Equivalences between isolated hypersurface singularities (Q1120704)

Let \({\mathcal O}_{n+1}\) be the ring of germs of holomorphic functions \(({\mathbb{C}}^{n+1},0)\to {\mathbb{C}}\). There are many important equivalence relations that have been defined on the elements of \({\mathcal O}_{n+1}\). \({\mathcal R}\)-, \({\mathcal R}{\mathcal L}\)-, and \({\mathcal K}\)-equivalence are well known in function theory. Each of these equivalence relations can be defined in terms of Lie group action on \({\mathcal O}_{n+1}\). For instance two functions are defined to be \({\mathcal R}\)-equivalent if they are the same up to a holomorphic change of coordinates in the domain. In this case the Lie group acting on \({\mathcal O}_{n+1}\) is the group of all holomorphic change of coordinates preserving the origin. Simple complete characterizations of when two functions are \({\mathcal R}\)-, \({\mathcal R}{\mathcal L}\)-, or \({\mathcal K}\)-equivalent were given by Yau and by Mather-Yau. \({\mathcal Q}\)-, \({\mathcal A}\)-, and \({\mathcal B}\)-equivalence come from singularity theory. These equivalence relations are defined on the basis of algebra isomorphisms. For example, we can associate a C-algebra \({\mathcal O}_{n+1}/\Delta (f)\), the Milnor algebra, to any \(f\in {\mathcal O}_{n+1}\), where \(\Delta\) (f) is the ideal in \({\mathcal O}_{n+1}\) generated by the partial derivatives of f. We say that two functions are \({\mathcal Q}\)-equivalent if their associated Milnor algebras are isomorphic. The purpose of this paper is two-fold. On the one hand, we give a necessary and sufficient condition for \({\mathcal R}{\mathcal L}\)-equivalence coincide with \({\mathcal K}\)-equivalence. This leads us to define the new notion of almost quasi-homogeneous functions. We suspect that the singularities defined by almost quasi-homogeneous functions may form a distinguished class of singularities which have some special properties shared by the quasi-homogeneous one. We discuss the relationship between \({\mathcal K}\)-equivalence and \({\mathcal Q}\)-equivalence. Perhaps the most striking result here is the theorem which provides us a lot of examples with two holomorphic functions having isolated critical points at the origin and identically the same Jacobian ideal, but their zero sets are not biholomorphically equivalent. We also give an example such that \[ \begin{matrix} \mathcal R(f) & \subsetneqq & \mathcal R \mathcal L(f) & \subsetneqq & \mathcal K(f) \\ && \cap\neq \\ && Q(f) \end{matrix} \] with Q(f)\(\not\subseteq K(f)\) and K(f)\(\not\subseteq {\mathcal Q}(f)\). This answers a question raised by G. M. Greuel who asked whether such function exists. The computation of this example is extremely difficult if not impossible at all by hand. In fact we have developed a computer program which allows us to check all these equivalence relations. The examples show the effectiveness of our criteria in checking whether the equivalence classes coincide or not. The results mentioned above together with results obtained by Shoshitaishvili, Mather-Yau and Yau finish completely the problem of equivalence between isolated hypersurface singularities which began in the sixties.