Relative parametrization of linear multidimensional systems

Abstract

In the last chapter of his book "The Algebraic Theory of Modular Systems " published in 1916, F. S. Macaulay developped specific techniques for dealing with " unmixed polynomial ideals " by introducing what he called " inverse systems ". The purpose of this paper is to extend such a point of view to differential modules defined by linear multidimensional systems, that is by linear systems of ordinary differential (OD) or partial differential (PD) equations of any order, with any number of independent variables, any number of unknowns and even with variable coefficients in a differential field. The first and main idea is to replace unmixed polynomial ideals by " pure differential modules ". The second idea is to notice that a module is 0-pure if and only if it is torsion-free and thus if and only if it admits an " absolute parametrization " by means of arbitrary potential like functions, or, equivalently, if it can be embedded into a free module by means of an " absolute localization ". The third idea is to refer to a difficult theorem of algebraic analysis saying that an r-pure module can be embedded into a module of projective dimension equal to r, that is a module admitting a projective resolution with exactly r operators. The fourth and final idea is to establish a link between the use of extension modules for such a purpose and specific formal properties of the underlying multidimensional system through the use of involution and a "relative localization " leading to a "relative parametrization ", that is to the use of potential-like functions satisfying a kind of "minimum differential constraint " limiting, in some sense, the number of independent variables appearing in these functions, in a way similar to the situation met in the Cartan-K\"ahler theorem of analysis. The paper is written in a rather effective self-contained way and we provide many explicit examples that should become test examples for a future use of computer algebra.