On a parabolic algebra \(P\) of \(\check U_q(sl_{n+1})\) and its semi-invariants with respect to the adjoint action of \(P\) (Q1272903)

For an \(n\times n\) Cartan matrix \(A\) let \({\mathfrak g}\) be the Lie algebra associated with \(A\) which in the paper under review is assumed to be simple and finite dimensional. For the related quantum Kac-Moody algebra \(U_q({\mathfrak g})\) generated by \(x_i,y_i,t_i\), \(i=1,\ldots,n\), and with the usual defining relations as in [\textit{A. Joseph} and \textit{G. Letzter}, J. Algebra 153, No. 2, 289-318 (1992; Zbl 0779.17012)], let \(\check U_q({\mathfrak g})\) be the Hopf subalgebra of \(U_q({\mathfrak g})\) such that, as a vector space, \(\check U_q({\mathfrak g})\cong U^-\otimes \check U_0\otimes U^+\), where the subalgebras \(U^-\) and \(U^+\) are generated respectively by \(x_i\) and \(y_i\), \(i=1,\ldots,n\), and \(\check U_0\) is generated by the images of the fundamental weights in the torus generated by \(t_i\), \(i=1,\ldots,n\). The main purpose of the paper under review is to prove quantum analogues of results shown in the classical case in [\textit{J. Dixmier}, Bull. Sci. Math., (2) 100, 57-95 (1976; Zbl 0328.17003)] and [\textit{A. Joseph}, Am. J. Math. 99, 1167-1192 (1977; Zbl 0378.17006)]. First the author constructs two subalgebras \(G\) and \(P\) of \(\check U_q({\mathfrak g})\) such that as a vector space \(G\otimes P\) is isomorphic to \(\check U_q({\mathfrak g})\). (This was also established by other methods in [\textit{M. S. Kébé}, C. R. Acad. Sci., Paris, Sér. I 322, 1-4 (1996; Zbl 0859.17009)] and in a yet unpublished paper by \textit{V. Baldoni, D. Barbasch} and \textit{P. Möseneder-Frajria}.) Then the author considers the case \({\mathfrak g}=sl_{n+1}\) and expresses the generators of the centre of \(\check U_q(sl_{n+1})\) as linear combinations of generators of \(G\) with coefficients in \(P\). Moreover, it is established that the vector space generated by the semi-invariants of \(P\) with respect to its adjoint action, is the polynomial algebra in one variable \(d\) with a description similar to that given by Joseph in the classical case. The author has mentioned that the methods used in the present paper are different from those used by Dixmier and the specific manipulations in the quantum case make some of the proofs easier than in the classical case.