Exponential codimension growth of PI algebras: an exact estimate (Q1283429)

The paper under review may be considered as the second part of the recent paper by the same authors [\textit{A. Giambruno, M. Zaicev}, Adv. Math. 140, No. 2, 145-155 (1998; Zbl 0920.16012)] reviewed above. If \(A\) is a PI-algebra over a field \(F\) of characteristic 0 and \(c_n(A)\) is the \(n\)-th codimension of (the multilinear polynomial identities of) \(A\), \(n=1,2,\dots\), the authors show the important result that \(\text{Inv}(A)=\lim_{n\to\infty}\root n\of{c_n(A)}\) always exists and is an integer. (In the above mentioned paper they have proved this well known conjecture for finitely generated PI-algebras.) By the theory of Kemer, there exists a finite dimensional \(\mathbb{Z}_2\)-graded algebra \(B\) such that the Grassmann envelope \(G(B)\) of \(B\) satisfies the same (ordinary) polynomial identities as \(A\). As in the first paper, assuming that the algebra \(B\) is given with its Wedderburn-Malcev decomposition as a \(\mathbb{Z}_2\)-graded algebra, the authors present an effective procedure which shows that \(\text{Inv}(A)=\dim C^{(0)}+\dim C^{(1)}\), where \(C^{(0)}+C^{(1)}\) is a suitable \(\mathbb{Z}_2\)-graded semisimple subalgebra of \(B\).