^ n\) to the Lie algebra \(\mathfrak t\) of the maximal torus converges to 0, (4) the sequence \(Z^ n\) converges. Evidently, the sequence \(Z^ n\) converges to an element of \(\mathfrak t\), and the limit represents the desired ``diagonalization of \(Z\). Finally, the algorithm is applied to obtain a new proof of Kostant's theorem that the orthogonal projection of any orbit of an element \(X\in{\mathfrak t}\) onto \(\mathfrak t\) is a convex polytope; namely, the convex hull of the orbit of \(X\) under the Weyl group. In contrast to most of the existing proofs, this new proof is conceptually simple; if neither uses Morse theory nor symplectic geometry.











This page was built for publication: Diagonalization in Compact Lie Algebras and a New Proof of a Theorem of Kostant

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3142869)