Darboux integrability and the inverse integrating factor. (Q1413199)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Darboux integrability and the inverse integrating factor.
scientific article

    Statements

    Darboux integrability and the inverse integrating factor. (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    16 November 2003
    0 references
    Consider planar (real or complex) polynomial vector fields \(X=P(x,y)\partial_x+Q(x,y)\partial_y\) having a Darboux first integral of the form \[ H= f_1^{\lambda_1}\cdots f_p^{\lambda_p} \left( \exp\left( {h_1}\over {g_1} \right) \right)^{\mu_1} \cdots \left( \exp\left({h_q}\over{g_q} \right) \right)^{\mu_q}, \] where \(f_i, g_i\) and \(h_i\) are polynomials and the \(\lambda_i\) and \(\mu_i\) are complex numbers. Recall that, given a first integral \(F\) of a planar differential equation, its associated \textit{inverse integrating factor} is given by \(V_F=-P/(\partial F/\partial y)=Q/(\partial F/\partial x).\) The authors prove that under some additional hypotheses on \(H,\) \(V_{\log (H)}\) is a rational inverse integrating factor of the vector field. They also study conditions to ensure that the differential equation has a polynomial inverse integrating factor. Two nice tools are used along this paper: the notion of \textit{remarkable values} of a rational first integral, a concept already introduced by Poincaré, and of \textit{nth-ecstatic curve,} \({E}_{n}(X),\) associated to a planar vector field \(X,\) that already appears in the work of the Russian mathematician M. N. Lagutinskii (1871--1915). Let us recall them: Given \(H=f/g,\) a rational first integral of \(X,\) it is said that a complex number \(c\) (or infinity) is a remarkable value of \(H\) if \(f+cg\) is a reducible polynomial in the space of complex polynomials. Here, if \(c=\infty\) then \(f+cg\) denotes \(g\). The nth-ecstatic curve is given by \[ E_{n}(X)=\det \left( \begin{matrix} v_1 & v_2 & \cdots & v_l \\ X(v_1) & X(v_2) & \cdots & X(v_l) \\ \vdots & \vdots & \cdots & \vdots \\ X^{l-1}(v_1) & X^{l-1}(v_2) & \cdots & X^{l-1}(v_l) \end{matrix} \right)= 0 \, , \] where \(v_1, v_2, \ldots, v_l\) is any basis of the vector space of complex polynomials in \(x,y\) of degree at most \(n,\) (hence \(l=(n+1)(n+2)/2\)), \(X^0(v_i)= v_i\) and \(X^j(v_i)= X^{j-1}(X(v_i))\). It can be seen that the above definition does not depend on the particular basis chosen. The interesting result is that for a planar vector field \(X,\) \(E_n(X) \equiv 0\) and \(E_{n-1}(X) \not\equiv 0\) if, and only if, \(X\) admits a rational first integral of exact degree \(n\).
    0 references
    0 references
    0 references
    0 references
    0 references
    Darboux integrability
    0 references
    inverse integrating factor
    0 references
    rational first integral
    0 references
    remarkable values
    0 references
    polynomial first integral
    0 references
    ecstatic curve
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references