scientific article; zbMATH DE number 2233304

From MaRDI portal
Publication:5709076





Under- or overdetermined systems of differential equations appear in many applications. Examples comprise in physics the Yang-Mills, Einstein or Navier-Stokes equations (more generally any kind of incompressible continuum or otherwise constrained system), in mathematics differential geometry is a rich source. The development of theories for treating such general classes of systems has started already more than a century ago; for more algebraic approaches one may mention in particular Janet, Riquier and Ritt (the founder of differential algebra), on the differential geometric side E. Cartan, Kähler and Kuranishi (using exterior differential systems). However, many researchers in differential equations seem to be unaware of the existence of these theories, as the author rightly remarks in the introduction of the book. This unfortunate situation is particularly evident in the Dirac theory of constrained mechanical systems or in the numerical analysis of differential algebraic equations: in both cases the rather straightforward analysis of ordinary differential equations has been rediscovered without any reference to the old works whereas the much harder treatment of partial differential equations mainly relies on ad hoc techniques inspired from the ordinary case which are however not sufficient in general. A key problem in dealing with an overdetermined system is the possible existence of hidden integrability conditions. While it is not difficult to exhibit some integrability conditions, it is much harder to decide whether one has found all of them. As these conditions represents obstructions to the construction of formal power series solutions, it is important to be able to make this decision. It turns out that for many purposes, like e.g. the proof of an existence and uniqueness theorem for analytic solutions, formal integrability is not sufficient but one needs the stronger concept of an involutive system. This notion appeared already in the 19th century but obtained its definite form only in the work of E. Cartan; the modern homological approach used in the book is mainly due to Matsushima and Spencer. The book may be divided into two parts. The first part, consisting of the first three chapters, contains an introduction into the formal theory and the concept of an involutive system. Chapter I discusses the required tools from commutative algebra like the Koszul complex and syzygies. It introduces the notion of an involutive module using the Koszul approach of Serre which is dual to the classical one via Spencer cohomology and permits a very simple finiteness proof. Furthermore, the homological version of the famous Cartan test for an involutive module is given. Chapter II starts the analysis of differential equations. The author uses throughout the language of global analytic geometry, i.e. all variables are assumed to be complex valued, all functions to be holomorphic and a differential equation is represented by a coherent sheaf of ideals. The characteristic module is introduced as the factor module of the symbol and the notion of an involutive system arises naturally during the discussion of the prolongation of formal solutions. Chapter III is concerned with the Cartan-Kähler theorem, the basic existence and uniqueness theorem for analytic solutions of analytic involutive differential systems. The given proof is not by reduction to the Cauchy-Kovalevskaya theorem; instead first the existence of a formal solution is shown and then its convergence under the assumption that the initial data are convergent. The second part of the book, consisting of Chapters IV and V, centres around the question of generic involutivity and is intended to elaborate on the rather vague statement by Cartan that ``under certain conditions which are not easy to make precise every system becomes eventually involutive. It combines and revises some recent articles by the author. As a basic tool the notion of a differential or \(D\)-variety (some authors also speak of a ``diffiety) is introduced and discussed. The main results of this part are two finiteness theorems. The first one states that ascending chains of pseudo-coherent radical differential ideals on a \(D\)-variety become stationary (this may be interpreted as a reformulation of the Ritt-Raudenbush theorem in differential algebra). The second one states that outside of a singularity set every differential system becomes involutive after a finite number of prolongations (thus it represents a variant of the Cartan-Kuranishi theorem in the context of \(D\)-varieties). A great problem in the theory of general differential systems is the existence of many different approaches using very different languages. A nice feature of the book is that the author tries to make transparent the relations between some of these approaches, which are usually much closer than many people think. In particular, a whole appendix is devoted to proving the equivalence of the formal theory of partial differential equations and the Cartan-Kähler theory of exterior differential systems. The first author of this book is one of the leading experts in the domain of general systems of partial differential equations and his insights should be useful for everybody interested in the topic. Applied mathematicians may find the language difficult, but for some emerging applications of the formal theory like a nonlinear differential Galois theory it is very useful. In conclusion, the book is a very welcome addition to the rather small literature in this field.




Cited in
(24)








This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5709076)