Differential codimensions and exponential growth (Q6168558)

Let \(A\) be a PI-algebra. If \(A\) has an additional structure, e.g. it is with an involution, group graded or a Lie algebra acts on \(A\) by derivations, then instead of ordinary polynomial identities one studies the polynomial identities which take into account the additional structure. In the paper under review the author studies the case when a Lie algebra \(L\) acts on the algebra \(A\) by derivations, i.e., when \(A\) is an \(L\)-algebra. One of the natural ways to measure the complexity of the polynomial identities of \(A\) is in terms of the asymptotic behaviour of the codimension sequence \(c_n(A)\), \(n=1,2,\ldots\) In the case of differential identities over a field of characteristic 0 it was shown in [\textit{A. S. Gordienko}, J. Algebra 393, 92--101 (2013; Zbl 1307.16019)] that the differential exponent \(\displaystyle \exp^L(A)=\lim_{n\to\infty}\sqrt[n]{c_n^L(A)}\) of a finite dimensional algebra \(A\) exists and is a non-negative integer and it was given an explicit way to compute it. This is an analogue of the celebrated result for ordinary polynomial identities established in [\textit{A. Giambruno} and \textit{M. Zaicev}, Adv. Math. 142, No. 2, 221--243 (1999; Zbl 0920.16013)]. It was conjectured in [\textit{A. S. Gordienko} and \textit{M. V. Kochetov}, Algebr. Represent. Theory 17, No. 2, 539--563 (2014; Zbl 1306.16015)] that if \(\dim(A)<\infty\), then the differential exponent \(\exp^L(A)\) coincides with the ordinary exponent \(\exp(A)\) and this was confirmed for \(L\) being a finite dimensional semisimple Lie algebra. \par The main result of the paper under review is to confirm the conjecture of Gordienko and Kochetov for any Lie algebra \(L\): If \(\dim(A)<\infty\) and the field is of characteristic 0, then \(\exp^L(A)=\exp(A)\). In the case when \(L\) is a finite dimensional solvable Lie algebra (and \(\dim(A)<\infty\)), the author classifies the varieties of \(L\)-algebras of almost polynomial growth, i.e., varieties of exponential growth such that any proper subvariety has polynomial growth. It turns out that there are only two such varieties and they are generated by the algebra of \(2\times 2\) upper triangular matrices: the algebra \(UT_2\) with the trivial action of \(L\) and the algebra \(UT_2^{\varepsilon}\), where \(\varepsilon\) is the inner derivation \(\text{ad}(e_{22})\).