Noncommutative Noether's problem vs classic Noether's problem (Q2193046): Difference between revisions

This paper investigates the relation between the Noncommutative Noether's problem and the classic Noether's problem. Let \(k\) be any field of characteristic 0 and let \(G\) be a finite group acting linearly on the ring of polynomials \(k[x_1,\dots, x_n]\). Extend the linear action of \(G\) to the action on the field of rational functions \(K_n = k(x_1, \dots , x_n)\). The Classical Noether's Problem (CNP for short) asks whether \(K_n^G\) is a purely transcendental extension of \(k\) or, equivalently, whether the quotient variety \(\mathbb A^n(k)/G\) is birational to \(\mathbb A^m(k)\) for some \(m\). Passing to a noncommutative case consider the \(k\)-algebra of differential operators on the polynomial ring \(k[x_1,\dots, x_n]\), which is the \(n\)-th Weyl algebra \(A_n(k)\), and extend the action of \(G\) to a linear action on \(A_n(k)\). The algebra \(A_n(k)\) is a simple Noetherian Ore domain which admits the skew field of fractions which we will denote by \(F_n(k)\). The action of \(G\) extends naturally to \(F_n(k)\). An analog of the Noether's Problem for the Weyl algebra then asks whether \(F_n(k)^G\) is isomorphic to \(F_m(L)\) for some \(m\geq 1\) and some purely transcendental extension \(L\) of \(k\) of transcendence degree \(t\geq 0\). This we call the Noncommutative Noether's Problem (NNP for short). It was first considered by \textit{J. Alev} and \textit{F. Dumas} [Prog. Math. 243, 21--50 (2006; Zbl 1102.16015)]. In the first main result of this paper it is proven that for any field \(k\) of zero characteristic and any linear action of a finite group \(G\) if the quotient variety \(\mathbb A^n(k)/G\) is rational then the NNP holds for \(G\), that is the CNP implies the NNP. With this result are immediately recovered all previously known cases with positive solution for the NNP, and many new examples when NNP holds. In the case of the complex field the authors generalize the latter result for the rings of differential operators on any affine irreducible variety. Let \(X\) be a complex affine irreducible variety, \(G\) a finite group of automorphisms on \(X\) and let \(D(X)\) be the algebra of differential operators on \(X\). The action of \(G\) on \(X\) extends naturally to the action of \(G\) on \(D(X)\). In the second main result it is proven that if the quotient variety \(X/G\) is birational to an irreducible affine variety \(Y\) then \(D(X)^G\) is birationally equivalent to \(D(Y)\). The authors also give an explicit proof of their main result for all pseudo-reflections groups and an arbitrary field of characteristic zero which allows them to find the Weyl generators by a simple algorithmic procedure. Finally the obtained results are applied to establish the birational equivalence for the cross products.