Identity bases for finite cyclic semigroups (Q6879559)

The article contains a proof of an old result by the author, which gives identity bases for finite monogenic semigroups. The result was obtained already in 1984 and was published without proof in the survey article [\textit{L. N. Shevrin} and \textit{M. V. Volkov}, Sov. Math. 29, No. 11, 1--64 (1985; Zbl 0629.20029); translation from Izv. Vyssh. Uchebn. Zaved., Mat. 1985, No. 11(282), 3--47 (1985)]. However, the proof remained unpublished and the present article fills this gap.\N\NGiven two positive integers \(h\) and \(d\), let \(C_{h,d} := \langle a \mid a^h = a^{h + d}\rangle\) be the finite monogenic semigroup of index \(h\) and period \(d\). The theorem proved in the article says that an identity basis for \(C_{h,d}\) is formed by the identities \(xy = yx\) and \(x^d x_1 \cdots x_h = x_1 \cdots x_h\) in case \(h \leq d + 2\); in case \(h > d + 2\), an identity basis is formed by the same two identities together with the identities \(x^r y^{r + d} x_1 \cdots x_{h - 3r - d} = x^{r + d} y^r x_1 \cdots x_{h - 3r - d}\) for \(r = 1,\dots,\lfloor(h - d)/3\rfloor\). This also sharpens the result, as formulated in [loc. cit.], by omitting identities that turn out to follow from the remaining ones.