Unique normal form of a class of 3 dimensional vector fields with symmetries (Q2249234): Difference between revisions

Bogdanov-Takens singularity is commonly known as a non-semisimple two dimensional nilpotent singularity given by \[ \begin{aligned} &\dot{x}=y+\sum^\infty_{i+j=2} a_{ij} x^iy^j, \\ &\dot{y}=\sum^\infty_{i+j=2} b_{ij}x^iy^j \end{aligned} \] and is among the most difficult and challenging problems in normal form theory. There have been numerous research results dealing with hypernormalization, unique (simplest), orbital and parametric normal forms and their implementations in computer algebra systems for practical applications; for recent results see [\textit{M. Gazor} and \textit{M. Moazeni}, Discrete Contin. Dyn. Syst. 35, No. 1, 205--224 (2015; Zbl 1308.34046)], [\textit{E. Stróżyna} and \textit{H. Żołądek}, J. Differ. Equations 179, No. 2, 479--537 (2002; Zbl 1005.34034)], [\textit{E. Stróżyna}, Bull. Sci. Math. 126, No. 7, 555--579 (2002; Zbl 1027.34099)], and [\textit{E. Stróżyna} and \textit{H. Żołądek}, J. Differ. Equations 193, No. 1, 239--259 (2003; Zbl 1039.34033)] and the references therein. There still exist open problems on this singularity. Very recently, Stróżyna and Żołądek through a very powerful approach in [\textit{E. Stróżyna} and \textit{H. Żołądek}, The complete formal formal form for the Bogdanov-Takens singularity, to appear in Moscow Mathematical Journal] and [Zbl 1039.34033] have claimed the complete normal form and orbital normal form classification of two dimensional Bogdanov-Takens singularity. For singular systems (reduced on a center manifold) with state dimensions higher than two, the unique normal form problem is substantially more complex than those systems with two dimensional state space. This explains why there have been a few research results on simplest normal forms of such singularities; for example see [\textit{A. Algaba} et al., Int. J. Bifurcation Chaos Appl. Sci. Eng. 8, No. 10, 1857--1887 (1998; Zbl 0996.37062)], [\textit{G. Chen} et al., C. R., Math., Acad. Sci. Paris 336, No. 4, 345--348 (2003; Zbl 1054.34067)], [\textit{M. Gazor} and \textit{F. Mokhtari}, Nonlinearity 26, No. 10, 2809--2832 (2013; Zbl 1283.34038)], [\textit{M. Gazor} et al., J. Differ. Equations 254, No. 3, 1571--1581 (2013; Zbl 1266.34066)], and [\textit{P. Yu} and \textit{Y. Yuan}, Dyn. Contin. Discrete Impuls. Syst., Ser. B, Appl. Algorithms 8, No. 2, 219--249 (2001; Zbl 0985.34029)]. This paper deals with the unique normal form classification for a three dimensional non-semisimple nilpotent singularity \[ \begin{aligned} &\dot{x}=y+f_1(x, y, z), \\ &\dot{y}=\alpha xy +\beta x^3+\gamma xz^2+f_2(x, y, z),\\ &\dot{z}=f_3(x, y, z), \end{aligned}\tag{1} \] with a symmetry defined by \[ f_1(x, y, -z) = f_1(x, y, z), \quad f_2(x, y, -z) = f_2(x, y, z), \quad f_3(x, y, -z) = -f_3(x, y, z), \] for which the \((x, y)\)-plane is an invariant manifold. Further, the authors assume that the nonlinear functions \((f_i)\) for \((i=1, 2, 3)\) do not have terms of \((\delta)\)-grade lower than or equal to \((2)\) for an appropriately defined linear grading function \((\delta)\). The paper under review presents the unique normal forms of the differential system (1) and shows how the two dimensional Bogdanov-Takens normal forms given in [\textit{H. Kokubu} et al., J. Differ. Equations 132, No. 2, 293--318 (1996; Zbl 0876.34040)] can be obtained by their results through restricting their normal form system on the invariant \((x, y)\)-plane. The paper is a very well-written paper, although many subtleties of the theory are suppressed to attract a broad audience. The symmetry, generic and degenerate conditions are designed such that their associated Lie algebra would contain a Lie-isomorphic copy of that of the earlier results on the two dimensional Bogdanov-Takens singularity. This enables them to generalize the results in [Kokubu et al., loc. cit.] to a three dimensional non-semisimple nilpotent singularity. Here, they use a novel block matrix notation that makes it a feasible task to handle the large size matrices appearing in the computations.