Normal vectors on manifolds of critical points for parametric robustness of equilibrium solutions of ODE systems. (Q1809901)

Considering the parameter space of a physical system described by a nonlinear autonomous ODE, this paper introduces the notion of parametric distance to ``critical'' manifolds. Such manifolds include both the bifurcation sets and the point sets at which the state (or ``phase'' in the dynamics terminology) variables constraints, or output constraints, are violated. The interest of this original approach is to define a physical system by choosing a point of the parameter space insuring the maintenance of a given qualitative behavior with low sensitivity to unavoidable parameter variations due to environmental modifications and the system aging. The point so chosen must be sufficiently far from the ``critical'' manifold, leading to what is called ``robustness'' in engineering, i.e., a kind of structural stability. For this purpose, the authors use normal vectors on manifolds of critical points to measure the distance between these manifolds and a given steady state of the corresponding dynamical system. So, they follow a suggestion of I. Dobson who has given systems of equations to calculate normal vectors for codimension-1 bifurcations. The author's contribution consists more in an extension to ``critical'' manifolds (taking into account physical constraints), than in a generalization to bifurcations of arbitrary codimension. Compared to a previous work, this approach also has the advantage to reduce the number of equations for normal vectors on the Poincaré-Andronov bifurcation manifolds (here, wrongly called ``Hopf'' bifurcation manifold), and to simplify the proof for the regularity of the normal vector system. Illustrative examples in engineering applications are given, formulating an optimization problem, leading to system synthesis.