Artinian level algebras of codimension 3 (Q659899)

The paper under review fits in the research field that wants to determine whether a given numerical sequence is a level 0-sequence. This question arises frequently in the literature (e.g. [\textit{D. Bernstein} and \textit{A. Iarrobino}, Commun. Algebra 20, No. 8, 2323--2336 (1992; Zbl 0761.13001); \textit{A. Iarrobino}, Trans. Am. Math. Soc. 285, 337--378 (1984; Zbl 0548.13009); \textit{A. V. Geramita, T. Harima, J. C. Migliore} and \textit{Y. S. Shin}, Mem. Am. Math. Soc. 872, 139 p. (2007; Zbl 1121.13019)]). The authors of the present paper have already treated this topic in [\textit{J. Ahn} and \textit{Y. S. Shin}, J. Pure Appl. Algebra 210, No. 3, 855--879 (2007; Zbl 1121.13018)] where they proved that an Artinian algebra with Hilbert function \(H = (1, 3, h_2 , \dots , h_{d-1} , h_d , h_{d+1} )\) with the condition \(h_{d-1} > h_d = h_{d+1}\) cannot be level if \(h_d \leq 2d + 3\) and if \(h_d \geq 2d + 4\) then there is a level 0-sequence of codimension 3 with Hilbert function \(H\) for some value \(h_{d-1}\). In the paper under review they study Artinian level algebras of codimension 3. By the cancellation principle, one knows that \(H = (1, 3, \dots, h_s )\) cannot be a level 0-sequence if \(\beta_{1,d+2} (\mathrm{Gin}(I)) < \beta_{2,d+2} (\mathrm{Gin}(I))\) or if \(\beta_{1,d+2} (I ^{\text{lex}} ) < \beta_{2,d+2} (I ^{\text{lex}} )\) for some \(d < s\). Because of this reason they focus their attention to those Artinian level algebras of codimension 3 with the condition \(\beta_{2,d+2} (I ^{\text{lex}} ) = \beta_{1,d+1} (I ^{\text{lex}} )\). They prove that in this case, \(R/I\) can be an Artinian level \(k\)-algebra only if either \(h_{d-1} < h_d < h_{d+1}\) or \(h_d = h_{d+1} = d + 1\). Moreover they show that if \(h_{d-1} = h_d < h_{d+1}\) then if \(h_d \leq 3d + 2\), then \(H = (1, 3, \dots, h_{d+1} )\) cannot be level and if \(h_d \geq 3d + 3\) then there is a level algebra with Hilbert function \(H\) for some value of \(h_{d+1}\) .