The bernsteim polynomial in one variable

From MaRDI portal





The study of the structure of \(D\)-modules led I. N. Bernstein to the nice existence theorem about differential operators with polynomial coefficients, referred as the lemma of \(b\)-functions: there exists a polynomial \(d(\lambda)\), with coefficients in the Weyl algebra \(A_n(K)[\lambda]\) of \(n\) variables, and a nonzero polynomial \(h(\lambda)\in K(\lambda)\) such that \(d(\lambda)*p^{\lambda+1}=h(\lambda)p^\lambda\), where \(p\) is a fixed polynomial of \(n\) variables. Recently an algorithm of computing \(b\)-functions was provided [see: \textit{T. Oaku}, Duke Math. J. 87, No. 1, 115-132 (1997; Zbl 0893.32009)]. In the paper under review the case of one variable \(x\) is considered. More precisely, let \(K\) be a field of characteristic zero and \(p=p(x)\) be a non-constant polynomial in \(K[x]\). We shall consider the set \(\{P_1,\dots,P_f\}\) of all distinct monic irreducible factors of \(p(x)\). For each \(r\), \(1\leq r\leq j\), we set \(c_r(\lambda)=(n_r\lambda+n_r)\) \((n_r\lambda+n_r-1)\dots(n_r\lambda+1)\). Let \(Q\) be the subfield of rational numbers of \(K\), then the authors' result says that if \(c(\lambda)\in Q[\lambda]\) denotes the last common multiple of the polynomial \(c_r(\lambda)\), then \(d(\lambda)\bullet p^{\lambda+1}=c(\lambda)p^\lambda\), where \(d(\lambda)\in{\mathcal A}(K)[\lambda]\). By \(A(K)[\lambda]\) is denoted the algebra of polynomials in \(\lambda\) with coefficients in the Weyl algebra \(A(K)\) of differential operators in one variable over \(K[x]\). It is shown how to construct \(d(\lambda)\). After this the authors prove that \(c(\lambda)\) is a \(b\)-function, i.e. \(c(\lambda)=b(\lambda)\).











This page was built for publication: The bernsteim polynomial in one variable

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4224020)